From Wikisource
Jump to navigation Jump to search
Scriptorium Scriptorium (Help) Archives, Last archive
The Scriptorium is Wikisource's community discussion page. This subpage is especially designated for requests for help from more experienced Wikisourcers. Feel free to ask questions or leave comments. You may join any current discussion or a new one. Project members can often be found in the #wikisource IRC channel (a web client is available).

This page is automatically archived by Wikisource-bot

Have you seen our help pages and FAQs?


Academic journal articles[edit]

Is there any standard way to format the metadata for academic articles? I'm thinking something like Wikiversity:Template:Article info. If not, would it be OK to just copy over that template? HLHJ (talk) 01:52, 20 May 2019 (UTC)

Normally we would just put that info in Wikidata, and link to it from there. —Beleg Tâl (talk) 16:01, 22 May 2019 (UTC)
Thank you, Beleg Tâl. I'm trying to figure out the best method for adding entire copyleft fulltext academic articles to Wikisource. I have an application for this, discussed on the Wikiproject Medicine group (the group is currently supplying Internet-in-a-box wiki subsets to off-the-net medical facilities). Is there a good way to link such a fulltext to the Wikidata metadata? Presumably you'd want to display that metadata here somehow, but hotlinking might be better. Is there an extant example I could look at to figure out how to do this? HLHJ (talk) 22:20, 1 June 2019 (UTC)
Texts on Wikisource are linked to items on Wikidata using the interwiki links functionality on Wikidata. See for example d:Q15625490 which is linked to Biodiversity Assessment of the Fishes of Saba Bank Atoll, Netherlands Antilles. —Beleg Tâl (talk) 20:31, 5 June 2019 (UTC)
Thank you very much, Beleg Tâl, that's just what I wanted. I have looked through it and am looking at the tools credited in its edit history. I also uploaded another rather random non-medical article, but now I look at it more closely, I'm not sure the source (a journal which seems to be a research group's own publication) is reliable... I'll come back to this. HLHJ (talk) 04:10, 21 June 2019 (UTC)

Fixing problematic code (alignment in captions)[edit]

I've been informed that a number of my edits have introduced an HTML problem of some kind. (I'm no HTML expert, so I don't fully understand what it's about.) I'll put an example below. Can anybody suggest a better way to code such captions? I'd like to continue to use {{FI}} template if possible.

Letters from an Oregon Ranch p. 8.png

Copyright, Kiser Bros., Portland, Ore.

"We can plainly see Mount Jefferson" (page 46)

This applies to most of the images in Letters From an Oregon Ranch and a number of other works as well. Of course, it would also be good to reduce the amount of space between the image and the caption...maybe a solution could address that as well? -Pete (talk) 19:26, 18 June 2019 (UTC)

@Peteforsyth: Some comments if I may? The caption parameter of {{FI}} always tries to enclose its content in an HTML paragraph (<p>) tag which regretably is somewhat at odds with use of {{right}} as that template in turn uses a division (<div>) tag to do its magic. This is one of the relatively few verboten HTML constructs which may be the origin of the original complaint…and also indirectly explains the gap betwixt image and caption you mentioned! (Oh and {{c}} adds yet another <div>… but by now the damage has already been done and cue parable spilt milk &c.)

May I humbly propose something like:

|file=Letters from an Oregon Ranch p. 8.png
|caption={{sm|Copyright, Kiser Bros., Portland, Ore.}}</p>
"We can plainly see Mount Jefferson" (page 46)}}
which displays as:

Letters from an Oregon Ranch p. 8.png

Copyright, Kiser Bros., Portland, Ore.

"We can plainly see Mount Jefferson" (page 46)

Now for a little explanation (as I have cheated a little bit!) talign=right is used solely to align the "Copyright" line and its effect ends at the </p> fragment. tstyle=margin:0 is used to eliminate padding around and within the caption (everybody forgets that paragraphs always add typically around 7px margin-top and margin-bottom!) Finally {{p|ac|m0}} adds a new paragraph with center aligning and margins (re)suppressed to hold the main part of the caption; i.e. "MOUNT JEFFERSON &c.
Hope this is useful? 08:56, 19 June 2019 (UTC)
Many thanks, this helps me understand the problem and I think it's a good solution. In an ideal world, I tink there would be a little more space between the (c) notice and the actual caption (and/or a little less space between the caption's two lines). Maybe if I study the template you introduced I can find a way to accomplish this; but I wouldn't say it's mission-critical. Much appreciated! -Pete (talk) 19:12, 19 June 2019 (UTC)
Also, just out of curiosity -- won't this "cheat" introduce an extraneous closing paragraph tag? I assume that's simply ignored, and therefore not a problem -- but curious if I'm interpreting the code correctly. -Pete (talk) 19:13, 19 June 2019 (UTC)
Yes it does create another issues with mismatched tags, the REAL solution would be to fix {{img float}} so it can handle the block level elements correctly, but given past experience I don't see that happening certain contributors are aggressively dissuaded from "too clever by half" solution , as opposed to actual repairs to the underlying templates. ShakespeareFan00 (talk) 19:24, 19 June 2019 (UTC)
@Peteforsyth: Regarding the line spacing in the caption issue; is this more to your liking?
|file=Letters from an Oregon Ranch p. 8.png
|caption={{sm|Copyright, Kiser Bros., Portland, Ore.}}</p>
{{p|class=imgCaption|ac}}MOUNT JEFFERSON, FROM HOOVER'S BUTTE<br />
"We can plainly see Mount Jefferson" (page 46)}}
which displays as:

Letters from an Oregon Ranch p. 8.png

Copyright, Kiser Bros., Portland, Ore.

"We can plainly see Mount Jefferson" (page 46)

I dropped the m0 as I misunderstood your requirement—which reintroduces that 7px inter-paragraph margin I referred to before. As for the concern about extraneous closing paragraph tags the balance between opening and closing tags is maintained thus:
  • <p> is generated internally by {{FI}} just after image and in front of caption—furthermore this is the paragraph which conveys the talign and tstyle information into HTML, and is closed by…
  • </p> this explicitly coded tag ends the Copyright line fragment and also ends right-alignment…
  • A second <p> is generated by {{p}} (Also a side note here: I added class=imgCaption as I missed the (slight) issue of {{FI}} marking the caption block with this particular class name and thought I might as well make the two paragraphs as similar as possible in this respect.)
  • </p> is generated internally by {{FI}} to close off the caption block. This is the cheat in that {{FI}} "thinks" it is encapsulating a single parargraph's text and in fact here it has been handed two. This is not an issue because {{FI}} furthermore encloses the whole image+caption entity within a <div> block and thus the whole mess hangs together!
@ShakespeareFan00: I hope the above exposition allays your concerns. {{FI}} is a surprisingly flexible and robust template and if anything ought to act as a model for modifications to other templates such as {{img float}}… However the two templates do not satisfy the same purposes at all and are not generically interchangeable. {{FI}} is superior for situations where a (possibly variable size) image+caption stands alone either within a page or floats off to one side. {{img float}} lends itself better to text winding around a captioned image of fixed dimensions. Not at all the same animals!
Thank you for prompting this little stroll down memory lane… 22:51, 19 June 2019 (UTC)
Thanks very much, this is all very helpful -- both to the immediate task at hand, and also giving me a nice lesson in the underlying issues. I'll start implementing this shortly! -Pete (talk) 07:25, 21 June 2019 (UTC)

{{Ditto}} and {{hi}} incompatibility...[edit]

Page:The Prince (translated by William K. Marriott).djvu/303

Using {{hi/s}} {{hi/e}} causes {{ditto}} to fail. Please fix the templates, or provide a STABLE way of doing this. ShakespeareFan00 (talk) 08:52, 21 June 2019 (UTC)

{{hi/s}}<span style="display:inline-block; text-indent: 0;">{{ditto|Here is a STABLE way of doing this.}}</span>{{hi/e}} Beleg Tâl (talk) 03:53, 22 June 2019 (UTC)
That is implemented as : , I've also tried it with a table.

I'd like a second opinion on which approach would be better to use before formatting all the Notes pages. ShakespeareFan00 (talk) 08:29, 22 June 2019 (UTC)

Template styles to reduce calls to ts...[edit]

Template:Os Lusiadas (Burton, 1880)/errata.css This is a useful technique, Is someone considering documenting this so it can be used more widely? ShakespeareFan00 (talk) 15:23, 21 June 2019 (UTC)

@ShakespeareFan00: (and CC Inductiveload) There is documentation for TemplateStyles itself at mw:Help:TemplateStyles. It's not much used here yet because until very recently there were serious bugs in MediaWiki when it was used, and that broke stuff here on enWS. Those have now been fixed so as a general rule TemplateStyles should be safe to use.
However, note that the intended use for mw:Extension:TemplateStyles is in actual templates, and use elsewhere is not necessarily a good idea. Even if it does not create any other technical issues, I am uncertain we would want a work, like The Lusiads (tr. Burton), in mainspace to have an associated custom stylesheet over in the Template: namespace (i.e. Template:Os Lusiadas (Burton, 1880)/errata.css). Until some kind of practiice and guidelines is figured out for this use case I would recommend using it only very sparingly and only when the benefits are significant (avoiding a large number of {{ts}} calls might be one such case).
The method does look like a very useful one with many potential benefits for the Wikisources though. Perhaps it would be worthwhile to poke Tpt and Anomie and see if something tailor-made for Wikisource could be made based on ProofreadPage and TemplateStyles? There is that "CSS" field in the Index: that only kinda sorta does work sometimes maybe. Perhaps it would make sense to have per-work custom CSS in a subpage of that work's Index:, and use the "CSS" field on the work's index to load it?
Whether or not we would want to actually allow per-work CSS is a different matter, and one that would need discussion on the Scriptorium to determine. I imagine the benefits might be obvious for larger works like the DNB, but more controversial for most smaller individual works. --Xover (talk) 16:23, 21 June 2019 (UTC)
@Tpt: @Anomie: , Let's start the discussion.. I know with some works I did that I was using a LOT of {{ts}} calls, when some form of advanced per work classing would be more appropriate, assuming it was possible..
It's definitely a nice idea. I have opened a task about it. Tpt (talk) 17:38, 21 June 2019 (UTC)
FYI, you should be able to do <templatestyles src="Index:Os Lusiadas (Burton, 1880)/errata.css" /> to include Index:Os Lusiadas (Burton, 1880)/errata.css. You should be able to convert that Index page to the "Sanitized CSS" content model using Special:ChangeContentModel, or by moving the existing Template:Os Lusiadas (Burton, 1880)/errata.css with that content model to the new title. Anomie (talk) 13:39, 22 June 2019 (UTC)
@Anomie: - Thanks, but I think Index: pages generate their own edit form, which would need to be DISABLED if whats being seen is in fact the per work styles.. 13:44, 22 June 2019 (UTC)

Note that any transcluded styles for Page: 's would also need to be made in such a way that the CSS concerned could also be used in Main namespace, but I'm not sure how you would be able to write CSS that was transclusion aware.. ShakespeareFan00 (talk) 16:28, 21 June 2019 (UTC)

ShakespeareFan00 (talk) 16:28, 21 June 2019 (UTC)

(Aside) It would be nice in time (i.e when anyone cares to implement it) if {{ts}} was re-written slightly, so that adding new codes to it and other templates like {{tf/s}} {{span}} {{p}} etc. could be done WITHOUT having to add new clauses to the MASSIVE switch statements inside them. ShakespeareFan00 (talk) 16:45, 21 June 2019 (UTC)
If necessary, you should be able to use body.ns-0 as part of a selector to apply only on main-namespace pages, and similarly body.ns-104 as part of a selector to apply only on Page-namespace pages. Anomie (talk) 13:39, 22 June 2019 (UTC)
In terms of name-mangling if needed, could you do something like __indexpage_id__localclassname?

ShakespeareFan00 (talk) 13:44, 22 June 2019 (UTC)

In terms of what's required to use this right now:
  • If it's acceptable to put the CSS styles in the Template namespace, nothing needs to be done right now.
  • If it's preferred to put per-work styles in the Index: namespace (which I think is probably a good idea) then any of these will suffice:
    • Set $wgTemplateStylesNamespaces so that Index namespace CSS pages get the Sanitised-CSS content model by default (IMO, this is best), or
    • Permit non-administrators to change page content models (to work around the lack of default). This can be either by a change of config, or a resolution of phabricator:T85847.
    • Put them in the Template: namespace and move them later when one of the above is done.
In terms of nice-to haves, a resolution of phabricator:T226275 would allow the <pages/> tag to pull in a style that every page of a work uses. But this can be worked around for now by manually placing the shared <templatestyles/> tag before the <pages/> tag and adding to the auto-header in the index page. THis would be easy to bot out later, and harmless if left in place.
Inductiveloadtalk/contribs 14:19, 2 August 2019 (UTC)

Changing the title of a page[edit]

I just edited the page for John Donne's epigram "Hero and Leander" - it was incorrectly spelled as "Hero and Leader." I changed the title in the page entry, but "Hero and Leader" is still appearing, and a corrected link goes to the Marlow poem instead of the Donne epigram. Can anyone tell me how to fix this (or do I just need to wait for the change to percolate through)

Thanks unsigned comment by Dkohen (talk) .

Yes check.svg Done , all taken care of now —Beleg Tâl (talk) 03:45, 22 June 2019 (UTC)

Page:The Prince (translated by William K. Marriott).djvu/311[edit]

The approach I am using here works, but adding {{p}} tags for each entry would be inefficient. @Xover: is this an approach where a TemplateStyles approach to redefine the P handling for each block of entries could be implemented for this index? It's mostly a simple margin collapse, and combination of that with the CSS style in the dent. It would be nice to be able to customise the indentation and paragraph break spacing behaviour, but I'm not sure TemplateStyles was advanced enough to handle paramatised generation of CSS per called block yet. ShakespeareFan00 (talk) 08:51, 22 June 2019 (UTC)

@ShakespeareFan00: Provided I understand correctly what you mean, this sounds like it could be reasonably solved by using <p class="no-vertical-space"></p> and a per-work stylesheet that removed the margins. That is, in the page markup you use a template that tags it as a particular class name, and then write custom CSS that styles all instances of that class however you want. Provided each paragraph is styled the same, you would not need actually parametrized CSS for this.
For practical reasons we would probably want to require some kind of prefix on the CSS class name (to avoid collisions), or some magic in MW (ProofreadPage and TemplateStyles) that takes care of the scoping. At small volume, for simple cases, and CSS that is used in a single work, this is not likely to be an actual problem; but worth keeping in mind. --Xover (talk) 09:13, 22 June 2019 (UTC)
Yes but I'd like to avoid having to type the P class=foo stuff individually for each paragraph, Ideally what I'd like is {{p-collapse/s}} and {{p-collapse/e}} I can apply over an entire block, which is at most 4 calls per page as opposed to a considerable number of {{p}} calls. ShakespeareFan00 (talk) 09:18, 22 June 2019 (UTC)
Probably doable by using <div class="no-vertical-space"></div> around all the paragraphs and then using a stylesheet that applies the style to > p. But since the actual paragraphs are then generated by Mediawiki there is some fragility here that may or may not be a problem in practice. --Xover (talk) 10:24, 22 June 2019 (UTC)
The other concern, is that the final paragraph of the block would need to revert back to normal behaviour at the bottom. ShakespeareFan00 (talk) 10:42, 22 June 2019 (UTC)
See - Page:The Prince (translated by William K. Marriott).djvu/311 there is a gap between the A and B block entries.
Also- [1] Why is the bottom entry mis-aligned, I was using the /s /e variants to accommodate the split layout?
It seems that at the end of the page it fails to wrap the end paragraph. This is a LONG-standing issue that no-one seems to have wanted to tackle prviosuly. Using {{nop}} in the footer did not change the behaviour. I've also asked at least TWICE for the whitespace behaviour to be PROPERLY DOCUMENTED so I'm not playing hunt the "pedantic minutiae" MANY, MANY, TIMES. ShakespeareFan00 (talk) 11:16, 22 June 2019 (UTC)
@Xover: can you take a look at the experimental version, and tweak? Thanks..ShakespeareFan00 (talk) 13:40, 22 June 2019 (UTC)
@ShakespeareFan00: You'll need to give me more to work with. What pages am I looking at; what am I looking for there; and what needs tweaking? --Xover (talk) 13:50, 22 June 2019 (UTC)
@Xover: Template:P-collapse/s/style.css being the style-sheet ShakespeareFan00 (talk) 14:03, 22 June 2019 (UTC)
On Page:The Prince (translated by William K. Marriott).djvu/311 The text in the scan has a gap between the A and B block of entries. In the code I've ended one P-collapse block and opened a new one.
What needs to happen in the CSS style, is that the margin behaviour for the top margin/padding reverts to normal for the first entry in a collapsed block (i.e the first entry of the B block) ,and conversely needs to revert to normal for the bottom margin in the final paragraph of a block (i.e the end of the paragrpah representing the last entry of the block of A entries.) I had a possible approach commented in the style sheet, but wasn't sure how to cancel a previously defined style back to the normal/default inherited values. ShakespeareFan00 (talk) 14:03, 22 June 2019 (UTC)
You should be able to do that by simply adding a margin to the div itself. I've done that: take a look and see if that does it? --Xover (talk) 14:20, 22 June 2019 (UTC)
Thanks, Also added some rules for the first and last P elements, hope it doesn't add additional space, I an re-comment them if needed. Seems to work so far.

Text Indentation handling[edit]

@Xover: Relating to Page:The Prince (translated by William K. Marriott).djvu/313 Currently I am using {{dent/s}}, {{dent/e}} to set up the text indent, however here there is a slight issue with this, given that the start of the page is a continuation of the previous entry and so as such doesn't need the indent. The slight discrepancy can be worked with for now, but in time it would be nice to be able to have an expanded {{dent/c}} that gives a more accurate behaviour (My current attempt doesn't actually work properly yet). ShakespeareFan00 (talk) 14:40, 22 June 2019 (UTC)

Alternatively, {{Dent/s}} that could be amended to have a suppression parameter for when it's used in Page: namespace headers? ShakespeareFan00 (talk) 14:40, 22 June 2019 (UTC)
The templates make it kinda hard to see, but all of them are really just spitting out HTML div elements with some associated CSS. Most of the /e templates spit out a plain </div>. Which means what you're really doing here isn't some special Mediawiki syntax: it's just various shortcuts for inserting HTML code. And as such, anything you want treated specially needs to be marked up specially.
I can't immediately think of any way to make {{dent/s}} automatically correct for being in the Page: namespace and set an indentation that is correct relative to whatever you had it set to on the previous page (and the pages may even be stored out of order). But simply wrapping the special bit in a separate {{dent/s}}, ending it when the special bit is done, and then adding the regular dent for the rest of the page is relatively cheap. --Xover (talk) 15:17, 22 June 2019 (UTC)
Good thinking in regard to the modified wrapping, but it would have caused problems for the nesting of HTML tags. For now I've just taken out the indentation entirely, and until {{dent}} or another template can be repaired to work correctly over multiple paragraphs and inside other "classed" DIV's without resetting stuff. ShakespeareFan00 (talk) 16:14, 23 June 2019 (UTC)
For whatever reason putting {{dent/s}}{{dent/e}} inside the {{p-collapse/s}}{{p-collapse/e}} causes the P handling to revert back to the default behaviour. Well at least we tried to to get it working. ShakespeareFan00 (talk) 16:19, 23 June 2019 (UTC)
The style is defined as DIV -> P but where the dent is put in it would be DIV-> DIV -> P, I'm not sure there's a way in CSS to say DIV -> (any number of sub levels) -> P Sigh :( ShakespeareFan00 (talk) 16:30, 23 June 2019 (UTC)
Problem exists between user and keyboard, Seems I'd got the CSS rule wrong :) ShakespeareFan00 (talk) 17:44, 23 June 2019 (UTC)

Bulk replacement request[edit]

Index:The Life of the Spider.djvu Currently this uses {{RunningHeader-centered}} in the header and footer of a number of pages.. The code given could be easily replaced with a much simple {{center}} given that the single element present is in the center location, but to do it for all pages manually would be time consuming. Is there someone here that can make the appropriate change with a script, AWB or bot rapidly? ShakespeareFan00 (talk) 09:08, 22 June 2019 (UTC)

Done manually, but it would be nice to have a response sometimes. ShakespeareFan00 (talk) 11:22, 22 June 2019 (UTC)
Expecting a response for a highly technical request in only 2.5 hours is extremely optimistic. —Beleg Tâl (talk) 16:47, 22 June 2019 (UTC)
@ShakespeareFan00: Are you familiar with the AWB software? If you're using Windows, it's a great option for stuff like this. I think it's pretty straightforward to figure out, but I'm happy to answer questions if needed. -Pete (talk) 19:25, 22 June 2019 (UTC)
I am aware of AWB, but feel I lack the communities trust to use it here at Wikisource. ShakespeareFan00 (talk) 19:29, 22 June 2019 (UTC)

Something is corrupt on this page The War with Mexico/Volume 2/Appendix[edit]

Can someone please take a look and advise. Thanks.Ineuw (talk) 06:27, 23 June 2019 (UTC)

Not to panic. When page 562 is formatted correctly no doubt all will come good! 08:00, 23 June 2019 (UTC)
Rather than using a table, had you considered using {{p-collapse/s}}{{p-collapse/e}}? ShakespeareFan00 (talk) 15:37, 23 June 2019 (UTC)
I considered all formats but settled on tables, being the most versatile.Ineuw (talk) 20:30, 23 June 2019 (UTC)
PS: @ Thanks, the table ended on Page:The War with Mexico, Vol 2.djvu/542 and the main namespace page is fixed.Ineuw (talk) 22:02, 23 June 2019 (UTC)
@Ineuw: You are quite correct. I did not word it well but meant only after the entire set of transcluded pages is (at least) at proofread standard — if the main-space problem remains then is the time to start seriously bug-hunting. I missed the fact Appendix A (and thus the table, too) ended on page 524! 04:44, 24 June 2019 (UTC)
Understood what you said about the set. Instead, there are a lot of small tables within the section.Ineuw (talk) 08:13, 24 June 2019 (UTC)

Page:Atlas of the Munsell color system.djvu/9[edit]

Can someone come up with a way to convert the Munsell colors back to RGB for the table cell backgrounds? I tried using a convertor/picker here [2] but the colors picked bear little resemblance to those in the (faded) chart.

I am out of my depth and so an expert to write a conversion script would be VERY much appreciated. ShakespeareFan00 (talk) 14:42, 23 June 2019 (UTC)

I couldn't find anything better than the link you provided. Maybe w:Wikipedia:WikiProject Color will be able to point you in the right direction? —Beleg Tâl (talk) 13:05, 24 June 2019 (UTC)
You could also just open the scan image in Paint or GIMP or something, and use the colour picker tool to identify an approximate colour to use. —Beleg Tâl (talk) 13:06, 24 June 2019 (UTC)
Due to a 'misunderstanding' I'm not editing on Wikipedia right now, otherwise I would have asked an appropriate project there already. Nothing to stop another contributor from doing so though. ShakespeareFan00 (talk) 13:42, 24 June 2019 (UTC)


A number of the subpages for this translation are giving rise to "fostered content" , which typically means a table has been misformatted somewhow. Can an experienced contributor take a look and provide a generalised fix that could be applied across all the pages?

(side note, The translation is marked as incomplete). ShakespeareFan00 (talk) 09:08, 24 June 2019 (UTC)


The second example uses a score.

However, the SCORE extension generates a DIV block wrapping the formatted score output. Per HTML structuring conventions, you can't put a DIV inside a SPAN.

Can someone advis on how this might be repaired, or do I need to raise a Phabircator ticket like I did for a related issue with the POEM tag? ShakespeareFan00 (talk) 13:41, 24 June 2019 (UTC)

Use a different template that does not require the contents to be inline elements? —Beleg Tâl (talk) 02:09, 25 June 2019 (UTC)
Whilst the conventional rule is a normal SPAN may not wrap a normal DIV that is not quite what is happening here. Strictly a SPAN modified via display attribute into an inline-block is attempting to wrap a block DIV element which appears to be permissible in some browser renderings. I pose the question (as I do not know myself) if this is a legal (if rather borderline) construct. Any expert opinions? 10:25, 25 June 2019 (UTC)

British Calendar Act of 1751[edit]

Can some PLEASE explain why the formatting on this is COMPLETELY *&&^ED up? ShakespeareFan00 (talk) 22:01, 24 June 2019 (UTC)

Can you please explain what part and in what manner is the formatting completely ***ed up? —Beleg Tâl (talk) 02:08, 25 June 2019 (UTC)
Prior to my subsequent updates, the 'sidetitles' overlapped the main body text, in the portion that represented the original Act. This was determined to be due to it using a version of {{cl-act-t}} which was subsequently edited considerably. Replacing it with {{cl-act-h}} and cleaning up the excessive DIV based formatting in that section solved the issue. The rest of the text could be 'standardised' though.

ShakespeareFan00 (talk) 08:34, 25 June 2019 (UTC)

Page numbers do not display correctly..[edit]

See :-- The Prince (translated by William K. Marriott)/How Many Kinds of Principalities there are, and by what Means they are acquired In Firefox, the page numbers are wrong or mis-aligned with respect to other content on the page. ShakespeareFan00 (talk) 12:39, 26 June 2019 (UTC)

It should be OK now, I excluded the empty pages. --Jan Kameníček (talk) 12:57, 26 June 2019 (UTC)
Thanks, it would be nice if the pages tag had a a noblanks option so the page concerned didn't need to be listed individualy in this use case.ShakespeareFan00 (talk) 13:23, 26 June 2019 (UTC)
Hmm. I though it automatically excluded pages marked as "Without text"? --Xover (talk) 14:02, 26 June 2019 (UTC)
No it doesn't. They take up zero vertical space when they are empty, but they are still transcluded. Comes in handy sometimes, for example when transcluding pages that use {{iwpage}} to pull content from another wikisource, but which have no local content that needs proofreading —Beleg Tâl (talk) 15:35, 26 June 2019 (UTC)

[SOLVED] A database error bug in Save a Gadget setting - already reported[edit]

Unchecked the Easy LST option which generated a database query error on Save. Could someone verify it? Ineuw (talk) 19:33, 26 June 2019 (UTC)

I just checked and can't reproduce. If there is a deployment underway currently then it was probably something transient related to that. --Xover (talk) 20:53, 26 June 2019 (UTC)
Thanks for checking. It was repaired/resolved within a few minutes by the powers that be. Ineuw (talk) 23:01, 26 June 2019 (UTC)

Deleting a duplicate book[edit]

Uploaded this file only to find an identical file already validated in pdf format. The difference is that my upload (.djvu) came from Internet Archive, and the .pdf came from Google books. Downloaded the .pdf and replaced the Google claim with a blank page.

I would like to delete the .djvu version. Unless I am mistaken, deleting the index page does not delete the pages. How can I go about removing the pages as well? Ineuw (talk) 06:40, 30 June 2019 (UTC)

@Ineuw: So far as I know there is no special magic involved: both the index and the pages are just wiki pages that happen to be in the Index: and Page: namespaces, and are deleted just like every other wiki page. But to avoid the tedium you will probably want to enable the MassDelete tool at the bottom of the Gadgets section of your Preferences so that you can use Special:MassDelete to do them all in one batch (temporary bot flag is probably a good idea too). --Xover (talk) 06:59, 30 June 2019 (UTC)
Much appreciated. Will tackle it tomorrow.Ineuw (talk) 07:10, 30 June 2019 (UTC)

Migrating from Index:A Chinese and English vocabulary, in the Tie-chiu dialect.djvu to Index:A Chinese and English vocabulary, in the Tie-chiu dialect.pdf[edit]

Are there any bots that can be used to ease the process of migrating to Index:A Chinese and English vocabulary, in the Tie-chiu dialect.pdf?

(If not, could I personally use AWB?) Suzukaze-c (talk) 02:35, 1 July 2019 (UTC)

Would be preferable in my opinion to repair Index:A Chinese and English vocabulary, in the Tie-chiu dialect.djvu and continue to use it. —Beleg Tâl (talk) 11:09, 1 July 2019 (UTC)
As in converting the PDF to DJVU? I think I like the general quality of the PDF more as well. (although, MediaWiki seems to be returning images of lower-quality than what I can see in PDF.js or pdfimages.) Suzukaze-c (talk) 21:30, 1 July 2019 (UTC)
Eh, using the DJVU file format doesn't cause MediaWiki to ruin the image resolution (??????), so I took the time to assemble a new DJVU. 🤷 Suzukaze-c (talk) 04:51, 2 July 2019 (UTC)

Okay why doesn't something work?[edit]

I created {{table class}} recently:-

|Line 1<br />Line 2
|Line 3
Line 1
Line 2
Line 3

works, compared to :-

{|{{ts|bc}}{{table class|cell_vtp }}
|Line 1<br />Line 2
|Line 3
Line 1
Line 2
Line 3

where the vertical alignment option is moved to a table class to apply it universally doesn't. WHY? Is Mediawiki being pedantic about where it expects to see things, as with SO many other issues, making it much more complicated than it needs to be? (sigh)

ShakespeareFan00 (talk) 09:36, 3 July 2019 (UTC)

The reason for creating {{table class}} was so I did NOT need to use a multiple {{ts|vtp}} calls. But if it's not going to work properly... as evidenced here - Page:Miscellaneous Babylonian Inscriptions.djvu/79 then why should I bother trying to make things less costly in terms of parser calls etc....

ShakespeareFan00 (talk) 09:40, 3 July 2019 (UTC)

It seems that for the moment I need to import the relevant style sheet outside the table {{table class/import}} ... hmm ShakespeareFan00 (talk) 10:42, 3 July 2019 (UTC)
The answer as to why is simple. Look at the table of parameters in the documentation for {{ts}}. "vtp" has a "no" for table, but a "yes" for row and cell. The solution is therefore logical, and I leave it to the reader to derive. Beeswaxcandle (talk) 08:17, 4 July 2019 (UTC)
So my styles were incorrect. :) Thanks. and I'd like someone to review the new templates (see Scriptorum) anyway. ShakespeareFan00 (talk) 08:36, 4 July 2019 (UTC)

The Christian's Secret of a Happy Life[edit]

This says there are misnested SMALL tags, but despite going over the alleged mistaken sequences MANY times, I can't find the mistake, Can someone please indicate precisely WHERE and WHY the relevant passages are misnested? because right next to them are nearly identical consturctionswhich DO NOT CAUSE the Linter tool to identify them as a mistake. I am getting extremely tired of running around playing hunt the minutiae, because certain tools don't provide enough SPECIFIC infrmation to actually repair a page.ShakespeareFan00 (talk) 13:44, 5 July 2019 (UTC)

My advice: don't even bother. That work is a garbage copydump and contains way more problems than linter mistakes. Don't waste your time playing whack-a-mole with a text that should be replaced wholesale instead. —Beleg Tâl (talk) 16:26, 5 July 2019 (UTC)
I did manage to repair it after a LOT of search and replace. However, as you say a properly scan-backed version would be much much better. ShakespeareFan00 (talk) 17:15, 5 July 2019 (UTC)
I have a theory about what might have gone wrong in respect of the SMALL misnesting though, see the "Pathological HTML test case" example, in my Userspace sandbox, Not that it's something LintError would directly tell you to look for :( ShakespeareFan00 (talk) 17:16, 5 July 2019 (UTC)

Page:Mrs Beeton's Book of Household Management.djvu/633[edit]

This for SOME reason stopped working correctly, when an update was made to the underlying module, I reverted the changes to the module, but I suspect they were made in good faith. What broke it and why?

ShakespeareFan00 (talk) 17:02, 5 July 2019 (UTC)

What module? What changed? What was broken? —Beleg Tâl (talk) 22:36, 5 July 2019 (UTC)
Module:Aligned table, removed the code which suppressed the generation of a header and footer portion of the generated table, which I'd used on the page concerned in an attempt to simplfy the table generation. ShakespeareFan00 (talk) 08:31, 6 July 2019 (UTC)

Page numbering, Dynamic layouts... etc...[edit]

Is there a reason why the pages tag doesn't have a layout=foo.css option?

I was looking at the markup here Equitation/Chapter 28 which does a custom layout, which to me (given that it's being applied to the whole work would be better suited as being a 'layout') or a stylesheet.

It's also confusing what needs to be defined in a layout specfically, some more documentation on this would be nice. ShakespeareFan00 (talk) 21:17, 5 July 2019 (UTC)

You'd probably have to ask the devs for why the proofreadpage extension is the way it is. There is also a CSS field in the Index page form which as far as I can tell does nothing at all. Using a template to apply styles to a work as a whole is a clever workaround. I don't think layouts are the right tool for this type of effort. —Beleg Tâl (talk) 22:43, 5 July 2019 (UTC)
The CSS field is supposed to apply that style in the Page namespace, iirc. But, no, it seems, essentially useless in practice. The idea in task T226275 though, might make the style effectively configurable from the Index: page. Or it might make sense to add the |layout= parameter you suggest here.
The page numbers / dynamic layout stuff is a great idea, but it's not really "finished". The layouts are hardcoded inside a javascript that is unconditionally loaded from common.js. I'm trying (off and on; I keep hitting dead ends) to figure out some way to first turn it into a Gadget—so that it can be temporarily disabled by anyone working on a new version of it—and then to clean it up and refactor it. One of the first tasks in that would be to move the actual layouts out of the javascript and into separate .css files that are actually editable by anyone that knows CSS. At that point we might start running into conflicts between the dynamic layouts and a hypothetical custom style provided by ProofreadPage, but that should be manageable if define what gets used for which thing well. It's entirely possible such changes to ProofreadPage will make the local dynamic layouts mostly obsolete. --Xover (talk) 08:48, 6 July 2019 (UTC)

Philosophical Review/Volume 1/September 1892/Der menschliche Weltbegriff[edit]

This apparently has a badly formed header. Can someone with the time and expertise please figure out why the header template is seemingly unable to cope with SIMPLE italics in a title, causing it throw lint errors?

I think the header template needs a complete rebuild anyway, given that a VERY large number of LintErrors I have had to fix recently are ultimately caused by very specfic minutiae of how whitespace and pairs of markup tags are dealt with. As I've said REPEATEDLY it is unreasonable to expect contributors to know these precise miniutiae every single time they might want to use a template. Is the header template sufficiently complicated that it can't actullay be made to work consistently from a contributors perspective, or converted to a Module that is better able to cope with the more unusual things contributors supply as parameters to it?

ShakespeareFan00 (talk) 20:48, 6 July 2019 (UTC)

I don't see any problems in the header, though the title and section field are swapped. I don't expect contributors to know or care about lint errors either (I don't care about them myself). If the devs want us to care about lint errors, they can make tools to prevent lint errors from being introduced in the first place. — I do, however, think it would be worth revisiting the header template, aligning it with modern responsive web design principles, integrating it more heavily with Wikidata, and so forth. I do acknowledge that it would be extremely hard to preserve backwards compatibility if this were done, and backwards compatibility is extremely important given the number of "hacks" that are in place currently to get the desired display in the current structure. —Beleg Tâl (talk) 23:32, 6 July 2019 (UTC)
I care instead and I think it should be enforced and become an habit. I spent quite a lot of effort in cleaning up the backlog. "lintHint" is very quick and quite helpful (most of the times ... and except that in Page ns, page or page-preview behave differently). Immediate clean-up is N times more efficient than cleaning up afterwards. And most of the times issues are wrong nesting of tags, due to misuse of div-based vs. span-based templates. — Mpaa (talk) 20:31, 8 July 2019 (UTC)
I would only support enforcing it if the Wikisource interface were to automatically by default check every edit for lint errors and clearly display exactly what needs to be fixed, in terms any noob editor can understand. Otherwise it is completely unreasonable to expect anyone to do this. Wikisource is a steep enough learning curve already. —Beleg Tâl (talk) 20:52, 8 July 2019 (UTC)
Without making too complex, I think it would be enough to make the gadget mandatory and add a sentence at the bottom of the page in Preview mode to ask (or suggest, as there might be some hard cases to figure out, i.e. it is not mandatory but encouraged) to run it. Errors are quite well explained at mw:Help:Extension:Linter, which could also be added near the message. IMO, it is reasonable to ask at least experienced editors to chip in to maintain the work as clean as possible. Not being visible does not mean we should ignore it.— Mpaa (talk) 22:27, 8 July 2019 (UTC)

Page:Sintel movie 4K.webm/8[edit]

Perhaps someone with sharper eyes can figure out EXACTLY where something is misnested? (because neither LintErrrors or the LintHint script could identify precisely WHICH template call in a overly complex nested set of templates was ACTUALLY responsible). ShakespeareFan00 (talk) 18:21, 7 July 2019 (UTC)

This was an unclosed small-caps template. For future reference, an easy way to fault-find these issues, even on very large pages, is:
  • Copy-paste the content to a sandbox page (optional, but allows you to save "checkpoints")
  • Run the LintHint tool from the edited page, it should show the same errors
  • Remove self-contained (i.e. don't break templates) chunks of text one at a time, checking LintHint each time.
  • When an error disappears, the culprit was in the last block removed.
  • Ctrl-Z to remind yourself what the block was. If the block was very large and the error is still not obvious, subdivide the block and do it again. Saving the sandbox at this point can be helpful, as you use "Show changes" to see the diff.
Hope that helps in future. Inductiveloadtalk/contribs 10:17, 8 July 2019 (UTC)

Sidenotes don't work directly in mainspace?[edit]


This uses sidenotes, but the formatting to do them correctly and the appropriate layouts is never directly loaded because it forms part of the page numbering script used when transcluding pages with a <pages> tag. Can someone come up with a way to use them directly? or provided an alternative format for use here? {{pline}} might be what was intended. ShakespeareFan00 (talk) 09:11, 8 July 2019 (UTC)

{{Pline}} would probably suffice. This edition is a copy of a copy of a copy so I wouldn't put too much bother into it; if {{pline}} works and is easy to implement, go for it. —Beleg Tâl (talk) 14:52, 8 July 2019 (UTC)
I did a quick test, and it doesn't look right. (sigh). Rhetorical: When will the contributors with technical expertise on this project actually get around to ACTUALLY fixing the many problems I've REPEATEDLY mentioned in terms of getting core things like sidenotes, like badly nested templates &c. &c. ShakespeareFan00 (talk) 16:13, 8 July 2019 (UTC)
Maybe try {{np2}}? —Beleg Tâl (talk) 16:32, 8 July 2019 (UTC)
Possibly, but I am going to take a break from trying to solve this problem right now.ShakespeareFan00 (talk) 18:29, 8 July 2019 (UTC)


The image isn't actually centered despite the stylesheet saying to do the margins automatically. Suggestions before I start sending carefully worded emails with choice comments to certain contributors? ShakespeareFan00 (talk) 15:54, 8 July 2019 (UTC)

The {{img float}} template is for floating images, which are put into spans so they can be inserted without breaking paragraphs. This template has never been intended or worked really well for centred images for that reason (centring a span doesn't mean an awful lot, you have to set text-align:centre on the container, not the span). In fact, when I wrote it way back in 2010, I didn't give it a centre option. The documentation of the template suggests two alternatives, one of which is not to use a template at all, because the "raw" File: method already has a centre argument. Alternatively, specifically for centred images, {{FreedImg}} has "smarter" behaviour for resizing, etc, but there are a few quirks there too (like an automatic 94% font size). Inductiveloadtalk/contribs 16:18, 8 July 2019 (UTC)
Okay so why is the documentation saying something is possible when it CLEARLY isn't? ShakespeareFan00 (talk) 16:21, 8 July 2019 (UTC)
The first suggested approach in the documentation DOES NOT WORK and it would be helpful if at some point it was in fact removed.ShakespeareFan00 (talk) 16:27, 8 July 2019 (UTC)
Hmmm - I wonder if this is what broke it. I recently had to revert another of this users good faith efforts to "update" something.... ShakespeareFan00 (talk) 16:31, 8 July 2019 (UTC)
{{Img center}} is the start of an experimental image centering template, It still uses DIV's though, so can't yet be used to solve issues elsewhere, If you want to improve it feel free, currently it just passes through a File syntax option, but removes one of the borders. Needs much improvement. (sigh) ShakespeareFan00 (talk) 19:12, 8 July 2019 (UTC)
Why does it need a new template anyway? File: and a {{center}} will work for the most common case, and {{block center}} for the rest. And if that's still not enough, {{FreedImg}} should handle it. Inductiveloadtalk/contribs 21:39, 8 July 2019 (UTC)
{{FI}} is not necessarily inline, which means it can't be embedded inside a SPAN or P based template, or inside a P created by transclusion, Imgcenter is yet another attempt to work around this limitations (sigh) Sometimes I wish there were more people willing to fix these problems in the backend instead. ShakespeareFan00 (talk) 22:02, 8 July 2019 (UTC)
Anyway : Category:Images using center float, That's about 1100 images using a center imgfloat, which DOES NOT WORK, Perhaps you would be willing to get started on finding a CONSISTENTLY behaving LONG-TERM fix, so as with so many other issues, the same concerns do not have to be raised REPEATEDLY here, with no credible approach emerging. I am increasingly fed up, with being told to use various work-arounds that are potentially unstable or building up future problems. Unfortunately to really resolve issues and concern like 'inlined images' would need someone to hold down a developer until they actually respond to concerns like this instead of developing even more controversial bits of wizard gadgetry for a prestige project like Wikipedia. rant rant &c.

ShakespeareFan00 (talk) 22:22, 8 July 2019 (UTC)

On a related note - {{FIS}} doesn't work properly either as evidenced here -A_history_of_the_gunpowder_plot/Chapter_1 when I tried it. Unless someone comes up with a credible fix by the end of the week, I will consider just disabling or removing both templates, because they just DO NOT WORK correctly anymore.ShakespeareFan00 (talk) 22:35, 8 July 2019 (UTC)
Here {{FIS}} screws up the page numbering - A_history_of_the_gunpowder_plot/Chapter_24 at the point the image appears. ShakespeareFan00 (talk) 23:11, 8 July 2019 (UTC)
There is no calign parameter on {{FIS}}. Why are you not following the documentation and using {{FI}} as recommended for centred images? 01:41, 9 July 2019 (UTC)
According to the documentation for {{FIS}}, for centred images, you must ensure the width + margin-left + margin-right sums to 100%. You did not do this, so the span's cleared width is only the width of the image + padding + margins. As it is a span, it is embedded in the overall container, of course it doesn't "claim" more block-wise space. The example on the page uses percentages: 30% + 35% + 35% = 100%. You used px, so it's impossible get 100% width with hardcoded values (as 100% could be any number of px). You could do it with CSS calc() function, if it is acceptable WRT browser support (Chrome 26, Firefox 16, IE 9, Safari 7). This works for me:
 | width=600px
 | margin-left=calc((100% - 600px) / 2);
 | margin-right=calc((100% - 600px) / 2);
I'd appreciate input as to the acceptability of this, is there a minimum browser version targeted? Perhaps some e-readers wouldn't appreciate it? Do we care? If so, it could be useful to allow the template to apply a default margin-left/-right of calc((100% - imgWidth) / 2), as this avoids the template caller having to do this.
I do not think issuing an ultimatum threatening to disable/remove a template used by something like 16000 pages is constructive. Moreover, these templates have (relatively complete) documentation, even if they clearly have some issues. If you have beef with the docs, might I also suggest that they do say that charity begins at home? Inductiveloadtalk/contribs 07:26, 9 July 2019 (UTC)
Agreed. And thanks for reminding me to add the documentation for those templates. ShakespeareFan00 (talk) 07:59, 9 July 2019 (UTC)
And thank you taking the time to respond with the above suggestion regarding calc. I would agree with the auto option for the margin handling as it does exactly what was intended. [[3]] looks EXACTLY as intended now. ShakespeareFan00 (talk) 08:11, 9 July 2019 (UTC)
Glad it's working.
Be careful with the word "auto": in CSS it has a different meaning, which doesn't really make sense for inline elements (even with display:inline-block). The current {{FIS}} does actually use auto for margin-left and -right. This doesn't work as it does with {{FI}} because it produces this:
<span id="cSpan1" style="display:inline-block; width:100%; margin-right:auto; margin-left:auto;>
This only works for width=100% because there is no space on the floated line for anything else. As soon as width is less than 100%, you'll find the cSpan1 span only has whatever width you give it and no margins. This allows other elements on the floated line, or the embedding container, to creep up next to the image.
WRT to making this automatic, we'd have to agree that CSS3 calc() is "OK" for general use. Unless this is already agreed ({{div col}} already uses CSS3 properties). Inductiveloadtalk/contribs 09:10, 9 July 2019 (UTC)
It seems in some instances Mediawiki or other Wikisourced templates may already be using CSS3 type features.. Examine the HTML generated for the poem in [[4]] for example? ShakespeareFan00 (talk) 14:07, 9 July 2019 (UTC)

Enough is enough: WHERE is the mistake? i.e where the's flipped formatting?[edit]

According to the LintErrors page there's a flipped over DIV SPAN combination SOMEWHERE in a transcluded page for this Nicholas_Nickleby/Chapter_24, however, I've run ALL the relevant pages through the Lint hint script INDIVIDUALLY and it did NOT find a concern on ANY of the transcluded pages. So can someone PLEASE indicate EXACTLY WHERE LintErrors is deciding to be pedantic before I send a very strongly worded e-mail to the maintainers of the Linter tool about it's increasingly frustratingly INACCURATE false-detections ? ShakespeareFan00 (talk) 23:38, 8 July 2019 (UTC)

Could it be this? I found it quickly using Special:ExpandTemplates + an HTML validator.
<div style="display:table; position:relative; margin:0 auto; width:auto;">
<span style="font-size: 83%;"><div class="poem">
<p>Sing, God of Love, and tell me in what dearth <br />
Thrice-gifted <span style="font-variant:small-caps">Snevellicci</span> came on earth, <br />
To thrill us with her smile, her tear, her eye, <br />
Sing, God of Love, and tell me quickly why.
Suzukaze-c (talk) 02:43, 9 July 2019 (UTC)
Thank you, I'm very suprised that when I checked that page, the checker script didn't find it initially. Is this something that a localised script could check for and repair automatically, as the repair looks straightforward. @Mpaa:. ShakespeareFan00 (talk) 07:58, 9 July 2019 (UTC)
I don't know about an automated repair tool, as wikitext is notoriously tricky to parse, but finding lint failures in some unknown subpage is easy with Pywikibot: listpages -linter:html5-misnesting -prefixindex:Page:Archaeological_Journal -intersect
    1 Archaeological Journal, Volume 1.djvu/8
   10 Archaeological Journal, Volume 5.djvu/5
 10 page(s) found
This won't find lint errors that manifest only when the pages are transcluded together, but it would presumably have found this one. Hope it helps, Inductiveloadtalk/contribs 10:49, 9 July 2019 (UTC)
Thanks, I am not in the best position to ask, but would you also be willing to mention the pywikibot angle in the phab ticket at ? If this check could be an additional "check transcluded pages" button available from Special:LintErrors it would be a rather useful addition. ShakespeareFan00 (talk) 14:05, 9 July 2019 (UTC)

Is there a bot that checks for ligatures?[edit]

I have just proofread and transcluded Sin_and_Crime:_Their_Nature_and_Treatment. I found quite a few f ligatures in the scan, as listed in the Index talk:Sin_and_Crime.pdf, but they're hard to spot with my eyes and I'm thinking they would be quite easy to catch with a bot? Zoeannl (talk) 10:06, 9 July 2019 (UTC)

They are also very quickly spotted if you use certain fonts as the edit-box font. (I use Bedstead (which is based on a teletext font precisely because it's very easy to spot obviously different things like ligatures. ShakespeareFan00 (talk) 15:19, 9 July 2019 (UTC)
Done.— Mpaa (talk) 21:28, 9 July 2019 (UTC)


Hi people! Is ok to use these books by Santos-Dumont on their respective language Wikisources (English, French and German)? All the archives are from before 1920s and I remember to saw the French one on the fr.Wikisource sometime ago but I couldn't find now. Thanks, Erick Soares3 (talk) 15:23, 9 July 2019 (UTC)

@Erick Soares3: All works published anywhere in the world before 1924 are in the public domain in the US (95 years after publication). In Brazil, the term of protection is 70 years after the death of the author (pma. 70). Works by Santos-Dumont thus entered the public domain in Brazil in 2003. That means the scans can be hosted on Commons, and the English language works can be transcribed on English Wikisource. I am not familiar with the policies on frWS and deWS, but I would be very surprised if they were stricter than Commons. They may however have other restrictions (taking into account the jurisdiction of the transcriber, say) that affect the issue, so the only way to be sure is to ask on their respective Scriptoriums. --Xover (talk) 16:10, 9 July 2019 (UTC)
@Xover:, thanks! I will upload the English book and then research about the French and German Wikisources. Erick Soares3 (talk) 17:30, 9 July 2019 (UTC)
@Xover, @Erick Soares3: The French WS policy is PD in country of origin and the German WS is PD in EU (death+70), which are both fine for Santos-Dumont. —Nizolan (talk) 17:31, 10 July 2019 (UTC)
@Nizolan: Thanks! The German version was translated by Ludwig Holthof that I can't find any biographical information, but seeing that his works date from late 19th early 20th century I think that is easy to assume that are PD. Erick Soares3 (talk) 17:49, 10 July 2019 (UTC)
I just found that the translator died in 1911]. So, is PD. Erick Soares3 (talk) 17:52, 10 July 2019 (UTC)

Page:The Elements of Euclid for the Use of Schools and Colleges - 1872.djvu/31[edit]

I'm not entirely convinced calling lots of float rights is the best approach? Is there a way to combine the functionality here into a single template or indeed a stylesheet that could be universally applied per page? This works, so hardly a priority. ShakespeareFan00 (talk) 22:50, 9 July 2019 (UTC)

The postulate titles are floated right, so {{float right}} is exactly the best possible approach. You could use {{rbstagedir}} to apply the other formatting along with the floats if desired. However, if by "single template" you are thinking of creating some sort of {{eeusc-header-fl-r}} type monstrosity, please don't. It can do no better than the current solution, and it will be vastly more difficult to maintain. As for stylesheets, this should be accomplished by modifying {{float right}} rather than creating an alternative template whose only difference is that it uses classes instead of inline styles. —Beleg Tâl (talk) 12:19, 10 July 2019 (UTC)
(Aside) Float-right is SPAN based, which may be the reason a number of misnesting errors are being identified in other works. Block based content should not use a float-right in it's current form. Not sure how in some instances this would be fixed without having to do some pusedo-P hacks, to allow a DIV based float, but hiding the paragraph breaks, which I am very unwilling to do for maintenance reasons. unsigned comment by ShakespeareFan00 (talk) .
I didn't notice it is span based. We should create a separate template for block elements (maybe modify {{block right}}?). In the case of the Elements however, the floated content is inline so {{float right}} is fine. —Beleg Tâl (talk) 14:38, 10 July 2019 (UTC)
There's also the issue of precisely where Mediawiki decides to "tidy", or end paragraphs which seems to be a dark art to decipher at times. this is despite REPEATED requests for someone to actually sit down and FORMALLY document what it's SUPPOSED to do. ShakespeareFan00 (talk) 16:14, 10 July 2019 (UTC)

CharInsert no longer displays the User defined row[edit]

Every page opened for editing I must re-select the "User" defined row of characters. I tried working with the "Special characters" built into the Toolbar, but this lacks ligatures among other things. Is there anything can be done to fix the Charinsert.Ineuw (talk) 08:13, 10 July 2019 (UTC)

Not seeing this behaviour in CharInsert for me (Monobook skin, no toolbar (since they stole it), and FF 67). Beeswaxcandle (talk) 08:37, 10 July 2019 (UTC)
@Beeswaxcandle: Are you saying that your "User" Charinsert selection shows up automatically when you open a page for edit? Also, please clarify "no toolbar (since they stole it), and FF 67)". To emulate your setup I witched everything to Monobook (from Vectra).
  • Logged out from WS and deleted the en.wikisource cookie from FF 67.0.4.
  • Logged back, and in the Page namespace I edited a page.
  • Set Charinsert to "User", and saved the edited page.
  • Opened the following page in edit mode, and my Charinsert reverted to "Insert" row.Ineuw (talk) 20:25, 10 July 2019 (UTC)
In preferences, editing tab, I have "Enable the editing toolbar" off. I loathe this one and manage without it. The only thing I can't do is initiate OCR of a page. Beeswaxcandle (talk) 05:54, 11 July 2019 (UTC)

POEM inside REF tags?[edit]

Can Poem not appear inside a REF tag? Because I can't find another reason why King_Solomon's_Mines/Chapter_XVI should be giving a DIV SPAN swap warning. Yet sometimes in other works this behaviour works, and sometimes it doesn't (Sigh!)

Do I actually have to do something drastic to get things PROPERLY fixed? ShakespeareFan00 (talk) 16:03, 10 July 2019 (UTC)

Examining one of the poem based references : here :
<ol class="references">
<li id="cite_note-1"><span class="mw-cite-backlink"><a href="#cite_ref-1" aria-label="Jump up" title="Jump up"></a></span> 
<span class="reference-text">
<div class="poem">
<p><span style="float:left; text-align:right; margin-left:-moz-calc(0em - 3em); margin-left:calc(0em - 3em); width:3em;">"</span>Now haste ye, my handmaidens, haste and see<br>
How he sits there and glowers with his head on his knee."

Which is very obviously malformed HTML. You clearly can't put a DIV in a SPAN. As this is the behaviour of Extension:Cite itself, this isn't going to be a problem that can be solved easily, given that for it to work, what Extension: Cite actually generates will need updating. Extension:Cite however is a fairly core template... (worried look). I'm also suprised no-one noticed this previously. ShakespeareFan00 (talk) 16:44, 10 July 2019 (UTC)

Looks like it has been noticed previously, (phab:T49544 "<references/> list item must not wrap the text in <span>"), but some kind of server error is preventing me from seeing the content of that ticket. —Beleg Tâl (talk) 16:56, 10 July 2019 (UTC)
I hope by "do something drastic" you mean "create lots of tickets at Phabricator" because that would very likely be the best solution for parser-level issues like this. —Beleg Tâl (talk) 16:59, 10 July 2019 (UTC)
I think I had raised several tickets of this nature prveviously. ShakespeareFan00 (talk) 17:12, 10 July 2019 (UTC)
If I haven't, maybe I should open a new one, linking here. ShakespeareFan00 (talk) 17:38, 10 July 2019 (UTC)
I am not using HTML but have no problem creating tables and poems within <ref></ref> tags with our templates.Ineuw (talk) 20:32, 10 July 2019 (UTC)
You're missing the point a little bit. The Cite extension will absolutely allow you to create tables and poems within ref tags, but every time you do this the Cite extension will generate malformed HTML regardless of what markup you use. —Beleg Tâl (talk) 21:01, 10 July 2019 (UTC)
I stand corrected Beleg Tâl. :-) Never used the Cite template.Ineuw (talk) 21:37, 10 July 2019 (UTC)
If you've used ref tags, you've used the Cite extension. The Cite extension is what provides us with the ref tag, the references tag, and probably a couple of other things. —Beleg Tâl (talk) 23:18, 10 July 2019 (UTC)

China Review scans from HKU[edit]

HKU has uploaded scans of the articles from the 25 volumes of The China Review (1872–1901), all of which should be PD. Unfortunately they only include the articles and not the frontmatter and other non-article pages of the journal volumes/numbers (notably there's no scan of the ToC). Are these OK to upload article by article? The alternatives aren't much: IA has a couple of issues of CR but not more than that; HathiTrust seem to have scans of the whole thing but their nutty copyright policy—they view anything after 1879 as copyrighted by default and allow only pdfs of individual pages to be accessed, and from the US only—makes it technically difficult to transfer them. —Nizolan (talk) 18:57, 10 July 2019 (UTC)

(Worth noting a contents page for The China Review already exists but the external links are all to the mostly unusable(?) HT scans. —Nizolan (talk) 21:21, 10 July 2019 (UTC))
HathiTrust only treats everything after 1879 as copyrighted because you're outside the US; as a general rule, everything before 1924 is free on HathiTrust. Their agreement with Google means you can't download the whole PDF of Google-scanned files (most of them), but they don't make any serious effort to stop people for downloading them page by page. I can do it for any work you need.--Prosfilaes (talk) 07:16, 11 July 2019 (UTC)
@Prosfilaes: I figured, my complaint was more that, afaik, there's nowhere in the world that has a "publication + 140 years" law so it's a pointlessly overrestrictive policy on their part. The articles I'm interested in at the moment are pre-1879 so I should be able to get the volumes from Google for now, but I might take you up on the offer in the future, thanks. —Nizolan (talk) 12:37, 11 July 2019 (UTC)
In 2010, the earliest known work under copyright in the UK was from 1859. In 2011, it was from 1865.[5] So if one wants to be absolutely correct about life+70 law, publication + 151 years is necessary. For a major author, Author:Bertrand Russell's first work will be 145 years old when his works leave copyright in life+70 countries. Restricting, but not pointlessly overrestrictive.--Prosfilaes (talk) 22:52, 11 July 2019 (UTC)

Sometimes Mediawiki makes me want to scream...[edit]

Page:The_Church,_by_John_Huss.pdf/353, A misnesting glitch is PURELY down to whitespace handling such as occurs at the end of a page between the text and the footer. And this means what should be a paired opening and closing SPAN {{hin/s}} (in the body) and {{hin/e}} (in the footer), are incorrectly tidied up, making what was supposed to be an intended template USELESS in one of it's intended use cases, and causing unecessary Linter warnings Or am I not seeing an obvious repair? ShakespeareFan00 (talk) 21:38, 10 July 2019 (UTC)

If the transcluded output renders correctly, I would not consider the template useless. Does MediaWiki always generate lint errors if a span is split this way? If so, this is a bug and the devs have to fix it, because open/close span templates are not going anywhere. —Beleg Tâl (talk) 23:31, 10 July 2019 (UTC)
I'm not sure it transcludes nicely, and I couldn't see any obvious errors in the direct HTML view in the browser. I think Linter MIGHT be misdetecting this very specfic use case... ShakespeareFan00 (talk) 00:30, 11 July 2019 (UTC)

Page edit problems[edit]

The Page header appears below the textbox on top of the footer see display.

Also, the OCR button (activated in Gadgets) is frozen, although the "Site: General utilities needed by the templates and portals of this wiki project." is active.Ineuw (talk) 21:32, 10 July 2019 (UTC)

I tested these in both Windows 10 and Linux Mint 19.1 in Firefox 68 and the latest stable Vivaldi.Ineuw (talk) 21:32, 10 July 2019 (UTC)

Works for me, Windows 10 v1903, Firefox 68 and latest stable Vivaldi. I assume you've tested it while logged out? —Beleg Tâl (talk) 23:38, 10 July 2019 (UTC)
I did and it didn't work, so after taking a long break, now everything works.Ineuw (talk) 07:00, 11 July 2019 (UTC)
@Beleg Tâl: My software is identical to yours but again the same, . . . the OCR stopped working and the header position is again at the bottom. When logged out, the page layout is fixed, but cannot test the OCR (but I am sure it's good). Cleared all cookies. Logged in and after two pages the problem returned. I am convinced that it is the software, because as I said earlier it is the same in Vivaldi as well. Any ideas? Ineuw (talk) 07:37, 11 July 2019 (UTC)
@Ineuw: Do you have any messages in the Javascript Console? Have you tried blanking your user scripts to check whether something in there is affecting this? There was a new Mediawiki release deployed yesterday. I don't see anything obviously related, but in the interaction between changes to Mediawiki and custom user CSS and JS there is a lot of potential for stuff to break. --Xover (talk) 11:11, 11 July 2019 (UTC)
@Xover:, That's the one thing I forgot, because I use it very rarely. Thanks for the reminder. As for my custom scripts, I will remove and re-install them one by one.Ineuw (talk) 20:01, 11 July 2019 (UTC)

@Xover: Tracked down the source of the error. It is the pathoschild family of scripts. Emptied vector.js and .css, Then restored the .css and everything was fine. Then, one by one, I added the scripts until the errors popped up. This was when I added the first script pointing to pathoschild tools on wmflabs in the vector.js

A maximum of 6 errors (below) showed up when opened a Page for editing. Otherwise, 3-4 various "Content Security Policy" errors everywhere always, with numerous warnings about outdated load.php issues. Should I file a bug report?Ineuw (talk) 02:54, 12 July 2019 (UTC)

Content Security Policy: The page’s settings observed the loading of a resource at (“script-src”). A CSP report is being sent.

Content Security Policy: The page’s settings observed the loading of a resource at (“script-src”). A CSP report is being sent.

Content Security Policy: The page’s settings observed the loading of a resource at (“script-src”). A CSP report is being sent.

Content Security Policy: The page’s settings observed the loading of a resource at (“script-src”). A CSP report is being sent.

ReferenceError: hookEvent is not defined index.php:390:1

Content Security Policy: The page’s settings observed the loading of a resource at blob: (“script-src”). A CSP report is being sent.
@Ineuw: Hmm. The CSP messages can be ignored: they're output for everything hosted on and it's just reporting.
The remaining message, though, is a genuine symptom of an error somewhere. .hookEvent() is a function for attaching multiple event handlers at once that was provided by the old "wikibits.js" library that has been slowly being phased out since before the introduction of ResourceLoader. .hookEvent() in particular, was deprecated in MediaWiki 1.17 (released in 2011) and removed in MediaWiki 1.29 (released in 2016). It seems unlikely that the new MediaWiki version deployed on Wednesday (1.34/wmf.13) would have made any change that affected that one way or another. I also copied your vector.js and failed to reproduce both the header positioning issue and the error message in the console, so the browser would seem to be one factor playing into it. I tested in Safari and Firefox, and neither had the header positioning issue; but Firefox did point to some potential issues in your personal copy of the nop-inserter script (don't we have a Gadget for that now?).
I don't think we have the source of the problem pinpointed enough to make a good bug report yet, so I'll keep poking around to see if I can find anything else useful. You might want to try tacking &debug=1 onto the end of the URL on a page where you see the problem: it tells MediaWiki to send you each javascript separately instead of bundled together through ResourceLoader, so error messages in the console will give you a better chance of tracking down exactly which line in which script is triggering the error. --Xover (talk) 07:18, 12 July 2019 (UTC)
Your help and advice is much appreciated. My copy of the nop-inserter.js highlights the page navigation arrow in green (for a few seconds) to indicate success. The general nop-inserter in Gadgets flashes a message. Both of these options exist in the public Gadget but the community preference was for the message and I was used to the colour highlight. PS. will post the results of &debug=1.Ineuw (talk) 07:33, 12 July 2019 (UTC)

Documentation markers for charcter(SPAN) vs paragraph(block) based templates..?[edit]

I'd like these to categorise the parent template they are on into an appropriate category, but not the Doc page they appear in?

Is this possible? Thanks. ShakespeareFan00 (talk) 09:37, 11 July 2019 (UTC)

@ShakespeareFan00: You might try putting your templates on the /doc page but inside <onlyinclude>{{div-based-template}}</onlyinclude>. The /doc page is transcluded onto the parent template page, but wrapped inside <noinclude>…</noinclude>. The net result ought to be that the template has no effect on the /doc page (onlyinclude); spits out its contents (i.e. the category) on the parent template page; but won't affect whatever other page where the parent template gets transcluded (noinclude). However, this will affect both the category and the visual output, so it might be better then to simply add the category to the doc page directly. --Xover (talk) 11:05, 11 July 2019 (UTC)

Rant part 4...[edit]

Worth filing a Phabricator ticket for better inline images support? ShakespeareFan00 (talk) 12:35, 11 July 2019 (UTC)

Phab ticket raised - Please comment.

ShakespeareFan00 (talk) 12:54, 11 July 2019 (UTC)

It seems to me that your edit to Page:History of Hudson County and of the Old Village of Bergen.djvu/45 is the correct way to handle such cases - "frameless" and "center" do not make sense for inline images. —Beleg Tâl (talk) 13:29, 11 July 2019 (UTC)
@ShakespeareFan00:I know that this is an old post but we were discussing the {{FIS}} template further on, perhaps you want to look at the change I made in the page mentioned above.

By the way, this is my starting layout generated by AutoHotkey. From here on, I delete the non-applicable parameters and modify the rest.

 | file         =
 | width        = 500px
 | cstyle       = margin-top:0px; margin-bottom:-5px
 | float        = left|right
 | margin-right = 7px
 | margin-left  = 7px
 | tstyle       = padding-top:.4em;  font-variant: small-caps; 
 | talign       = center
 | caption      =

Ineuw (talk) 22:45, 11 July 2019 (UTC)

OCR button not working[edit]

I am having an issue in which the OCR button on the toolbar isn't working. I encountered this when I was proofreading Index:A History of Wood-Engraving.djvu. For some reason, it worked fine for the first few pages with text, but after that, it simply refused to generate a text layer for the other pages despite the words being clearly readable. Is there a fix for this? DraconicDark (talk) 14:32, 11 July 2019 (UTC)

[CC: Ineuw] @DraconicDark: I just checked in Safari and the old OCR button is non-functional there too. Checking the javascript console it is throwing an error:
[Error] Error: Syntax error, unrecognized expression: An error occurred during ocr processing: /tmp/52004_20706/page_0011.tif
	error (load.php:20:880)
	select (load.php:37:233)
	find (load.php:41:184)
	init (load.php:142:753)
	jQuery (load.php:2:505)
	hocr_callback (Script Element 1:594:825)
	fire (load.php:45:980)
	fireWith (load.php:47:174)
	done (load.php:126:628)
	(anonymous function) (load.php:130)
This indicates a server-side problem: the phetools hOCR tool is returning an error message and the local javascript here that is calling it is choking on the error.
The Google Vision-based OCR tool that you can enable in your preferences ("Development (in beta)" section) seems too work though, and is a decent workaround. --Xover (talk) 15:35, 11 July 2019 (UTC)

Rant part 5 (Scoring points...)[edit]

Page:How to Write Music.djvu/21 Here a {{float box}} is used to wrap the scoring elements, which would work if the Score tag provided span-based output. It currently provides a DIV based output (not unexpectedly) which means that in this page malformed HTML get's generated. I'm starting to wonder if many of the common floats and shoulders Wikisource contributors use, should really instead be local classes in the core Mediawiki styling for the site, like certain of the table classes I've authored recently. ShakespeareFan00 (talk) 18:51, 11 July 2019 (UTC)

It doesn't wrap the score, it's wrapping a {{c}} (div-based). That's the div-span-flip error. Using {{float box}}'s own align parameter allows to avoid this. Inductiveloadtalk/contribs 19:37, 11 July 2019 (UTC)
Thanks. Perhaps someone should start documenting Wikisource:Formatting howlers?
Related is the example in the {{FIS}} documentation here -Template:FreedImg/span/doc, That's also showing up as 'flipped' warning, and there the SCORE tag IS being placed directly into a SPAN ShakespeareFan00 (talk) 19:56, 11 July 2019 (UTC)
What I actually see (in Firefox for that example is :-
<span id="cSpan2" class="freedImg" style="display:block; width:100%;"><div class="mw-ext-score" data-midi="//" data-source="//"><img src="//" alt="{ \clef bass \time 3/4 g,4 (b,4 [d4 f4 a4]) r4 \bar &quot;|.&quot; }" width="205" height="48"></div></span>
which is much more clearly malformed...
The actual invocation is :
 | type  = user
 | width = 50%
 | file  = <score>{ \clef bass \time 3/4 g,4 (b,4 [d4 f4 a4]) r4 \bar "|." }</score>

There the Score extension is CLEARLY generating a DIV based wrapper isn't it? ShakespeareFan00 (talk) 19:56, 11 July 2019 (UTC)

@ShakespeareFan00:, Try {{FI}} instead of {{FIS}}? Remove the "S" from FIS and try it again. FI was written first, but it could not float, So, FIS was created with <span></span> instead of <div></div>.Ineuw (talk) 20:11, 11 July 2019 (UTC)
I've checked, in the equivalent documentation for {{FI}} which I suspect is the same documentation as {{FIS}} although written previously, the score example works without generating the warning. Next question would be how to put a simple score (typically a short extract) inline (which might need yet another Phabricator ticket...) There's an open ticket on Phabricator about converting lillypond to MusicXML but any kind of support for doing that is still in the very early stages of development.. ShakespeareFan00 (talk) 20:18, 11 July 2019 (UTC)
For that you'll need to identify a text which has a short extract inline, such that this feature would be necessary for proofreading that text correctly. —Beleg Tâl (talk) 20:48, 11 July 2019 (UTC)
Something like A Dictionary of Music and Musicians/French Sixth or A Dictionary of Music and Musicians/Dot which are inline in the mainspace, but aren't in the Page: namespace? As it works in the mainspace, I don't worry about what happens in the Page: namespace. Beeswaxcandle (talk) 05:16, 12 July 2019 (UTC)

Rant Part 6 (The clever men at oxford?)[edit]

{{Cite DNB|vb=yes|author={{DNB PWJ}}|title=Beranger, Gabriel (DNB00)|work=[[Dictionary of National Biography]]|volume=04|pages=322|url=,_Gabriel_(DNB00)}}


[[File:PD-icon.svg|12px|alt=|link=]]&nbsp;This article incorporates text from a publication now in the [[w:public domain|public domain]]:&nbsp;<span class="citation encyclopaedia"><div></div><p id="DNBfooterInitials" style="clear:both; text-align:right;">[[Author:Patrick Weston Joyce|P. W. J.]]</p> (1885). [,_Gabriel_(DNB00) "Beranger, Gabriel (DNB00)"].  In [[Author:Leslie Stephen|Stephen, Leslie]]. ''[[Dictionary of National Biography]]'' <b>04</b>. London: Smith, Elder & Co. p.&nbsp;322.</span><span title="ctx_ver=Z39.88-2004&" class="Z3988"><span style="display:none;">&nbsp;</span></span>

There's a DIV in a SPAN, but as this template is used on a LOT of pages, I didn't feel happy trying to attempt a fix. ShakespeareFan00 (talk) 00:20, 12 July 2019 (UTC)

Create a template named Template:SF00, copy and correct the original template's error. Then, use it as a test template on pages which are linked to the original. If the test template works well everywhere, then open a discussion in the Proposals section of the Scriptorium. Ineuw (talk) 01:20, 12 July 2019 (UTC)
Sounds of manic laughter as Sf00 is led away in a virtual strait-jacket, [:] ONE misplaced line-feed. :lol :sigh :scowl ShakespeareFan00 (talk) 15:12, 13 July 2019 (UTC)
And it's that the DNB author initials {{DNB PWJ}} aren't a PURE span, they have code that generates the DIV. Is there an alternative template that should have been used?ShakespeareFan00 (talk) 15:22, 13 July 2019 (UTC)

Rant part 7[edit]


The float left item should be centered, I thought setting a style would do this, but text-align seems to be inapplicable to spans? ShakespeareFan00 (talk) 11:40, 12 July 2019 (UTC)

Also note to other contributors, typically style= when supplied to a template doesn't have wrapping quotes like it would when being supplied directly to an HTML tag..ShakespeareFan00 (talk) 11:55, 12 July 2019 (UTC)

Rant Part 8 ( A Table of concerns...)[edit]

I cannot seem to find a way to generate this row without a warning about fostered content or the resultant transclusion being incorrect.


As we other issues, I've REPEATEDLY said the way to PROPERLY solve this is to have a table syntax that supports proper continuations, but despite raising concerns there's been no further action taken on finding a STABLE long term solution. (sigh) ShakespeareFan00 (talk) 14:32, 13 July 2019 (UTC)

@ShakespeareFan00: Instead of complaining, ask for help, or take the initiative to read the history of the template. Template:Markup/row was originally imported from WP. So I replaced our version with the latest WP copy. At first glance, they look the same. Please try it because I don't know what it is for.Ineuw (talk) 01:09, 14 July 2019 (UTC)

Rant part 9 (Listing yet another concern!)[edit]

Despite there being no whitespace (other than line feeds, Mediawiki seems unable to handle multiline content that's supplied to a SPAN based template in a clean way... Collapsing the whitespace made the error go away, but the parser should be much more robust than that? I thought I'd ask here if there is another less drastic way of doing this as opposed to collapsing ALL whitespace because Linter (if not the parser) seems to throw a fit, if it doesn't see exactly what it expects... ShakespeareFan00 (talk) 15:55, 13 July 2019 (UTC)

Use a a div-template.— Mpaa (talk) 18:58, 13 July 2019 (UTC)

That's what I figured as well, but as I said the parser should be more robust. ShakespeareFan00 (talk) 19:43, 13 July 2019 (UTC)

Page tools link in sidebar is broken[edit]

In the Page edit mode the sidebar TemplateScript/proofreading.js link in the "gear" icon (on right of the 'Page tools' title) is broken. Can someone fix this link and check if the contents of that page are still relevant? Ineuw (talk) 20:36, 13 July 2019 (UTC)

Rant part 10 ( Something that really SHOULDN'T have worked.)[edit]

I'm head-scratching... Something I did here worked... And it shouldn't have, as I understood the layout model concerned :(


Perhaps another very experienced contributor can review what's going on here so I know why it worked when it did not previously? It would be nice to generalise this down a bit so it can be used elsewhere. I have a nasty feeling it's not stable though :( ShakespeareFan00 (talk) 23:02, 13 July 2019 (UTC)ShakespeareFan00 (talk) 23:12, 13 July 2019 (UTC)

Why do you think this should not have worked? This is exactly the solution I suggested to you the last two times you asked for help making it work, and it worked both those times. I also updated Template:Dropinitial#Technical notes with this suggestion last time you brought it up. —Beleg Tâl (talk) 11:01, 14 July 2019 (UTC)
Thanks, The tweak in this particular instance was to adjust the margins by the width, and to use the DIV based version of float-left, which I did not know existed.

That gets me thinking about sidenotes again.. ShakespeareFan00 (talk) 15:08, 14 July 2019 (UTC)

And I said it wasn't stable, but looking at User:ShakespeareFan00/sandbox ( fourth heading) , thanks to some template logic I wrote a while back, I've got the shift when it's transcluded. Next step would be the logic in a template like {{lrpaged}} to read the page number in mainspace and then I don't have to worry about left or right anymore. :)

...ShakespeareFan00 (talk) 15:30, 14 July 2019 (UTC)


This was leaking a SPAN, I've re done the logic in a sandbox and would appreciate someone else reviewing the code and the testcases before it's hopefully swapped in? ShakespeareFan00 (talk) 16:24, 14 July 2019 (UTC)

@ShakespeareFan00: I took a look and found even your improved version needlessly complicated. I took a stab at simplifying while guaranteeing valid output, and as best I can tell it works as intended now. If you agree then feel free to update the live template. --Xover (talk) 08:30, 15 July 2019 (UTC)
In the original the wraping lang in the span is only applied if parameter en is supplied. In your version it was supplied regardless, which is not the original logic as I understand it. Resolved this in the sandbox and swapped it in. ShakespeareFan00 (talk) 08:50, 15 July 2019 (UTC)

Quoted paragraphs...[edit]


Here each line of the quote is precceded by a ` to mark each line of the quote.

My thought was that this could be styled to insert this with a CSS rule, but no such selector for the start of line exists.

Also some discussions on the #css IRC channel at freenode indicated that it could not currently be done in purely CSS, however a contributor there suggested a Javascript method might work. Would such an approach be feasible, or for simplicity do we just drop this archaic formatting? ShakespeareFan00 (talk) 20:21, 14 July 2019 (UTC)

I'm going to say that this is not something we reasonably can reproduce, nor something we should reproduce. Granted it sometimes obliges us to find an alternate formatting to convey the same information, but that's not usually been a big problem in my experience. We have all the usual typographic conventions at our disposal, including things like indentation and margins. The important factors are that it is clear and understandable for our readers, and that it reproduces the meaning of the original formatting (not its mere visual instantiation). --Xover (talk) 07:14, 15 July 2019 (UTC)
The "traditional" solution to this situation is to enclose the quoted section within <blockquote></blockquote> tags. Granted, the representation looks nothing like the running-left-quotes and maybe never will but at least the semantic structure is captured and for now that has always been regarded as "good enough." 07:38, 15 July 2019 (UTC)
Not only is not something we can "reasonably" reproduce, it's not something we should reproduce. Typesetters of printed works that use the device know what the line length will be on every page of their book and can insert the appropriate slug as required. Given we are reproducing works in a non-printed medium for multiple devices that use many different line lengths, the benefit of doing the work to make it behave is minimal (particularly as the situation where a quoted section goes across page breaks will be particularly messy to deal with in Javascript). The nearest equivalent in modern typography is to block indent the text and place a grey bar on the left margin. The suggestion of the IP to use blockquote tags is pretty close and there is a family of templates ({{quote}}) that allow for such that go over pages. I would expect each paragraph within the blockquote to begin with a quote mark and the final paragraph to terminate with one (as per long-established practice) even though they are set off. Beeswaxcandle (talk) 09:32, 15 July 2019 (UTC)
Which is what I eventually did . ShakespeareFan00 (talk) 10:13, 15 July 2019 (UTC)

Identifying Header and Footer in AWB[edit]

Hi all. I am just starting out using AWB on US Statutes at Large Vol. 33, Parts 1 and 2. I've successfully done some very simple search/replace of obsolete template calls, but now - no doubt preparing to trip for the first time over my new and exciting AWB shoelaces, I am being more ambitious! I would love to be able to insert the necessary elements for USStatHeader in the several hundred pages of these two documents. My problem is my ignorance of regex - I thought I could search for <noinclude>*</noinclude> to replace it with my USStatHeader string, but no dice. It then occurred to me that I need a text-layer means of differentiating between header and footer, anyway, which my initial approach would have failed to achieve. Is what I am trying to do possible? If there are any AWB wizards out there, if it's possible to use a search/replace control file whereby for a given page file name, it pastes the next (perfectly formed) line of wikitext from the input file into the header, that would be AWESOME. Thanks, and apologies if this is all an achingly basic question. CharlesSpencer (talk) 11:52, 16 July 2019 (UTC)

An empty header generally takes the form <noinclude><pagequality level="0" user="Beleg Tâl" /></noinclude>. I generally find it helpful to use to build and test regexes that I will use in AWB. The CSVLoader plugin should be able to handle your search/replace control file requirement; see w:Wikipedia:CSVLoader/Find and replace for details. —Beleg Tâl (talk) 12:16, 16 July 2019 (UTC)
That sounds VERY interesting - thank you very much indeed. I shall report back! CharlesSpencer (talk) 12:51, 16 July 2019 (UTC)

The Poems of Henry Kendall[edit]

Would someone with access to HathiTrust be able to grab this scan for me? Thanks. —Beleg Tâl (talk) 19:04, 18 July 2019 (UTC)

@Beleg Tâl: Will this do? --Xover (talk) 22:07, 18 July 2019 (UTC)
@Xover: thank you! —Beleg Tâl (talk) 00:34, 19 July 2019 (UTC)

Inserting a literal asterisk (*)[edit]

I need to begin a paragraph with an asterisk. How does one keep it looking like an asterisk? Klarm768 (talk) 10:36, 19 July 2019 (UTC)

I found the answer... {{*}} embarrassed to admit I asked the question over a year ago. And had forgotten the answer. Klarm768 (talk) 10:57, 19 July 2019 (UTC)

Surprised by the structure of a URL and a blue link leading nowhere—is there something wrong?[edit]

Reviewing the External links at w:The Chaos, I opened to find it is a redirect to:

Out of editorial curiosity, I opened:


Both pages said: Wikisource does not have a text with this exact name

Also on Wikisource, clicking the link on the page of The Chaos/De Chaos resulted in the same message in a language from the Netherlands (nl), I trhink. So I opened and a search for "Ruize" returned no instances. Google Translate could not translate that term. I don't edit on Wikisource much at all, so I was surprised by the structure of the URL and by a blue link that leads nowhere. There may be absolutely no issue here, but in case there is... Hey! Here's what I found odd!

All I can do is post here and probably not come back due to real-life limitations. I edit en.WP (including technical documentation), en.WT, en.C and occasionally other languages and projects. I mention my WM CV so you can maybe garner from this what an experienced editor might find odd on WS, but I did not check for an existing WS for WP editors page, like the one on WT. There's my data dump! Thanks in advance, Geekdiva (talk) 15:12, 19 July 2019 (UTC)

@Geekdiva: this is why it behaves this way:
  • nl:Ruize-rijmen is a book written in Dutch, which is why the link in the header of Ruize-rijmen/De Chaos points to the Dutch Wikisource.
  • The book nl:Ruize-rijmen has not yet been added to Dutch Wikisource, which is why it says (in Dutch) that this work does not exist.
  • The poem Ruize-rijmen/De Chaos is written in English, which is why this particular page is on English Wikisource instead of Dutch Wikisource.
  • The book nl:Ruize-rijmen is not in English, so the rest of the book should not be added to English Wikisource. This is why it says "Wikisource does not have a text with this exact name" when you go to en:Ruize-rijmen.
  • Finally, and this is the real issue: wikilinks that point to other wikis (including Dutch Wikisource) are always blue, even if the page does not exist on the other wiki. Links to nl:Ruize-rijmen are essentially redlinks, but they appear blue because nl:Ruize-rijmen is on a different wikisource.
Let me know if you have any other questions. —Beleg Tâl (talk) 15:51, 19 July 2019 (UTC)


I'm asking for someone with experience to review the approach used on this page with {{sn-year}} before deploying it more widely.

I have concerns that the rendering may not be entirely stable. ShakespeareFan00 (talk) 22:58, 19 July 2019 (UTC)

And here I've found another problem, namely Template:Sn-year/sandbox that it seems to be next to impossible to actually emulate what the page numbering script is creating (in respect of dynamic layouts to fake the structure sufficiently to allow for testing of layouts in other than mainspace from a direct transclusion.
It would of course be nice if for once, someone actually sat down (independently of content work) and implemented a PROPER and FULLY DOCUMENTED explanation of what certain core Wikisource functionality ACTUALLY does, so that templates like {{sn-year}} and {{Fake layout}} can be implemented in a logical, consistent and reproducible way, and be used for thier intended purpose, but given the time available to a volunteer project like this, I don't see that happening with any priority. Instead there are various work-arounds, hacks and 'too clever by half' approaches, that aren't stable (including some I've written). This is NOT an approach sustainable in the long-term, in terms of maintainability.ShakespeareFan00 (talk) 23:26, 19 July 2019 (UTC)
Looks good to me —Beleg Tâl (talk) 12:47, 20 July 2019 (UTC)

OCR failed and the button is frozen[edit]

On several pages, including this, I clicked on OCR in edit mode, and nothing happened. The old text remained. So, I saved the page and the text is gone. Ineuw (talk) 04:16, 20 July 2019 (UTC)

I tried using my alternate account and it was the same. Unfortunately, an unregistered user has no access to the OCR button.Ineuw (talk) 04:23, 20 July 2019 (UTC)

Emptied the vector.js and .css files and repeated the OCR action . The result is the same as with the files installed. — Ineuw (talk) 06:09, 20 July 2019 (UTC)

I have the same problem. --Jan Kameníček (talk) 06:14, 20 July 2019 (UTC)
@Ineuw, @Jan.Kamenicek: See #OCR button not working above. This is a server-side issue with Phe's OCR tools: the javascript gadget that implements the OCR button here calls a service on the toolserver that in turn attemps to perform OCR on the page requested, and that service is choking on something in certain DjVu files and returning an error message to the javascript gadget here. Since the gadget here doesn't specifically handle that error it just gets logged in the web browser's "Javascript Console" and the script just terminates (which looks like "nothing happened" to the user).
The quickest workaround is to just enable the Google OCR gadget in your preferences and use that. If it produces results that are unusable (it usually does at least almost as well as the old OCR gadget in my experience) it's also possible we can regenerate the whole DjVu file with a text layer generated by Tesseract 4 (Phe's tool uses Tesseract 3, and Google's uses the Google Vision API). It depends entirely on the work, how many pages are affected, whether original scan images are available, and so forth, whether it's worth doing the not-overwhelming but not-quite-trivial amount of work involved. For this particular work it doesn't seem obvious that regenerating the DjVu would produce any better results than the existing OCR text layer, but if you would like to test the results I'd be happy to try it. --Xover (talk) 09:04, 20 July 2019 (UTC)
@Xover:, Should I file a bug report? — Ineuw (talk) 09:49, 20 July 2019 (UTC)
@Ineuw: Phe isn't active on-wiki or on Phabricator, so it'd probably have to be at Github. --Xover (talk) 10:13, 20 July 2019 (UTC)
I am not experiencing the problem with a .djvu, but with a .pdf file. It has got an OCR layer, which is unfortunately quite bad, and so are the results with the Google gadget :-( I hope the OCR button will be repaired soon... --Jan Kameníček (talk) 16:40, 20 July 2019 (UTC)
I just checked and it's the same error, so whatever the issue is it affects both DjVu and PDF files. --Xover (talk) 18:37, 20 July 2019 (UTC)


@Xover: Thanks again for your kind help. Unfortunately, Phe's Github activity and his activity on French WS, ceased two-three years ago. I don't know what is Phabricator's policy in such a case, so I will post the question in their community forum. Also noticed a bug report about the .pdf scanner. task T224355 — Ineuw (talk) 20:24, 20 July 2019 (UTC)

The OCR gadget is an inseparable part of my work with the proofread extension and I am really surprised that its maintenance depends on one particular person... That is not good. --Jan Kameníček (talk) 06:55, 22 July 2019 (UTC)
I have announced the problem at phabricator. Please add there any relevant information that I missed. Thanks. --Jan Kameníček (talk) 13:46, 24 July 2019 (UTC)
It seems to work. — Ineuw (talk) 04:53, 25 July 2019 (UTC)
I have just tried it at Page:Magdalen by J S Machar.pdf/106 and it does not :-( --Jan Kameníček (talk) 06:26, 25 July 2019 (UTC)
Just tested it again, and it now doesn't work for me as well. Wish I could track down the version of 04:35, 25 July 2019 (UTC). — Ineuw (talk) 18:37, 25 July 2019 (UTC)

Section Labels and known problems[edit]

I have started publishing pages, but not transcluding any of them yet. I see in Help:Transclusion that "Use of letters and numbers is usually a recommendation, and other extended characters may cause some issues when transcluding." Section Labels published so far contain the following characters: underscore, hyphen, colon, ampersand, and period (_ - : & .) I am hoping that someone has used these without problems. Or that someone can attest to having used other non-alpha, non-numeric characters with impunity?

I need to decide whether I should I start over and republish all the pages done so far. I have over 25000 Section Labels yet to be inserted. ¯\_(ツ)_/¯ Klarm768 (talk) 21:07, 20 July 2019 (UTC)

In my humble opinion, just use a character and a number, and without space. In this manner, enclosing it in quotes is superfluous. In my scheme of things I use the following syntax. For the ending section "E" = End and "289" equals djvu page number and "B" = Begin of next section. Do not use page numbers, they can be either missing or duplicated. Djvu numbers are always unique.
<section begin=E289 />
<section end=E289 />

and for the second 

<section begin=B289 />
<section end=E289 />

— Ineuw (talk) 00:41, 21 July 2019 (UTC)

I have not found that to be a useful process for biographical works as you have to track too many things, and too many possibilities for mistakes and omissions. I have found it significantly more helpful to use a section name that matches the article name (subpage) and for the article to match the identifying component. You will see that approach in many of our biographical works (though noting that in have been involved in most).

This also enables a quick and easy transclusion process, as I have two sidebar scripts that insert things like



which means that if you have a ToC to work from then you are just inserting page numbers (first example where it spread over multiple pages), or one page number, when the article is contained within a page. Noting that this method as it can have spaces means you need to wrap the section name within quotes in the Page: ns.— billinghurst sDrewth 00:55, 21 July 2019 (UTC)

I guess that my projects are easy because having two sections on the same page is very rare. — Ineuw (talk) 06:48, 21 July 2019 (UTC)
Thank you both for your thoughts... although I'm a bit uncertain whether I rightly untangle the comments. I have revised 689 section labels to contain two "words"... entirely alpha-numerics. I do not understand whether there is additional advantage to truncating to a single "word" but I want to understand before resuming.

The work (Imperial Dictionary of Universal Biography) is a three volume set comprising 3940 page images. The 3672 pages-to-publish I have assigned into 25,740 sections, using a spreadsheet. There are 256 "signatures." Within my personal dissection I assign these "signatories" to 240 "identities" or "Contributors." This work has no de facto ToC. My sequencing spreadsheet is as close as probably ever will exist. I presume it is not eligible to be published with the IDUB, but it might have a usefulness in transclusion. I do not grasp the reference to "sidebar scripts," but I like the idea of facilitating transclusion. My background is not in computers. My mental-model is a mail-merge involving a "namespace template letter" into which fields are assigned and populated from each row of the spreadsheet to create templates to paste for transclusion. I'm sure that there are sophisticated methods but I work with the tools I have. Klarm768 (talk) 21:21, 21 July 2019 (UTC)

@Klarm768: Which project(s) are you referring to? — Ineuw (talk) 19:10, 23 July 2019 (UTC)
@Ineuw: I'm not sure what the word "project" means in your question's context. If there is an "official" Wikisource project regarding "Imperial Dictionary of Universal Biography" I do not know of its existence. As far as I know, billinghurst has created a few pages at the beginning of Volume 1 such as "List of Contributors" and other preparatory pages. Otherwise, as far as I know, I am the only person who has ever transcluded any portion. That occurred due to quest to track down as-complete-as-possible-works of James Frederick Ferrier. I choose to pursue the entire work. My reasons are personally motivated... but available on request.

Whether nobody else would choose to make it a project is not my concern. As far as a Wikisource IDUB project... which, if there were one, would have to mean I am deemed unwelcome/uninvited. I have spent many hundreds of hours getting the entire text in fair-copy-MS-Word... getting every entry & article sequenced and titled... getting the contributors disambiguated to considerable degree (actually one of the most confusing aspects). I would prefer to have a Wikisource-pilot steering my trireme... so many Scyllas and Charybdises. (Perhaps the only one who attempted the role got in trouble because of my errors... choices made before I ever discovered that Ferrier contributed to IDUB.) Besides... helping me with a project of 5,845,911 word-count is not Wikisource priority. In a year or two I hope to have IDUB in fair-copy-Wikisource and transcluded. The Wikisourse Mysteries are reserved for the initiates. Klarm768 (talk) 21:17, 23 July 2019 (UTC)

For what it's worth, there is one page of IDUB with 38 Sections on a single page. There are hundreds of redirects & biographies with fewer syllables than the subject's name. Klarm768 (talk) 21:24, 23 July 2019 (UTC)
@Klarm768:If there is more than 1 section on a page then you must use user:billinghurst's method. As for "project" I meant the title, just to take a look. — Ineuw (talk) 07:57, 24 July 2019 (UTC)
@Ineuw: I'm not sure what the phrase "user:billinghurst's method" means in your context. I have already gone back and converted all Section Labels to a single 'word' of alpha-numerics. The numeric portion is based on the image# & and the alpha portion reflects/preserves the sort order of the Sections. Is that the "method?" I scrolled through the "user:billinghurst" site but found nothing I could otherwise apply myself. I have previously transcluded articles from pages with multiple sections on a page. Does that mean I used the method" at that time? Or is the "method" something new in the past 18 months?

BTW, I do appreciate your taking time to respond. Hopefully I will eventually have more comprehension and fewer confusions. Klarm768 (talk) 15:22, 24 July 2019 (UTC)

We use the word project to collectively to describe major transcription efforts, be it formal or informal, though usually something that sits within Wikisource:Projects. Ineuw has done a lot of work within PSM and I have done work within DNB and other works within Wikisource:WikiProject Biographical dictionaries. In PSM there is usually an end of one and start of another article on a page, hence a rhythmic simpler use of sections. Whereas a biographical work like Index:The Dictionary of Australasian Biography.djvu or Index:Men-at-the-Bar.djvu can have many multiple sections. — billinghurst sDrewth 15:39, 24 July 2019 (UTC)
@Billinghurst: Thank you for your input. As I'm sure you are aware, Imperial Dictionary of Universal Biography has no ToC of its own. Is it conceivable that my Section-sequenced spreadsheet could be adapted (eventually) to generate the "records" to populate transclusion data. The Indexed Images are PDF rather than DJVU. Despite repeated attempts, I could not upload the entire Volume 3 in one piece, so I settled for making it into Volume 3a & 3b. Have I compromised automating the transclusion process? Klarm768 (talk) 20:14, 24 July 2019 (UTC)

@Klarm768: None of that changes my commentary about the preferred means to label sections with the resulting (subpage) article name. With regard to

  • the name list that you have, if you dump it somewhere as a text list, one name per line, I will convert it into a linked list for you, that is an easy task. The change that we make is to make it overt that OUR table of contents is not part of the publication, it is our generation. Men-at-the-Bar, Alumni Oxonienses, etc. are all in that boat.
  • DjVu vs PDF, again no issue, beyond DjVu has some advantages for our work
  • reupload as a split file, again no issue, the transclusion will care not, and as I suggested about ignoring volumes for your presentation, it makes no difference.

billinghurst sDrewth 22:02, 24 July 2019 (UTC)

@Klarm768: Some additional comments.

  • If your material is in spreadsheet format it can easily be used to create a ToC. You can create a ToC in the main namespace, as it was done in PSM. I don't think that its against the rules of WS if one adds relevant material, which does not exist in the original. The rule is not to change the original itself when proofreading.
  • In reference to section tags, you should familiarize yourself with billinghurst's method of sectioning the text. I think that the second option is the simpler. If help is needed, just ask.
  • If the text is in MS Word format, there are people who use it for editing, spell checking, and then copying and pasting as plain text.
  • Finally, "Seeing is believing". Create a couple of pages and ask for input here until you get up to speed. — Ineuw (talk) 21:46, 24 July 2019 (UTC)
@Billinghurst: @Ineuw: If it would be prudent to perform a test batch on an IDUB subset I could make X-Y-Z available soon. When inductiveload helped with a ToC for Institutes of Metaphysics we used and I exported a delimited text file. I presume I still can use that when the time comes. The columns (fields) in my spread sheet are entitled thus:

Volume# | Image# | Page# | Entry-Sequence per page | ABC | Contributor# | Contributor's Text-Signature | * | Title of Entry | word count | Disambiguation/Clarification | SectionLabel (which is built from Image#&Sequence)

I can arrange or reorder to your specifications. Yes, I have "Word Counts" on 25,740 sections and assorted clarification errata which I hoped to insert into the "notes" area at transclusion. Contributor# is my creation, an assigned number signifying the sequence in which a contributor-signature is first observed in IDUB. Then if a new signature is discovered which is defensibly thought to be the same contributor I assign that signature to the existing Contributor#. I try to include the reasoning behind that assignment in the Disambiguation/Clarification field. The * field is for IDUB biography titles whose names are preceded by * signifying the subject was thought to be alive. This indicator is nowhere explained on any page of IDUB but is findable in the IDUB advertisements in other books.

I have a whole other spreadsheet upon which I have researched all the contributor Signatures and IDUB "Lists of Contributors" and wherein I pursue the disambiguation process. I don't know if that can be found useful, but it exists. Klarm768 (talk) 11:29, 25 July 2019 (UTC)
Just go ahead and do a test page or two according to your plan. A Wikisource test page is = 1,000 words in the Scriptorium. — Ineuw (talk) 18:32, 25 July 2019 (UTC)
@Ineuw: I shall presume that you mean right here. Klarm768 (talk) 14:09, 26 July 2019 (UTC)
@Klarm768: Moved sample text to User:Klarm768/Sandbox01. — Ineuw (talk) 15:26, 26 July 2019 (UTC)

Chaucer's Works (ed. Skeat) Vol. II/Boethius Book I[edit]

Page numbering to the left with {{page break}} or {{pagebreak2}} is NOT compatible with {{block center}}.It would be NICE if someone eventually sat down and rewrote certain templates so that they have reliably predictable behavior regardless of where they are ACTUALLY used.

Other than stripping them out, I'd like a suggestion on how to get them to format as the original template intended? ShakespeareFan00 (talk) 21:15, 20 July 2019 (UTC)

Then don't use it. It is a very old template and it works for its tasks where it was designed to be used, block center came along later.

P.S. How about you stop the commentary of complaint? Keep things simple and they generally work fine. — billinghurst sDrewth 01:07, 21 July 2019 (UTC)

I am generally not happy taking out templates others have placed, but as I trust your technical expertise, I'll mention your above comment if someone queries my removal of 'incompatible' templates. ShakespeareFan00 (talk) 09:05, 21 July 2019 (UTC)
Is there a policy here about mass removal of 'incompatible' or nominally depreceated templates? ShakespeareFan00 (talk) 09:16, 21 July 2019 (UTC)
You and others, have previously suggested that before making certain types of changes, that I SHOULD consult more. If I don't ask about concerns when I can't find an obvious solution, then what do you suggest? You seem to be saying you would prefer a more neutral 'technical' request, in comparison to the percived venting that's sometimes happened. ShakespeareFan00 (talk) 09:30, 21 July 2019 (UTC)
In respect of the actual concern with the template "incompatibility" which is about 170 pages to check (most of which do not need any changes so possibly more like about 40-50 pages that need updating), any volunteers? ( the temporary solution is to convert |left to |hide.). A longer term re-design of the templates mentioned would be appreciated,

ShakespeareFan00 (talk) 11:10, 21 July 2019 (UTC)

A technical question arises, how does the PageNumber.js script get to get the page numbering to the side of the page consistently (I am wondering if it's a very specfic form of classing or style? Because if there was a way to arbitrarily call that code directly from mainspace, Could it be utilised by setting the relevant class and idents in {{Page break}} to get the desirable behaviour (of consistently left hand aligned page numbers)? I've not as far as I know had issues with block centered content and page numbers (generated) from a transcluded work vs trying to do it directly in Mainspace with {{page break}} As you are more familiar with the scripting code concerned, would it be possible to implement a way of calling it arbitrarily? ShakespeareFan00 (talk) 12:21, 21 July 2019 (UTC)
In PageNumber.js, all page numbers are listed before any of the text content, and each page number is manually positioned by the algorithm based on the calculated location of the transcluded content within the browser window. —Beleg Tâl (talk) 17:21, 21 July 2019 (UTC)
And is there any reason why that approach could not be used directly in mainspace? using Page break or Page break 2 to insert markers read by an algorithm? ShakespeareFan00 (talk) 14:00, 24 July 2019 (UTC)

I don't think it makes sense to use {{page break}} to simulate PageNumber.js on non-scan-backed works, and I would support removing all instances where it is used in this way. —Beleg Tâl (talk) 17:21, 21 July 2019 (UTC)

Little help?[edit]

Hi! Someone could help with the "Contends" and "List of images" from the book My Airships? I'd like to release each chapter when I finish each proofreading. Thanks, Erick Soares3 (talk) 17:55, 21 July 2019 (UTC)

@Erick Soares3: I took a stab at the ToC: here and here. Note in particular that I changed from roman to arabic numerals in the chapter subpage names. Looking quickly at the "List of images" it looks like the same approach would work there. There are, of course, also other ways to go about it, so please do not feel obliged to keep my changes if you prefer a different approach. --Xover (talk) 21:23, 22 July 2019 (UTC)
@Xover: Thanks! But I saw that something went wrong between page 11 and 12 (see My Airships). Have any idea? I will see what to do in the "List of images" - I'm still learning all that part of Wiki code. Erick Soares3 (talk) 21:34, 22 July 2019 (UTC)
@Erick Soares3: My apologies. It was that darn magic mystical {{nop}} that needs to be inserted in oftentimes seemingly random places. When we combine multiple separate pages, each containing wiki markup, into a single page with transclusion, it is always tricky getting the whitespace right for the markup where whitespace matters. For example inside a table, the "|-" that starts a new table row needs to start on a new line, but the transclusion removes the preceding newline because most of the time that page break occurs "within
a sentence". The "magic nop" template forces a new line there, which makes it work. This is about the fiddliest and arcane bits of wiki syntax you will have to care about. --Xover (talk) 22:05, 22 July 2019 (UTC)
Can I remind users that inside tables, that we should be using {{nopt}} (nop for tables) as it is <span> type. — billinghurst sDrewth 22:19, 22 July 2019 (UTC)
Thanks!! Erick Soares3 (talk) 22:26, 22 July 2019 (UTC)
I have never heard of {{nopt}} - smart idea. Could you perhaps fill out Template:Nopt/doc and update Template:Nop/doc so future editors will know about it? —Beleg Tâl (talk) 01:50, 24 July 2019 (UTC)

Lilypond line width[edit]

In OAW vol 1 page 36 and OAW vol 8 page 433, lilypond is generating scores that are much wider than the rest of the layout: this magazine looks best if fixed-width elements are no wider than 600px, preferably 500px. Page 433 could be improved by breaking the score into three lines; is there a way to do that? The same could be done on page 36, or the lines could be horizontally compressed (there is no reason the notes have to be spaced out that much). Levana Taylor (talk) 23:44, 21 July 2019 (UTC)

@Levana Taylor: Poking around like a bull in a china shop, but I did take a stab. Maybe some further tweaking will get you the result you're after? As best I can tell there is no way to make the Score extension generate images of a given web-ish dimension, much less dynamically size them relative to the browser or web page. It is possible something can be achieved by wrapping the score tags in a template that imposes some sizing: it'd be hacky and inconvenient, but might actually work. If this is an issue you run into often then feel free to give me a ping and I'll look into it. --Xover (talk) 21:57, 22 July 2019 (UTC)
When I read the Lilypond manual and didn't see any way of forcing the lines to break, I thought I must have overlooked it, but if you didn't find it either, it must not be there. Your method of globally reducing the size of the output works, though. It forces the user to choose between long lines or else tiny font size for the lyrics. I think the best compromise is "line-width = 130\mm #(layout-set-staff-size 13" which fits inside 600px (although 120/12 fits into 500px, it's ridiculously tiny). Thanks! Levana Taylor (talk) 01:08, 23 July 2019 (UTC)

Adding missing pages to a file[edit]

I uploaded vol. 31 of Poet Lore from HathiTrust, but then I found out that some pages were missing. HathiTrust has one more copy, which has other problems (missing index pages), but it has got the pages that are missing in the uploaded file, i. e. pages no. 328–329 and 371–374. Could someone who has got the needed software extract these pages from the other copy and add them to the uploaded file?

Thanks very much! --Jan Kameníček (talk) 21:36, 22 July 2019 (UTC)

@Jan.Kamenicek: I don't have tools for working with PDF, but I can probably generate a new composite DjVu file for you. It'd also get a completely new text layer which I cannot guarantee is as good or better than the existing one. In my limited experience it's as good or better than the alternatives but this will depend strongly on the work in question as well as a bit on individual preference. But if those caveats are acceptable to you I can take a stab at it. --Xover (talk) 11:38, 23 July 2019 (UTC)
@Xover:I do not insist on PDF, I prefer it mostly because it is usually easier to handle for me. So if you make a DJVU composite, it will help. Thanks very much! --Jan Kameníček (talk) 11:50, 23 July 2019 (UTC)
@Jan.Kamenicek: Done: File:Poet Lore, volume 31, 1920.djvu. Please check that I didn't mess up anything with the page doctoring.
I would also be very interested in your impressions of the OCR quality. Both overall; if there are specific features that are handled better or worse than expected; and whether there are particular pages where it does unexpectedly well or falls down completely. The OCR here is performed using different software than IA and that used in both the OCR button and the Google OCR button, and OCR results are to a certain degree dependent on the preprocessing of the page images (which is obviously different here), so I'm interested in such feedback to get a sense of how well my tools are doing compared to the alternatives (and it may be that there is further pre- and post-processing I can do to improve the results).
No need to make a study of it, but if you happen to notice anything while working on it I'd appreciate a note on my talk page. --Xover (talk) 15:45, 24 July 2019 (UTC)
Perfect, thank you, you have helped me very much again! For more, I am going to write on your talk page. --Jan Kameníček (talk) 17:17, 24 July 2019 (UTC)

A Record of Education[edit]

There is a book, A Record of Education: The Schools and Teachers of Dedham, Massachusetts, that is now in the public domain and available on Google Books. It would be very helpful for my current project on Wikipedia if it was in Wikisource, but I am not as familiar with this wikiproject, have limited tech skills, and have no idea how to add it to Wikisource. Could someone please help? Thanks! --Slugger O'Toole (talk) 00:12, 23 July 2019 (UTC) PS - It is also in Internet Archive, if that helps. --Slugger O'Toole (talk) 00:18, 23 July 2019 (UTC)

@Slugger O'Toole: Step one is to upload it to Commons. I'll do that now for you. Step two is to make an index. Step three is to start transcribing. (Easier said than done.) Have you read Help:Adding texts yet? Do you have questions about that guide? —Justin (koavf)TCM 00:29, 23 July 2019 (UTC)
Thanks, Justin. I had not read that help page. I was looking at what I presume to be an older page and was a bit daunted. I have created the index page. I guess the next step is to proofread. Seems like it will take some time. I'd appreciate the help, if you're inclined to give it. If not, thanks for getting me this far! --Slugger O'Toole (talk) 02:55, 23 July 2019 (UTC)
@Slugger O'Toole: Happy to help but I'm too swamped to promise anything. :/ —Justin (koavf)TCM 03:36, 23 July 2019 (UTC)
@Slugger O'Toole: By the way, we don't transcribe stickers, hand-written notes, stamps, etc. that weren't a part of the original work. —Justin (koavf)TCM 03:42, 23 July 2019 (UTC)
@Slugger O'Toole: It seems that the original OCR layer is quite bad and unfortunately our OCR button has stopped working for some reason a short time ago. So I recommend you to enable Google OCR button in your Preferences/Gadgets. --Jan Kameníček (talk) 07:36, 23 July 2019 (UTC)

Quality differences between PDF and DjVu[edit]

Following on from this thread

Hmm. This is odd. The OCR text showing in the Page: namespace here is pretty poor; but when I look at the original PDF and DjVu files at IA the OCR text looks just fine. I can even copy from the PDF into a text editor and get much better quality than what appears online here. So I uploaded copies of both files locally here (PDF, DjVu) and checked the same page in both. This is the result:



Jno BuUard, Robt. Grossman, Hen Wilson, Jno N[ewton], Edw. Colver. Hen. Smith, Nath Colborne, Nath. Aldus, Hen Phillips, Sam^

Dan Morse, Jno. Morse, Jos. Kingsbury, Jno. Dwite, Lamb.
Kemp, Edw. Richards, Tho. Leader, Geo. Bearstowe,

G[enery], Edw.

Jonath. Fairbanks,

Jno Frarey,


Mich Powell, Mich Metcalfe


Jno Metcalfe,

Lusher, R[obt] Hinsdell, Pet. Woodward, Jno Guyle,

Rich. Evered, Robt. Gowinge, &ce.


said Inhabitants, taking into Consideration the great necesitie

some means for the education of the youth in o^ s'd
Towne, did with an unanimous consent declare by voate their willingness to promote that worke, promising to put too their hands to provide
maintenance for a Free Schoole in our said Towne. And farther did
resolve and consent, testifying it by voate, to rayse the summe of
Twenty pounds p annu, towards the maintaining of a Schoole M^ to
keep a free Schoole in our s'd Town.
of providing


and consent to betrust the s'd 2(d£ per annu
Towne, formerly set apart for publique use, into
the hand of Feofees to be presently chosen by themselves, to imploy the
sd 20;^, and the land afors'd, to be improved for the use of the said
Schoole that as the profits shall arise from the said land, every man
may be proportionably abated of his some of the s'd 20£ aforesaid,


also did resolve

certain lands in o^


freely to be given to the use aforesaid.

have power





y* y® said

Feofees shall

a Rate for the necesary charg of improving the s'd

they giving account thereof to the Towne, or to those



John Hunting, E^^"^, Eliazer Lusher, Francis ChickerJohn Dwight & Michael Powell, are chosen Feofees and betrusted

should depute.

in the behalf of the school as afore said.

The school thus
was designed

established, to be

managed by Feoffees,

to continue seven years, as will appear in the

following record relating to the training ground
1644. 4^



Granted to the Feofees for the free schoole



ham for the use of the s'd schoole a parcel of the Training ground
much as shall be set out to them by the Towne, which said p'cel


granted from this present day unto the last day of the eighth month

which shall be in the year 1650.
Hen. Phillips deputed to set out the

Hen. Chickering,
s'd parcell of



Land above



Jno BuUard, Robt. Grossman, Hen Wilson, Jno N[ewton], Edw. Col- 
ver. Hen. Smith, Nath Colborne, Nath. Aldus, Hen Phillips, Sam^ 
Morse, Dan Morse, Jno. Morse, Jos. Kingsbury, Jno. Dwite, Lamb. 
G[enery], Edw. Kemp, Edw. Richards, Tho. Leader, Geo. Bearstowe, 
Jonath. Fairbanks, Mich Powell, Mich Metcalfe juno'^, Jno Metcalfe, 
Jno Frarey, Eli. Lusher, R[obt] Hinsdell, Pet. Woodward, Jno Guyle, 
Rich. Evered, Robt. Gowinge, &ce. 
The said Inhabitants, taking into Consideration the great necesitie 
of providing some means for the education of the youth in o^ s'd 
Towne, did with an unanimous consent declare by voate their willing- 
ness to promote that worke, promising to put too their hands to provide 
maintenance for a Free Schoole in our said Towne. And farther did 
resolve and consent, testifying it by voate, to rayse the summe of 
Twenty pounds p annu, towards the maintaining of a Schoole M^ to 
keep a free Schoole in our s'd Town. 
And also did resolve and consent to betrust the s'd 2(d£ per annu 
& certain lands in o^ Towne, formerly set apart for publique use, into 
the hand of Feofees to be presently chosen by themselves, to imploy the 
sd 20;^, and the land afors'd, to be improved for the use of the said 
Schoole : that as the profits shall arise from the said land, every man 
may be proportionably abated of his some of the s'd 20£ aforesaid, 
freely to be given to the use aforesaid. And y* y® said Feofees shall 
have power to make a Rate for the necesary charg of improving the s'd 
land : they giving account thereof to the Towne, or to those whom they 
should depute. John Hunting, E^^"^, Eliazer Lusher, Francis Chicker- 
inge, John Dwight & Michael Powell, are chosen Feofees and betrusted 
in the behalf of the school as afore said. 
The school thus established, to be managed by Feoffees, 
was designed to continue seven years, as will appear in the 
following record relating to the training ground : 
1644. 4^ 12°^*^. Granted to the Feofees for the free schoole in Ded- 
ham for the use of the s'd schoole a parcel of the Training ground so 
much as shall be set out to them by the Towne, which said p'cel is 
granted from this present day unto the last day of the eighth month 
which shall be in the year 1650. Hen. Chickering, Eli. Lusher & 
Hen. Phillips deputed to set out the s'd parcell of Land above said. 

OCR button

Jno Bullard, Robt. Crossman, Hen Wilson, Jno N[ewton], Edw. Col-
ver, Hen. Smith, Nath Colborne, Nath. Aldus, Hen Phillips, Sam’
Morse, Dan Morse, Jno. Morse, Jos. Kingsbury, Jno. Dwite, Lamb.
G[enery], Edw. Kemp, Edw. Richards, Tho. Leader, Geo. Bearstowe,
Jonath. Fairbanks, Mich Powell, Mich Metcalfe juno", Jno Metcalfe,
Jno Frarey, Eli. Lusher, R[obt] Hinsdell, Pet. Woodward, Jno Guyle,
-Rich. Evered, Robt. Gowinge, &ce. ~

The said Inhabitants, taking into Consideration the great necesitie
of providing some means for the education of the youth in o' s’d
Towne, did with an unanimous consent declare by voate their willing-
ness to promote that worke, promising to put too their hands to provide
maintenance for a Free Schoole in our said Towne. And farther did
resolve and consent, testifying it by voate, to rayse the summe of
Twenty pounds p annu. towards the maintaining of a Schoole MT" to
keep a free Schoole in our s’d Town.

And also did resolve and consent to betrust the s’d 20£ per annu
& certain lands in o° Towne, formerly set apart for publique use, into
the hand of Feofees to be presently chosen by themselves, to imploy the
sd 204, and the land afors’d, to be improved for the use of the said
Schoole: that as the profits shall arise from the said land, every man
; may be proportionably abated of his some of the s’d 20£ aforesaid,
freely to be given to the use aforesaid. And y* y® said Feofees shall
have power to make a Rate for the necesary charg of improving the s’d
land: they giving account thereof to the Towne, or to those whom they
should depute. John Hunting, E’", Eliazer Lusher, Francis Chicker-
inge, John Dwight & Michael Powell, are chosen Feofees and betrusted
in the behalf of the school as afore said.

The school thus established, to be managed by Feoffees,
was designed to continue seven years, as will appear in the
following record relating to the training ground :

1644, 44 12™°, Granted to the Feofees for the free schoole in Ded-
ham for the use of the s’d schoole a parcel of the Training ground so
much as shall be set out to them by the Towne, which said p’cel is
granted from this present day unto the last day of the eighth month
which shall be in the year 1650. Hen. Chickering, Eli. Lusher &
Hen, Phillips deputed to set out the s’d parcell of Land above said.
Google OCR
Jno Bullard, Robt. Crossman, Hen Wilson, Jno N[ewton], Edw. Col-
ver, Hen. Smith, Nath Colborne, Nath. Aldus, Hen Phillips, Sam1
Morse, Dan Morse, Jno. Morse, Jos. Kingsbury, Jno. Dwite, Lamb.
G[enery], Edw. Kemp, Edw. Richards, Tho. Leader, Geo. Bearstowe,
Jonath. Fairbanks, Mich Powell, Mich Metcalfe juno", Jno Metcalfe,
Jno Frarey, Eli. Lusher, R[obt] Hinsdell, Pet. Woodward, Jno Guyle,
Rich. Evered, Robt. Gowinge, &ce.
The said Inhabitants, taking into Consideration the great necesitie
of providing
Towne, did with an unanimous consent declare by voate their willing-
ness to promote that worke, promising to put too their hands to provide
maintenance for a Free Schoole in our said Towne
some means for the education of the youth in oF s'd
And farther did
resolve and consent, testifying it by voate, to rayse the summe of
Twenty pounds p annu. towards the maintaining of a Schoole Mr to
a free Schoole in our s'd Town.
And also did resolve and consent to betrust the s'd 20£ per annu
& certain lands in or Towne, formerly
the hand of Feofees to be presently chosen by themselves, to imploy the
sd 20, and the land afors'd, to be improved for the use of the said
Schoole: that as the profits shall arise from the said land, every man
may be proportionably abated of his some of the s'd 20 aforesaid,
freely to be given to the use aforesaid.
have power to make a Rate for the necesary charg of improving the s'd
land: they giving
should depute. John Hunting, Eldr, Eliazer Lusher, Francis Chicker-
inge, John Dwight & Michael Powell,
in the behalf of the sch0ol as afore said.
set apart for publique use, into
And yt ye said Feofees shall
or to those whom they
account thereof to the Towne,
are chosen Feofees and betrusted
The school thus established, to be managed by Feoffees,
designed to continue seven
following record relating to the training ground:
will appear in the
1644. 4d 12mo, Granted to the Feofees for the free schoole in Ded-
ham for the use of the s'd schoole a parcel of the Training ground
much as shall be set out to them by
e Towne, which said p'cel is
granted from this present day unto the last day of the eighth month
which shall be in the year 1650.
Hen. Chickering, Eli. Lusher &
Hen. Phillips deputed to set out the s'd parcell of Land above said.

The immediate conclusion is that the best result comes from the OCR button, followed by the DjVu, followed by Google OCR, and with the PDF dead last. But since the text looks fine in the PDF locally, this means the problem must actually be the function that extracts the text from the PDF on-wiki!

I haven't done systematic testing of pages in this work or across works, but the results above are quite suggestive. For one it means we should very strongly prefer the DjVu to the PDF as a source. It also justifies the reliance on the OCR button for those who routinely use that. And it raises the question of what the difference is between the OCR button and the OCR at Internet Archive. The OCR button uses Tesseract, but IA uses a commercial product if I'm not mistaken. It would be interesting to compare the text produced by the OCR button with that produced by the latest Tesseract directly to see if the results are comparable. --Xover (talk) 11:06, 23 July 2019 (UTC)

Hm, that is very interesting, thanks for sharing this info. As for better OCR with DJVUs than with PDFs: besides various problems with DJVU format (common browsers cannot read it), e. g. HathiTrust does not offer scans in DJVU, only in PDF :-/ Many contributors are able to download a PDF and upload it to Commons, but they are not always able to convert it to DJVU prior uploading. So the best solution is not to stop using PDFs, but to make Wikimedia projects environment friendlier towards them. --Jan Kameníček (talk) 11:21, 23 July 2019 (UTC)
I meant, when you have both DjVu and PDF available, you should always prefer the DjVu. Creating a new DjVu when one is not already available is a fiddly, obscure, and highly technical task: I've created my own tooling that makes it somewhat bearable, but it's not something I would suggest anybody else tries to do unless they are otherwise undaunted by such work. --Xover (talk) 11:31, 23 July 2019 (UTC)
@Xover: I understand that DJVU is shown better in the Page namespace. Does it applie also to DJVU converted from PDF? I personally would expect that any further processing deteriorates the text, so it is really strange if it improves after the conversion. --Jan Kameníček (talk) 11:43, 23 July 2019 (UTC)
@Jan.Kamenicek: Sorry, I missed this question. There is nothing inherently lossy about converting between PDF and DjVu where the text layer is concerned. We're converting structured data from one format to another, not doing OCR on an existing OCR text or stacking lossy compression algorithms. If the extraction from PDF in Mediawiki gives poorer results than the extraction of the exact same text from a DjVu, then any process that moves the text layer as is to a DjVu should give better results. But, of course, if you use the same method that Mediawiki uses (which gives bad results from good data) and puts that into the DjVu, then the conversion will give equally bad results (or maybe even worse results).
But I'm not familiar with the way PDF stores text layers internally or the existing tooling for extracting it. It is entirely possible that actually doing this in practice is too complicated to be realistic. So at least for now, my approach will be to generate a new DjVu, with a new OCR text layer, from original page images, when this issue crops up. And I'd be happy to help out anyone who doesn't have the tooling available to do it themselves. Just with no promises on turnaround time: it's a somewhat fiddly multi-step process where each step takes a significant amount of processing time. --Xover (talk) 09:59, 28 July 2019 (UTC)

Adding to Wikisource[edit]

Hi there,

I made a translation into English from German of the poem "Wandere!" by Heinrich Heine. A user Huon (administrator of Wikipedia) suggested I contacted the Wikisource to inquire if I possibly should have this translation added to the Wikisource. That is why I am writing. I visited "Wandere!" page (! ) on Wikisource but I am not quite sure how to do all this (adding the text, should it be just a translation or the original poem in German too, etc.). The poem itself and the translation of it is currently on English Wikipedia Heinrich Heine page ( ) under the heading "Wandere!".Sergeismart (talk) 16:13, 23 July 2019 (UTC)

@Sergeismart: Danke, we do welcome translations, as long as you agree to freely license them. In the {{header}} template that begins the work, include yourself under the translator field and make sure to add the appropriate copyright tag at the bottom. It may be something like this: {{translation licence | original = {{PD-old}} | translation = {{CC-BY-SA-3.0}}}}. See {{translation license}} for more on that. Then, add a link to it to Author:Christian Johann Heinrich Heine and it seems like you're all set. I look forward to it being added here. —Justin (koavf)TCM 18:26, 23 July 2019 (UTC)
Hey, Justin (koavf), will you check it out here -! - if I did more or less fine (I mean if all spaces are filled correctly, etc.). And by the way I will appreciate your thoughts on the translation itself (if you have a spare second to read through it). Thanks.
Sergeismart (talk) 20:07, 23 July 2019 (UTC)
Ja, meine Duetsche nichte ist gut but I'll take a look. Odd about the translation notes below, because we definitely do have locally-hosted translations that are not in that namespace and that use the standard header template. —Justin (koavf)TCM 21:53, 23 July 2019 (UTC)
@Koavf: Independently published translations (like other independently published works) go in mainspace and use {{header}}. User-generated translations go in Translation namespace and use {{translation header}}. —Beleg Tâl (talk) 01:40, 24 July 2019 (UTC)
@Beleg Tâl: We do, in fact, have local pages that are not in the Translation namespace like Bible (Wikisource)/2 Chronicles. Likely an error but true nonetheless. —Justin (koavf)TCM 01:51, 24 July 2019 (UTC)
@Koavf: yes, that's an error, and thanks for identifying it. —Beleg Tâl (talk) 01:54, 24 July 2019 (UTC)
@Beleg Tâl: Lots more examples: Category:Wikisource translations of works in Thai. Not sure how interested you are in fixing them all but there are several dozen. —Justin (koavf)TCM 02:00, 24 July 2019 (UTC)
I've taken care of Category:Wikisource translations up to about P so far. —Beleg Tâl (talk) 02:12, 24 July 2019 (UTC)
I will just add that original translations by Wikisource contributors (unlike published translations)
For an example see Translation:Absolute.
If you have any further questions, do not hesitate to ask. And welcome at Wikisource :-) --Jan Kameníček (talk) 20:17, 23 July 2019 (UTC)
I just added " Category:Wikisource translations " at the bottom of an edit, couldn't do it through adding "Translation header" in the beginning of an edit (for some reason it didn't go through, it went very well though with just a "header" template). Hope now it is more or less fine. Enormous thanks for your help.
Sergeismart (talk) 20:56, 23 July 2019 (UTC)
@Sergeismart: I have made the following fixes:
Beleg Tâl (talk) 21:23, 23 July 2019 (UTC)
@Beleg Tâl: I was looking at other works in the translation namespace and it seems to me that the prevailing custom is to use the translated (English) name of the work in the name of the page after "Translation:" --Jan Kameníček (talk) 21:44, 23 July 2019 (UTC)
@Jan.Kamenicek: It appears to be at the discretion of the translator(s) what title they want to give their translation. —Beleg Tâl (talk) 01:40, 24 July 2019 (UTC)

Hey, Beleg Tâl, I greatly appreciate it! No need to add myself as an author of a translation into the "translation license", right? Just asking cause User:koavf told me previously to include myself in the "translator" field. Thanks again. Sergeismart (talk) 21:51, 23 July 2019 (UTC)

@Sergeismart: The {{translation license}} template is for license tags, not for authorship credits. User-submitted translations, like Wikipedia articles, are collaborative projects, so each edit is credited individually on the page history page instead of in the header. —Beleg Tâl (talk) 01:40, 24 July 2019 (UTC)
@Beleg Tâl: Thanks for your response and for your help. It is greatly appreciated. - Sergeismart (talk) 11:39, 24 July 2019 (UTC)
@Sergeismart: Two more things: (1) The author's name is not part of the poem text, and so should not be included in the poem or translation. The original certainly doesn't have it there. (2) There is no reason to italicize the entire poem or its translation. The original is not italicized. --EncycloPetey (talk) 18:25, 25 July 2019 (UTC)

Tables of Contents[edit]

In the past I have mostly relied on {{Dotted TOC line}} when building tables of contents, and in many cases that works fine.

I've seen others use a table-based approach, rather than a template on each line. I don't see anything under help:Table of Contents or WS:Table of Contents, and I'm wondering if somebody could help me learn this other system.

I'm currently looking at this page, and related ones: Page:American Historical Review vol. 6.djvu/9 What's the best approach for a page like that, with 3 columns for text entries? Maybe somebody could get it started on that page, and I could carry it to completion? -Pete (talk) 22:08, 26 July 2019 (UTC)

That page uses {{TOCstyle}}, you can read the documentation on its template page to get up to speed. I personally would use {{TOC row 1-dot-1}} and its related templates, which are also documented on their own template pages. —Beleg Tâl (talk) 12:46, 28 July 2019 (UTC)

OCR failure is not universal[edit]

I work in several books in the same session, and found an inconsistent OCR behaviour. In this volume, the OCR works, and in this volume, OCR deletes the text on save. — Ineuw (talk) 04:02, 27 July 2019 (UTC) P.S: Please test the OCR repeatedly on the page. — Ineuw (talk) 04:05, 27 July 2019 (UTC)

I tried it and confirm that it behaves the way you have described. --Jan Kameníček (talk) 07:33, 27 July 2019 (UTC)
@Jan.Kamenicek: It seems that Phabricator found the problem and by comparing the results of the two pages and I hope that the solution will soon be found and applied. — Ineuw (talk) 22:26, 28 July 2019 (UTC)

Need index verified[edit]

Index:The Life of the Spider.djvu is completely verified except for the index (pages 401–404). Could someone verify those 4 pages? Thanks! Kaldari (talk) 11:24, 28 July 2019 (UTC)

Yes check.svg Done [@Kaldari] --Xover (talk) 18:40, 28 July 2019 (UTC)

Including indexes in exports[edit]

Just finished validating my first Wikisource text, The Life of the Spider (with some help from Xover above). One thing I'm not sure about is linking to the book index from the Wikisource page. The book index isn't listed in the table of contents, so it isn't linked to from the Wikisource page, and thus isn't included in exported versions. Is there a way to fix that (aside from just transcluding the book index directly into the Wikisource page)? Kaldari (talk) 15:06, 29 July 2019 (UTC)

I think I would have gone with treating it as a chapter (its own subpage) and used the previous/next links in the header to link it in. And if the original table of contents is insufficient for some reason, {{AuxTOC}} is usually used to supply the want. But I don't think I've run across this specific issue previously so you may want to hold off and see if anybody else has a better or more definiitive answer. --Xover (talk) 16:55, 29 July 2019 (UTC)
Yep, I'd just stick an {{AuxTOC}} at the start/end (as appropriate) of the real TOC. For example, The Amazing Emperor Heliogabalus has a Preface that isn't listed in the real TOC. Its quite common for front matter and indexes. It could probably do with a note somewhere if this is actually good practice. Inductiveloadtalk/contribs 17:27, 29 July 2019 (UTC)
Thanks for the suggestions! I tried the {{AuxTOC}} template, but I didn't like that it created a separate TOC in the export that wasn't part of the original book. Instead I ended up just adding a hidden link to the Index from the existing TOC (Page:The Life of the Spider.djvu/10) which actually worked perfectly. Kaldari (talk) 02:00, 30 July 2019 (UTC)

The pages tag is supposed to add a Back and forward link?[edit]

It was my understanding that header=1 added forward and back buttons:- but here it's use here didn't immediately add them.

Where is the mistake so I don't have to run around playing hunt the quirk again, when trying to rely on what has and hasn't been purged in the backend yet? ShakespeareFan00 (talk) 17:12, 2 August 2019 (UTC)

For anyone trying to use this, the way it works is that when you add |header= to the <pages /> tag, ProofreadPage automatically inserts a {{header}} template with automatically generated variables for |previous= and |next=. To find the right previous and next chapter it uses the work's Index: page as a table of contents: every link to mainspace (i.e. ignoring links to Author: and Portal: pages) is assumed to be a chapter link, except the first link which is assumed to be the link in the title field to the work's main page. This last point was what tripped up ShakespeareFan00 here: in the Index: page the work's title wasn't linked, so ProofreadPage instead ignored the link to the first chapter. Thus the previous link on chapter 2 didn't get filled in (but subsequent chapters' previous links did). --Xover (talk) 18:11, 2 August 2019 (UTC)

How should I add incomplete indexes to Author pages?[edit]

Hello! I've recently upload scans of Nutcracker and Mouse-king by E. T. A. Hoffmann to commons and i've uploaded a Index to wikisource. I'm going to start transcription soon, but I would like some help. Can I add a link to the index on the authors page and if I can, how do I do that? If I can't add a link to authors page, where on Wikisource can I ask for help and/or boost the index's popularity? Pago95 (talk) 18:39, 4 August 2019 (UTC)

In my experience, you can choose one of three ways:
  1. If the work has been partially transcluded, place {{incomplete}} on the transcluded work page, and do not link to the Index page from the Author page.
  2. Otherwise, in general you would use {{small scan link}}; see Author:Dwijendra Nath Neogi for example.
  3. If there are a lot of works listed on the Author page and it looks too cluttered when you use {{small scan link}}, or if you want to distinguish between works that have scans and works which were added but do not yet have scans, you can use {{scan}}; see Author:William Schwenck Gilbert for example.
Beleg Tâl (talk) 19:15, 4 August 2019 (UTC)

Links to non-Commons files for speech transcriptions?[edit]

I'm interested in adding transcription of speeches by, and interviews with, US Government people (so no copyright issues). I believe that this type of content is welcome at Wikisource ( I have two questions.

1) How can I add a speech transcription entry to Wikisource without first uploading a file to the Commons? I understand from the Beginners Help how to create a Wikisource entry from a Commons upload, but what if I haven't done one? ETA: I realize now there is no requirement to have a Commons upload.

2) If I have a link to video or audio of the actual speech, but the link is not to the Commons, what is the format to include this link in the Wikisource entry? For example, this is a transcription of an Obama Weekly address which can be viewed at Youtube. I think the entry would benefit from a link to the video, am I wrong and such external links are to be avoided?

-Dennis.the.Anarcho-syndicalist (talk) 04:13, 5 August 2019 (UTC)

Disambiguate plus Create an Author Page[edit]

I MUST create an author page for a "James Orton." He is not the same one who populates the existing "Author:James Orton" identified as an American Naturalist. ALL the background ID available on my "James Orton" reads as follows:

  • James Orton, Dublin, Author of "The Enthusiast," "The Three Palaces," &c., &c.

    My Jame Orton contributed 4 brief biographies to the "Imperial Dictionary of Universal Biography." Both authors wrote in the mid to late 19th century. I have found a few references in periodicals but have not been able to add a middle initial or vital statistics. What is the recommended remedy?Klarm768 (talk) 17:05, 7 August 2019 (UTC)

Have you used Wikidata at all? One good first step would be to ensure that each has a thorough Wikidata entry. The naturalist has an entry: wikidata:Q6140665 It would be good to start another for the other Orton.
More specifically, it's my impression that English Wikisource policy is generally to use the most complete name of the author, as opposed to the name by which they are best known. So if there is a known middle name for either, that's the most straightforward way to distinguish them. If not, Help:Disambiguation#Authors says we should use birth and death years to distinguish. -Pete (talk) 17:30, 7 August 2019 (UTC)
I can find no dates for the author of The Enthusiast &c., so a floruit date would be the next best thing: i.e. Author:James Orton (fl. 1843-1859), and the existing author page should be moved to Author:James Orton (1830-1877). —Beleg Tâl (talk) 17:40, 7 August 2019 (UTC)
I can see a James Orton in Ireland in the 1850s, and early 1860s. James Orton married Julia Colthurst in 1853, he was of St. Margaret's Bay, Kent. (Dover?) Wife died 1857 (seems to be childbirth), and daughter 7 months later, at 7mo. This gent is Bray, Wicklow (selling up 1862). Directories and newspapers shows James Orton an inspector of branches (People's Provident Assurance Society) in Dublin and Bray in 1858. [Baptism of John Nicholas Colthurst Orton in Westminster in 1854, so this is a UK & Ireland search. JNCO cited as actor in census. Death of same, also called Charles Overton, 1898]
Thank you to all who have pitched in... particularly @Beleg Tâl:. I was going to need further help implementing BT's advice, but it appears BT has performed what was proposed. I wish I could say that understand what BT proposed and appears to have done. Perhaps then I could do the remainder myself. Wikidata is mysterious to me. HOWEVER I have 3 remaining disambiguate+create problem authors. Should I create a new Scriptorum Section for each or continue this one? Klarm768 (talk) 14:35, 8 August 2019 (UTC)
The steps are as follows:
That's really all there is. A wikidata item for the new author will need to be created, but you can ask someone else to do it for you, or you can look at d:Wikidata:Tours to learn how to do it yourself. —Beleg Tâl (talk) 14:42, 8 August 2019 (UTC)
Thanks. I've now started the Wikidata Tours. A number of the IDUB contributors are not published elsewhere, no dates of birth/death are provided, often possess rather common names, are distinguishable by their profession/city/academic&heraldic_credentials. Overall I am proud of my detective work on 230+ discoverable contributors. BUT My last three provide least of any distinguishable facts, their Author:names are already in use. My numerously-configured Google searches are not adequate. More successful search methods are not mine. The dates of contribution-composition are not provided so floruit date could be 1854-1876.

I can provide a titles of their contributions and/or the full content of contributions. The TOTAL contributor data points are these:

  • Rev. James Anderson, Edinburgh.
  • James Barr.
  • Frederick M. White, Aylesbury.
Klarm768 (talk) 16:09, 8 August 2019 (UTC)
Possibly useful: Frederick M. White contributions (numbering about 370) are mostly biographies of Spanish historical and then-living contemporary figures. Klarm768 (talk) 16:23, 8 August 2019 (UTC)
I think James Anderson (1804-1863) is the first one; see these ads in the Imperial Dictionary for reference. —Beleg Tâl (talk) 17:50, 8 August 2019 (UTC)
Frederic(k) Meriton White (1828-1895), the brother of w:Jessie White Mario, wrote letters on Italian affairs signed "Frederick M. White, Aylesbury". His page will be Author:Frederic Meriton White, because his full name is unique. —Beleg Tâl (talk) 18:05, 8 August 2019 (UTC)
James Barr, it appears, wrote only a short article on the Lutheran theologian Daniel Hoffman, which may be a clue —Beleg Tâl (talk) 18:31, 8 August 2019 (UTC)
Thanks for your help... BUT I failed to include parentheses and I created Author:James Anderson 1804-1863. So now I need advice on how to remedy that.

I find 9 contributions by Barr. Most of them are ecclesiastics; but not of any particular denomination, nation, or doctrinal theme that I recognize. I observe that not only Barr's but many of the signatures fall in a segment of the alphabet. I have surmised that they were contracted for a "duration" (sub-editor switches were numerous). The original IDUB was created from A-to-Z. This might explain why all of Barr's biographies start with G, H, I, J, or L. At least one of the less august contributors was employed while a University student. Perhaps contributions ceased with graduation or acquiring a better job elsewhere. Klarm768 (talk) 20:35, 8 August 2019 (UTC)

I finally figured out how to perform "Move" and fixed Author:James Anderson (1804-1863) myself. Klarm768 (talk) 21:19, 8 August 2019 (UTC)
I think that I performed the 4 steps described above creating Author:James Barr (fl. 1854-1863). Please advise of any deficiencies. Thanks! @Beleg Tâl:

I happen to possess a physical copy of the 1863 Edition and could confirm that Barr's contributions are there. His originals were probably first printed 1856-1860. I know of no digital version of 1863 Edition. Klarm768 (talk) 21:48, 8 August 2019 (UTC)

Mark above line[edit]

On this page, about the middle of the right column, there is an odd character: a rho with a wavy line above it. This squiggle is intended to reproduce a manuscript mark and looks to have been added to the type plate freehand. How should it be reproduced digitally? Levana Taylor (talk) 01:44, 8 August 2019 (UTC)

A Rho or a P with a tilde. For our grade of work, it should be suitable to use w:P̃ billinghurst sDrewth 05:05, 8 August 2019 (UTC)
Nonono don't use Latin letters to transcribe Greek! The correct character is Rho + combining tilde: ΧΡ̃Ο ΧΡ&#x303;Ο or Χ{{subst:cdm|Ρ|303}}ΟBeleg Tâl (talk) 10:36, 8 August 2019 (UTC)
I agree with Beleg Tâl. An alternate method (as if you need more!) could be: <math>\Chi\tilde{\Rho}\Omicron</math> which renders as: . Incidentally in case you are wondering the sequence in Greek is usually translated as Christo (as in Jesus)… 11:44, 8 August 2019 (UTC)
Cool, thanks for the math code. I had actually tried the combining tilde and noticed that it’s not suitable for capital letters because of rendering on the letter instead of above. But ‘math> can do so much. Levana Taylor (talk) 15:13, 8 August 2019 (UTC)
The combining tilde renders above the letter. Your system may not work right, but it works right on this system (a generic Windows 10 system), and hopefully the percentage of system that have problems with it will just decrease over time. The math code is less well supported and I'm pretty sure it won't work with screen readers, whereas I'm pretty sure that ΧΡ̃Ο will at least get spelled out.--Prosfilaes (talk) 01:10, 9 August 2019 (UTC)
Of the two browsers on my laptop, Microsoft Edge renders the combining tilde correctly and Firefox doesn’t. Can some other people try out Safari, Chrome, and mobile browsers? Levana Taylor (talk) 23:55, 9 August 2019 (UTC)
For me capital Rho + combining tilde as suggested by Beleg Tâl display well both in Firefox and Chrome. --Jan Kameníček (talk) 07:37, 10 August 2019 (UTC)
It's going to depend a lot on more things than just the browser. Edge and Chrome on Windows 10, and Chrome on my Android system all display it correctly.--Prosfilaes (talk) 07:47, 10 August 2019 (UTC)
OK, I should stop worrying then. Somehow I hit the one browser setup (Firefox + whatever addons on Windows 1) that wasn’t working correctly. Probably won't happen for anybody else. Levana Taylor (talk) 08:02, 10 August 2019 (UTC)

Add floruit range to Module:Author[edit]

Would it be possible for someone to add floruit ranges to Module:Author? The relevant properties are d:P:P2031 and d:P:P2032. I could probably figure it out, but it would take a while. —Beleg Tâl (talk) 12:53, 8 August 2019 (UTC)

Maybe Sam Wilson could have a look at it, we discussed it some time ago. --Jan Kameníček (talk) 20:55, 8 August 2019 (UTC)

Navaho Legends or Memoirs...?[edit]

I'd like to move this book from Commons over to Wikisource. I've created a few books before from Commons to Wikisource but I don't know what to name this one. This is what I named it on Commons "File:Navaho Legends.djvu" But maybe I was supposed to name it "Memoirs of the American Folk-Lore Society V". This one in Wikisource is from the same series.
Index:Memoirs of the American Folk-Lore Society XVII.djvu
Should I have named the original file, on Commons, Memoirs of the American Folk-Lore Society V ?
And then there's this also, already on Wikisource, with a red link for Navaho Legends: Memoirs of the American Folklore Society Please help with advice or please go ahead someone and do it because I have no idea how to proceed and I would like to start reading and working on the Navaho Legends. Also note that same book is available on except there is an image that doesn't show on the book in Commons AND images are in are in color. Thank you much.--The Eloquent Peasant (talk) 00:24, 9 August 2019 (UTC)

Sorry for answering my own question. I've requested the file in commons be moved to: File:Memoirs of the American Folk-Lore Society V.djvu so then I can create the index file here. I think this is the correct thing to do. Thanks! The Eloquent Peasant (talk) 02:03, 9 August 2019 (UTC)
Ultimately the name of the file doesn't really matter, but it's preferred (especially on Commons) if files in a series have consistent naming. Since the other files in this series have the format "Memoirs of the American Folk-Lore Society #.djvu", I think you were right to rename the file. —Beleg Tâl (talk) 14:53, 9 August 2019 (UTC)

Translation:Order No. 227 by the People's Commissar of Defence of the USSR[edit]

I recently found Translation:Order No. 227 by the People's Commissar of Defence of the USSR, which though its original text is in Russian public domain (as per Template:PD-RU-exempt). Yet the translation still cannot give out a license for translation. Is that piece of text eligible for deletion? Thank you.廣九直通車 (talk) 13:12, 9 August 2019 (UTC)

This is a Wikisource original translation and as such the translation is licensed under CC BY-SA. —Beleg Tâl (talk) 16:30, 9 August 2019 (UTC)

Problems with specific transclusion[edit]

Despite my repeated attempts, I have been unable to properly transclude this work, do to a number of incorrect line breaks added by the <pages /> command. Is there any way to fix this, other than to either fully correct the file, or to give all of the pages as templates (i. e., {{Page:}})? TE(æ)A,ea. (talk) 15:11, 9 August 2019 (UTC).

Took me a couple of goes, but the answer according to H:POEM is a <br/> before the poem tag end to force a stanza break. Inductiveloadtalk/contribs 16:05, 9 August 2019 (UTC)
  • That’s the problem; the poems are transcluding as intended. It is the unnecessary page breaks inside of a paragraph that are the problem. “[A]nd that no State could/either…” should be in one paragraph, not two (here). I’ve used (perhaps broken) <pages index="Marmor Norfolciense.djvu" include=25-55,64-65,56-63,82,66-78 />; that shows the correct “page” range. TE(æ)A,ea. (talk) 17:35, 9 August 2019 (UTC).
    • Since the paragraph that starts on page 55 is continued on page 64, I think that this include method is the only way to preserve that flow within the pages tag. —Beleg Tâl (talk) 18:26, 9 August 2019 (UTC)
      • That is what is currently being used; however, it is not working. With the “include” parameter, regardless of given order, it proceeds in numerical order. This causes problems, as the scan is not in numerical order. TE(æ)A,ea. (talk) 23:26, 9 August 2019 (UTC).
        • I see what you mean. I think I have an idea of how we could make it work, but honestly I think it would be better to just fix the scan. —Beleg Tâl (talk) 01:30, 11 August 2019 (UTC)
          • I have fixed the scan; there remains only to find images of pages 41 and 42 to replace the blank placeholders. —Beleg Tâl (talk) 02:32, 11 August 2019 (UTC)

Remove Index (Delete because it's missing images)[edit]

I'd like to have the Index:Memoirs of the American Folk-Lore Society V.djvu removed- it came from archives and is missing tons of images.
A better version of the same is on and it's not missing images.
--The Eloquent Peasant (talk) 03:56, 11 August 2019 (UTC)

Upload the replacement file over the top of the current one. Beeswaxcandle (talk) 04:06, 11 August 2019 (UTC)
@Beeswaxcandle: Would you be able to point me to instructions on how to do that? I don't know how to upload a .pdf from books google. Thank you.--The Eloquent Peasant (talk) 04:13, 11 August 2019 (UTC)
It is slightly complicated as you want to upload a .pdf file, while the current one is .djvu. I personally would upload the pdf version to Commons independently of the original djvu, and then started a new index page. (The old one can be deleted then).
To get the book to Commons you need first to download it from Google Books to your computer, and then you can upload it to Commons using their UploadWizard. Unfortunately I cannot do it for you, as Google Books limit the access to many publications of theirs (including this one) in my country. If you need any other help, just ask. --Jan Kameníček (talk) 07:15, 11 August 2019 (UTC)
I did it! Thanks! 1) I downloaded the Google version (with images). 2) I converted it to djvu format. 3) I uploaded it over the old version (without images). Now I'll fix the pagelist. Thank you.The Eloquent Peasant (talk) 12:36, 11 August 2019 (UTC)
Nice job! Thanks for getting this squared away. -Pete (talk) 18:14, 11 August 2019 (UTC)
Thanks everyone for your tips!--The Eloquent Peasant (talk) 16:41, 13 August 2019 (UTC)

Wrong order of pages[edit]

A scanned file has got two wrongly ordered pages (no. 111 and 112). Having checked it with a different copy I found out that the original book has the pages correct and so it is only a fault of this particular scanned file. Unfortunately, the other copy has many other flaws, and so it is not possible to replace it. Could somebody skilled with djvu files fix it and reorder the two pages, please? --Jan Kameníček (talk) 18:30, 12 August 2019 (UTC)

Done. Please check, just in case ...Mpaa (talk) 22:30, 12 August 2019 (UTC)
Perfect, thanks. Now I'll just have to wait until the index pages get updated as clearing the cash has not helped, but I guess it is only a matter of time. --Jan Kameníček (talk) 06:59, 13 August 2019 (UTC)

Help with botched page move[edit]

In the midst of some discussion about book titles (see here if you want the background), I moved Vol 1 History of Mexico by H H Bancroft to History of Mexico (Bancroft)/Volume 1. I expected to see a checkbox that would have permitted simultaneously moving all the sub-pages (the chapters) as well, but I didn't see that option. I foolishly forged ahead, hoping that the checkbox I remembered was on the next screen in the process; but of course, it wasn't.

(1) Where did I go wrong, and (2) (more importantly) Is somebody able to move the sub-pages to clean up my mess?

The answer to (1) is actually somewhat important, because there are 3 other volumes in this series that will need similar treatment.

Please note, in this case the works have existed for some time, so it's important to preserve redirects rather than deleting the original page names. -Pete (talk) 06:03, 13 August 2019 (UTC)

You didn't go wrong, you just don't have the ability to do it because you're not an admin. Let me know on my talk page the exact names of origin and target that you want the other volumes moved to. Beeswaxcandle (talk) 06:53, 13 August 2019 (UTC)
OK. Now moved with subpages. I'll leave you to fix the links to the old title and the internal links in the headers and TOC. Beeswaxcandle (talk) 06:59, 13 August 2019 (UTC)
Ah, of course. I guess I'm remembering from other wikis where I'm an admin. I just didn't realize that checkbox was an admin thing. Thanks for doing that, and yes, I'm happy to dig through and fix any remaining idiosyncracies. -Pete (talk) 07:48, 13 August 2019 (UTC)

My custom edit toolbar gone again...[edit]

...which I could live with if I could possibly bring the Insert/Wiki markup, etc. toolbar from the bottom of my editing space to the top. Is this possible? Thanks, Londonjackbooks (talk) 04:11, 14 August 2019 (UTC)

please include athorisms of G.C. Lichtenberg[edit]

G.C. Lichtenberg was Germany's first professor of physics who had laid the theoretical groundwork for the Xerox machine by the mid-eighteenth century. Severely hunchbacked , he nevertheless became one the great aphorists of the age, recording his thoughts and observations - such as "Everyone is a genius once a year. A real genius has his original ideas closer together." - in a series of alphabetical "waste books"(he'd reached volume "l" by the time he died in 1799). Can you include the waste books, please ? Thank you! —Preceding unsigned comment added by (talk)

Feel free to upload a public domain translation of this work: see Help:Beginner's guide to adding texts for how to get started. —Beleg Tâl (talk) 13:02, 14 August 2019 (UTC)
The only available book seems to be Lichtenberg's Reflections. I have uploaded it for you and created the index page, so if you want, you can start transcribing. I have also proofread two pages as an example, see e. g. Page:The reflections of Lichtenberg.djvu/28. If you need some advice during the work, just ask, somebody will always help you. Happy editting! --Jan Kameníček (talk) 15:41, 14 August 2019 (UTC)

History of Colonial Zimbabwe (Southern Rhodesia) during 1800 to 1921[edit]

My name is Dr. Clement Jumbe

My account number is: EB* Britannica 800-3231229 IL 84.69 USD @ 1.346600000000

Will you please send me the history of how the Native people in reserves where administered during the Company rule up to 1921

Clement Jumbe

IwPageSection generates badly formatted HTML...[edit]

<span class="iwpage" title="fr|Ruffhead - The Statutes at Large, 1763.djvu/206|viscount" id="fr:Ruffhead_-_The_Statutes_at_Large,_1763.djvu/206:viscount"><div class="mw-parser-output"><div class="pagetext"><div class="mw-parser-output"><p><br>
rien receu des dettes le Roi &amp; qe riens ne prendretz par quoi le Roi perde ou par quoi droiture soit desturbe ou la dette le Roi delae e qe loiaument freez retourner &amp; loiaument suire les brefs le Roi a vostre fen &amp; a vostre poair &amp; qe vous ne prendretz nul ballif en vostre service por qi vous ne voletz respoundre et qe vous swreez voz ballifs faire autiel serment come a eux apent et que nul brief ne receivretz, par vous ne par les voz souz seal des Justices forsqe en eyre ou autres justices assignetz en meisme la counte ou Justice de Neugate &amp; qe vous mettretz voz ballifs de plus loiaux du pais &amp; qe nul ballif ne ministre qi ad este od lautre viscounte retendretz en vostre service.
<span class="iwpage lang-en" title="en|Ruffhead - The Statutes at Large, 1763.djvu/206|p160-n1" id="en:Ruffhead_-_The_Statutes_at_Large,_1763.djvu/206:p160-n1" lang="en"></span>
NewPP limit report
Parsed by mw1290
Cached time: 20190816083951
Cache expiry: 2592000
Dynamic content: false
Complications: [vary‐revision]
CPU time usage: 0.008 seconds
Real time usage: 0.011 seconds
Preprocessor visited node count: 34/1000000
Preprocessor generated node count: 0/1500000
Post‐expand include size: 1917/2097152 bytes
Template argument size: 176/2097152 bytes
Highest expansion depth: 4/40
Expensive parser function count: 0/500
Unstrip recursion depth: 0/20
Unstrip post‐expand size: 0/5000000 bytes
Number of Wikibase entities loaded: 0/400
Transclusion expansion time report (%,ms,calls,template)
100.00%    7.960      1 Modèle:Lst
100.00%    7.960      1 -total
 25.16%    2.003      1 Modèle:IwpageSection

This is malformed as a DIV cannot be inside a SPAN. LintErrors doesn't detect this,(and currently there is no visible impact as far as I can tell) but as I expressed previously on a different issue trying to shorehorn everything into a SPAN is leading to longer term issues. It's time certain contributors with the expertise, took another look at this. ShakespeareFan00 (talk) 08:47, 16 August 2019 (UTC)

Updated ShakespeareFan00 (talk) 10:07, 16 August 2019 (UTC)
I agree that this does not look like a good place to use <span></span>. By its nature it will regularly and quite appropriately contain block elements, and this will lead to all sorts of weirdness and undesirable behaviour. My initial thought is that this should be changed to <div class="iwpage-section"></div>, and site CSS changed to treat that wrapper as .iwpage-section {display: inline}. No guarantee this will not break anything that relies on the old quirks, but I would be inclined to argue that those instances are in principle broken to begin with. --Xover (talk) 10:11, 16 August 2019 (UTC)
I was saying eslewhere, that the iw templates and {{Page}} could ideally be a module, to centralise the core behaviour and provide options to overcome some of the current limitations, like in some instances the desirability of getting the semi-'raw' markup back, so that a local wiki could reprocess it slightly (However as templates are widely divergent on other wikiks, that proabably too risky.).ShakespeareFan00 (talk) 10:16, 16 August 2019 (UTC)
As such, see Special:WhatLinksHere/Template:IwpageSection so far only 1 work is actively using this, so it can be considerably overhauled. ShakespeareFan00 (talk) 10:18, 16 August 2019 (UTC)
I also note - had some performance improvments susequent to the local version. Sometimes I wonder if certain core code like this should be in a Centrally updated repository. (hard look at heavens). ShakespeareFan00 (talk) 10:37, 16 August 2019 (UTC)
The templates don't do much that's worth modularising; they just spit out a span or div with the provided params concatenated into the title or id attributes. The actual work happens in JS at MediaWiki:InterWikiTransclusion.js (globally loaded from MediaWiki:Common.js), which is a three year old copy of mul:MediaWiki:InterWikiTransclusion.js (I don't think the intervening changes are particularly relevant to this discussion). The templates expose the provided parameters as html attributes (id/title) which the javascript then picks up, and then requests the relevant content from the remote wiki using the MediaWiki API, and finally inserts it inside the span or div generated by the template here. In other words, the only bit worthwhile centralising is the javascript, and that is already centralised (it's just insufficiently maintained, but that's a separate matter).
A central repository for templates and Gadgets that are available across projects is on a long-term wishlist at the WMF, but nothing much is being done with it that I am aware of. The problems are not primarily technical: separate projects have separate goals and practices, making sharing code often be challenging. --Xover (talk) 10:53, 16 August 2019 (UTC)
I will also note the concerns expressed here User_talk:Beleg_Tâl#Page:Ruffhead_-_The_Statutes_at_Large,_1763.djvu/206 concerning whitespace insertion, there may be a need for a new template or approach in terms of my efforts. ShakespeareFan00 (talk) 11:15, 16 August 2019 (UTC)
Attempted this , but it doesn't work : {{IwpageSection/sandbox}} ShakespeareFan00 (talk) 12:11, 16 August 2019 (UTC),
See -,_1763.djvu/206&oldid=9543090 ShakespeareFan00 (talk) 13:15, 16 August 2019 (UTC)

<pages>tag fails when placed inside <ref></ref>[edit]

See the reference for Page:The seven great hymns of the mediaeval church - 1902.djvu/31 ShakespeareFan00 (talk) 10:30, 16 August 2019 (UTC)
<ref>…</ref> is itself technically provided by an extension (mw:Extension:Cite) and as such has severe limitations in what you can do inside of it: in particular, executing another extension (mw:Extension:Proofread Page) inside it is highly unlikely to work. --Xover (talk) 11:03, 16 August 2019 (UTC)
So is there a credible alternative approach, it did work at some previous date as the page was validated. ShakespeareFan00 (talk) 11:08, 16 August 2019 (UTC)
According to the ref documentation, you can use the #tag parser function for the ref tag. {{page}} inside a normal ref tag might work too. Inductiveloadtalk/contribs 11:14, 16 August 2019 (UTC)
You're actually missing the mark a little. The <pages> tag works great when placed inside <ref></ref>. However, it does not work at all in Page namespace. That is why the transcluded reference does not appear on Page:The seven great hymns of the mediaeval church - 1902.djvu/31, but it appears just fine on The seven great hymns of the mediaeval church/The Celestial Country. —Beleg Tâl (talk) 11:23, 16 August 2019 (UTC)
Then the second suggestion works: <ref>{{page|The seven great hymns of the mediaeval church - 1902.djvu/66|section=note01}}</ref>. The {{page}} template defers to LST's #section parser function, so it's independent of the pages tag. Regardless, the current state is not actually "broken" per se, since it does transclude correctly (edit: kinda, it has linter errors) (and the original work doesn't have the note on that page anyway). Inductiveloadtalk/contribs 11:30, 16 August 2019 (UTC)
HTH (sigh) ShakespeareFan00 (talk) 11:55, 16 August 2019 (UTC)
The LintHint script is however claiming for [[6]] is there's a DIV SPAN swap sowmewhere though. but I can't see it visually in the output, so hardly a priority. ShakespeareFan00 (talk) 12:02, 16 August 2019 (UTC)
That would be because there is the following construction for each ref:
<span class="reference">
    <div class="prp-pages-output">
The solution above with {{page}} doesn't suffer from this, as there is no class="prp-pages-output" div.
You could also do this all manually with anchors and links if refs doesn't get you what you want. The reference extension is just a helper. Inductiveloadtalk/contribs 12:41, 16 August 2019 (UTC)

Pictogram voting comment.svg Comment Not sure why we are going around this roundabout yet again. It will never work successfully, and it is completely known that it will be problematic within wikis. So please stop the incessant complaining; it is getting beyond a joke.

See mw:Help:Magic words and #tag for the means to address these sorts of matters. Others use of #tag:ref and #tag:pages over many years has not been accidental. So please utilise the solution that has been around for years, and look at using template:authority reference as it will probably manage what you are doing; actively used in {{IrishBio ref}}. — billinghurst sDrewth 12:48, 16 August 2019 (UTC)

Thanks, if that's the approach that SHOULD be used, where would be an appropriate place to document it? ShakespeareFan00 (talk) 13:16, 16 August 2019 (UTC)
And I'm sorry if I appear to be raising the same issues repeatedly. ShakespeareFan00 (talk) 13:21, 16 August 2019 (UTC)
I think this is just phab:T49544 rearing its ugly head again. We should definitely not start removing block content from references just because the references algorithm hasn't been updated to accommodate them yet. In this case the current construction works exactly as intended. —Beleg Tâl (talk) 21:35, 16 August 2019 (UTC)

Why make this so ludicrously complicated? To use the existing symbolism:


<ref><pages index="The seven great hymns of the mediaeval church - 1902.djvu" from=66 to=66 onlysection="note01" /></ref>


<ref name="note01"> </ref>

N.B. Do not get clever and remove the space or use form <ref name="note01"/> as Cite is not without finicky limitations…


## note01 ##
"Le {{ls}}urnom de Bernard &c. &c.


<ref follow="note01">
"Le {{ls}}urnom de Bernard &c. &c.

Now was that so hard? No need for nested extension invocations at all! Oh and don't forget to rinse and repeat for each case. 23:07, 16 August 2019 (UTC)

Or, you could leave it alone since it's already validated and working as intended. I probably would have used this method if I had known about it when I proofread that work, but the method in place currently is fine. —Beleg Tâl (talk) 23:17, 16 August 2019 (UTC)
@Beleg Tâl: You speak as the person who tweaked the page twice (out of nine edits) subsequent to validating the thing. Even if what you were attempting unsuccessfully to do actually worked it would be wrong and result in a large amount of footnote content which should not appear in Page: space. So, umm—no. Do it right instead.

Recall I have no real dog in this fight so do not care if you persist in being stubborn, foolish or foolishly stubborn. 01:19, 17 August 2019 (UTC)

While it's great that you have given the valuable advice above and I guess everybody really appreciates it, the way of doing so reminds a golden rule from meta. --Jan Kameníček (talk) 07:19, 17 August 2019 (UTC)

Subpages in Page:namespace[edit]

Page:Ruffhead - The Statutes at Large, 1763.djvu/206/fr displays the first or last page not the actual scanned page. Working as designed but tiresome, because this would have been one approach to re-consolidating this page, or more complex layouts between many pages. ShakespeareFan00 (talk) 12:38, 16 August 2019 (UTC)

There is no scanned page, ProofreadPage simply has a default view for a non-existing page. Things are being displayed as expected, it is your actions that are not expected or within scope. Keep within scope, and things will work fine. — billinghurst sDrewth 13:20, 16 August 2019 (UTC)
As I indicated, this IS working as designed. What would be useful is the ability to transcribe different portions of a complex layout to different sub-pages, but using the same scanned image (an example would be a newspaper scan with many individual articles, in each 'frame' of the print layout ). This would not necessarily be the same as using sections. If this isn't currently supported fair enough, but it's a limitation.

ShakespeareFan00 (talk) 13:57, 16 August 2019 (UTC)

It's an interesting idea. Unfortunately, I don't think the devs would implement it; it would require a lot of work to implement and would not be used very much at all. I've seen the devs shut down ideas that are better than this and easier to implement. —Beleg Tâl (talk) 21:38, 16 August 2019 (UTC)
And sometimes, my seeming unreasonableness, comes from the fact that I do actually "care" (as you do.), I'm just more likely to bluntly express the frustrations I have about quirks and limitations, because there is seemingly (as you mentioned previously) not necessarily the community expertise for someone to sit down and take a closer look at the responsible code. In many instances Wikisource IS pushing the limits of what Mediawiki and wikitext was originally designed to do, and the flexibility offered means that slightly less experienced contributors (myself included) are only to able to create combinationally complex templates, that a more considered approach wouldn't have. If I seem to be complaining a lot, it's not because I like complaining, it's out of frustration at encountering the same concerns and temporarily lacking the recall of what had been suggested or decided on as a reasonable solution previously. Given this I am wondering if I should stop contributing, as I seem to unable at times to express 'technical' concerns without ranting, or the ability to reason if previous solutions, (even some I had suggested) would be applicable. ShakespeareFan00 (talk) 16:46, 16 August 2019 (UTC)
I'll strike this out, on re-consideration. ShakespeareFan00 (talk) 18:11, 16 August 2019 (UTC)