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?

Linking illustrations[edit]

I have just one bothersome issue with a work I've been chipping away at for a while. I need to finish the linking for the list of illustrations: Eleven years in the Rocky Mountains and a life on the frontier/Illustrations

It seems to me there are three possibilities:

  1. Link to the chapter, with an anchor to the location of the image
  2. Link to the "Page" namespace
  3. Link to the "File" namespace

Number 1 seems like the most appropriate, since it places the image in context, and will be generally familiar to the reader. Also, the link might have the chance (slim though it may be) of "surviving" a conversion to PDF or another format, and being functional in an offline version.

However, in a few cases there are two or three images in a chapter. (e.g., chapter 5.) So, unless there's a trick I'm missing, there's no way to link to the images with an anchor tag, since the anchor for both is the word "#plate". Both links would target the first image.

What's the best way to approach this? -Pete (talk) 04:07, 22 May 2018 (UTC)

In principle {{anchor}} is one possible solution. 07:55, 22 May 2018 (UTC)
What about using {{anchor+}} with the image as the visible text? Beeswaxcandle (talk) 08:15, 22 May 2018 (UTC)
Just link to the page number: Eleven years in the Rocky Mountains and a life on the frontier/Chapter 5#94 unless I misunderstand what you are trying to do? Londonjackbooks (talk) 10:17, 22 May 2018 (UTC)

Thanks all! I will use one of the anchor templates, where needed. @Londonjackbooks: that works fine for images that are on pages with text, but most of the images are on plates of their own, so there is no page number; the plate, for instance, comes between pages 5 and 6. I suppose I could link to page 5 in a case like that, but it feels...inelegant :) -Pete (talk) 22:41, 22 May 2018 (UTC)

@Peteforsyth: From the main namespace we should not be directly linking to the Page: ns—outside of the marginal page numbering. Readers will just confused and lost being dumped there. Our main namespace linking should only be to our presentation level pages. We have tried to have our linking overt and non-surprising, and even somewhere like author ns it is why things like {{ext scan link}} exist rather than hyperlinking a work name (our old practice). — billinghurst sDrewth 23:31, 22 May 2018 (UTC)
Thanks @Billinghurst:, sounds like my goal is in alignment with standard practice, then. Good to have that confirmation! Now that I have a plan, I'm hoping to wrap this last bit up in the next 24 hours. -Pete (talk) 23:56, 22 May 2018 (UTC)
@Peteforsyth: I was recently flipping through The Water Babies, and noticed that User:Cygnis insignis [ inelegantly ;) ] linked to previous pages before the appearance of plates in the illustration TOC for the work. It may have merely been a simple formatting choice, or there may have been a considered reason for such linking. I would guess the latter, but I would be prone to. Londonjackbooks (talk) 13:30, 26 May 2018 (UTC)


I am validating The Bet by Anton Chekhov. On pages 13, 21, and 23, I see "1", "B", and "B2" in their footers respectively. Should I insert them? 4nn1l2 (talk) 12:43, 25 May 2018 (UTC)

@4nn1l2: Those are binder's marks, not needed in digital version. Hrishikes (talk) 13:18, 25 May 2018 (UTC)

Reusability of a transcription CC0 published in a file[edit]

Hi there,

An acquaintance of mine working in a university of Taiwan has published here a transcription of this work Index:English-Chinese Vocabulary of the Vernacular Or Spoken Language of Swatow.djvu. I was wondering whether we could republish here, splitting and matching the pages? Do we need an authorization for that? In what form? He told me that the work of transcripting that dictionary was basically CC0.

Assassas77 (talk) 03:16, 26 May 2018 (UTC)

I think the answer is no, there isn't authorization needed -- here's how I see it, from what you've described:
  1. The work itself is in the public domain
  2. A strict transcription does not add any creative element (see "sweat of the brow" in copyright law), so -- according to my understanding (and I'm not a lawyer) it's not possible to attach copyright to the transcription of a work that is in the public domain.
  3. I'm pretty sure #2 is accurate, but even if I'm wrong -- if the transcription work is "basically" CC0, all you'd need is a written assurance that it is in fact CC0, published somewhere you can link to. -Pete (talk) 04:49, 26 May 2018 (UTC)
Taiwanese copyright rules are tricky in this case: c:Commons:Reuse of PD-Art photographs#Taiwan. 4nn1l2 (talk) 10:54, 26 May 2018 (UTC)
Though it only talks about about printing, and reproduction of that print. Electronic transcription, especially into a non-print format (a format not like PostScript or PDF), seems outside those rules.--Prosfilaes (talk) 23:06, 26 May 2018 (UTC)
Have you asked them if they are willing to release their manuscript under a Creative Commons license? ShakespeareFan00 (talk) 06:05, 27 May 2018 (UTC)
I'm gonna ask them more information about the license they are putting on this work. I mean... for now, it is only a raw file with no licensing information. I'll provide more information soon. Assassas77 (talk) 14:28, 27 May 2018 (UTC)

For Wikisource's purposes, we don't need a license for a simple transcription. We can use it as is, no matter what Taiwan's copyright laws say.--Prosfilaes (talk) 22:10, 27 May 2018 (UTC)

Index:Cricket, by WG Grace.djvu[edit]

Once the tables in this are proofread, this will be ready for validation, any takers for filling out the remaining tables which are largely a data entry task? ShakespeareFan00 (talk) 11:05, 29 May 2018 (UTC)

I've run into situations like this, where data that might have been highly interesting as data presents a major impediment to completion of an otherwise prose-oriented work. In this case, it seems to me that if people are truly interested in cricket data at this level, maybe a site analogous to has already imported the data into an actual database, or will. I am not sure that an HTML table is a meaningful or useful step toward that.
Would Wikisource editors consider it acceptable to simply include these pages as images, which would take vastly fewer person-hours (which could otherwise, I would think, be spent on more enjoyable and more useful tasks)? I kind of hope so, but I'm curious what others have to say. -Pete (talk) 03:03, 30 May 2018 (UTC)
(To be clear, if your preference is to enter them as data in this case, I have no quarrel with that, and I'll even try to pitch in a bit. My question is about the general standards and expectations.) -Pete (talk) 03:08, 30 May 2018 (UTC)
In the end I used {{aligned table}} for many of them, but a better way of doing tables like this (for possible later Wikidata use) is proabbly a longer disscussion.ShakespeareFan00 (talk)
Sorry if that was an unwelcome distraction. Now that I explore the work in more depth, I see that it's very stat-heavy and most of the work had already been done, so my question was clearly misplaced. -Pete (talk) 20:35, 30 May 2018 (UTC)
@Peteforsyth: For complex tables we agree with you, and it is a practice that has been undertaken. Stick in the image, mark it {{table missing}} and problematic, though it is a per work decision. Genealogies (which are a PITA to replicate) are somewhat similar, though we haven't been marking them as missing, just accept the image reproduction, and use good alt= description. — billinghurst sDrewth 23:20, 30 May 2018 (UTC)

Page break within {{dotted TOC line}} - ?[edit]

Some templates have ".../s" and ".../e" variants that make them work across page breaks. Not so with {{dotted TOC line}}. See the entry on the Book I/Chapter III entry here: The Early Indian Wars of Oregon How should I best deal with this? -Pete (talk) 02:58, 30 May 2018 (UTC)

@Peteforsyth: There are many pages in the TOC. This template easily gets burdened with template overload because of the dots. Better to shift to {{Dotted TOC page listing}}, which has more options, works across page breaks, and is more resistant to template overload. However, mobile view and epub are not good with dotted TOC templates. Hrishikes (talk) 03:06, 30 May 2018 (UTC)
With those templates they are managing <table> formatting, whereas the /s and /e templates are managing <div> formatting; so unfortunately not directly comparable; the consequences of html rules. — billinghurst sDrewth 23:29, 30 May 2018 (UTC)
@Peteforsyth: For the true believers(!) there was a whole slew of page-crossing-support {{dotted TOC page listing}} variants: see documentation here. 00:44, 31 May 2018 (UTC)

Unable to find the problem with this header template[edit]

Pasted this sample header and I am unable to find the problem. Could someone please look at it? — Ineuw talk 03:52, 2 June 2018 (UTC)

Yes check.svg Done Beeswaxcandle (talk) 03:54, 2 June 2018 (UTC)
Much thanks. History compare, what a wonderful tool. :-) — Ineuw talk 04:24, 2 June 2018 (UTC)

Okay head-scratcher as to why two templates behave differently....[edit]

{{cl-act-paragraph|split=none|layout=right|title=Example|text=An extened example paragraph. Blah blah blah blah blah etc...}}

{{cl-act-paragraph/x|split=none|layout=right|title=Example|text=An extened example paragraph. Blah blah blah blah blah etc...}}

Should BOTH generate the SAME output:



The second correctly generates the intended side-title, the direct (first call) doesn't suggesting a logic flaw somewhere.

I'm again a bit frustrated that templates fail to behave consistently despite there being little or NO difference between how they are coded.

ShakespeareFan00 (talk) 16:51, 2 June 2018 (UTC)

Logical flaw as suspected... Repaired manged to confuse the values on an equality test. We really really should have some people that can validate templates logically. :( ShakespeareFan00 (talk) 17:05, 2 June 2018 (UTC)

Template:cl-act-paragraph and family...[edit]

You make one experimental change and for some reasons it all works beautifully :)

However, it will at some point need someone to put {{sidenotes begin}} and {{sidenotes end}} (or some appropriate alternative) in the header and footer of all pages using it. :(

Can't win e'm all, but it's certainly working a LOT better than some previous approaches..

I'd like some comment on how this now much more powerful template could be improved further.

ShakespeareFan00 (talk) 19:53, 2 June 2018 (UTC)

We have talked about your (overly?) complex templates with a narrow focus, and you repeatedly bring this set here complaining about them. Nothing has changed. The rest of us simplify our templates, or use them in sets. — billinghurst sDrewth 02:02, 3 June 2018 (UTC)
Thank you the response, confirming your previous opnions on the subject. ShakespeareFan00 (talk) 20:29, 3 June 2018 (UTC)

{{cl-act-title}} which was the cause of most of the problems was converted to be a specific call to {{MarginNote}}, with the complexity inherent in it to do with handling the automatic position switching between PAGE namespace and a transclusion elsewhere. If using {{Outside L }},{{Outside L}}, {{Outside RL}}, in many instance extensive sidenotes overlapped to such an extent as to be unreadable, by comparison those doen with the current experimental version of cl-act-title, seem to adjust! This in effect solved the problem that lead to the original really complicated (and now abandoned approach with cl-act-paragraph.)

If the marginnote template was able to handle lrpage or rlpage values for it's margin parameter (i.e did the automated side switching between Page: and the mainspace transclusion), then in many many instances, cl-act-paragaph, as such could be replaced with direct calls to cl-act-title (given the changes). On an identical note the anchors generated automatically could be done using directly, with the section formatting being manually typed. Given that templates are essentially macro shorthands for repeated formatting calls I'm not sure what advanatges would be gained by coding all the formatting manually. This leaves the remaining functionality being the indentation of paragaphs based on their level. Currently this is done by using a call to {{p}} to set the margins. The variants {{left/s}} and {{left/e}} could be used alternately for this (but I am thinking these are DIV rather than p based). Your feedback is appreciated, and you've given me some things to examine. ShakespeareFan00 (talk) 08:37, 3 June 2018 (UTC)

Hmm, per User:ShakespeareFan00/Sidenotes testing It seems that {{MarginNote}} is lacking an ability to tweak it for the use of margin adjusting templates where it's called from INSIDE them. Back to the drawing board sadly :( ShakespeareFan00 (talk) 08:51, 3 June 2018 (UTC)
Subsequently I've reverted back to the version of {{cl-act-title}} that uses a sidenoted reference, It's STILL wrong,and broken etc, but at least it nominally works consistently.
All of this of course moot, if someone (re)-implemented the sidenote functionality so that the overlaps didn't occur, based on apathy here on Wikisource, I don't see that occurring any time soon.ShakespeareFan00 (talk) 08:53, 3 June 2018 (UTC)
A point that I think you might be overlooking is that the original printed pages are a set width and as a result the marginal notes always begin and end at the same points across the page and there is space vertically to have them not overlap. In the online resetting that you are attempting, the page width is variable depending upon the screen size of the reader—anything from a 42″ flat screen TV to a 3″ cell phone. Instead of trying to slavishly reproduce the exact printed page setting, I suggest you focus on how to represent the text with its notes in a usable on-screen way. For many of the Acts of Parliament, a lot of the marginal notes are navigational (this is where subsection iii.a.iv.b starts)—you probably don't need to reproduce these at all. Another large set are mentions of other Acts—do these as in-text wikilinks, not as marginal notes. A third group are essentially references or footnotes—change these to standard Wikisource style footnotes. This should leave you with a lot fewer marginal notes to decide what to do with. Or, to put it in another way, what's the simplest way of making the text of the Acts available and usable to a wide range of readers. Beeswaxcandle (talk) 09:56, 3 June 2018 (UTC)
Okay what we have in the originals are
  • Section headers ( I.E titles that are defacto section summaries this is what most level 1 side titles are, and these are explictly coded for.)
  • Citations (I.E a cross reference to another document or statute in this instance.
  • Explanatory notes.

The main version translates ALL but the titles to level 1 paragraphs into sidenoted references for clarity.

The version in sandbox, uses Margin notes to try and approximate the formatting of the older fixed width versions..

One minor change that would make simplification a lot easier to update {{Short-title/c0}} which was written to use some code in the Module: Short-title which called {{cl-act-title}} directly with a fixed format, that code can probbaly be removed or the usages replaced wth something more conventional. ShakespeareFan00 (talk) 11:33, 3 June 2018 (UTC)

A progress update - {{short-title/c0}} replaced with references.

ShakespeareFan00 (talk) 13:39, 3 June 2018 (UTC)

Unless there's some kind of concrete proposal in a week or so, I'll consider 'phasing out' this template family entirely, as it seems a "massive" breakage or conversion is the only way this issue is going to get the attention it needs. For a template that was on the cusp of being very useful, that would be a shame. ShakespeareFan00 (talk) 21:30, 3 June 2018 (UTC)

Plenty of people are prepared to say 'too complex' , but it seems bar one approach (whose logic is complex to my understanding) it seems there's an (inadvertent perhaps) apathy to actually provide suggestions on HOW to simplify this template family or actually resolve the issue. So far this template family
  • has been converted down from a DIV based approach to a P based approach because of the parser issues
  • has had sidetitles converted into references because that's apparently "more compatible" with dynamic layouts
  • has had multiple inline sidetitles moved to the top of a section (albiet in a currently cryptic way)
  • has had certain forms for different levels ( i.e paramater sets) coded in to specfic templates as shorthands...


So I will ask a blunt question. What will be acceptable to certain contributors such that this template family is

  • Understandable
  • Maintainable
  • Compatible


Sorry for sounding exasperated, but thats the way I feel right now.

ShakespeareFan00 (talk) 21:49, 3 June 2018 (UTC)

Okay... time to take a wiki-break, before I go beserk.. ShakespeareFan00 (talk) 23:21, 3 June 2018 (UTC)

Use of text attached to plates[edit]

Esme Shepherd (talk) 10:30, 5 June 2018 (UTC) This query arises probably because I do not know the full technicalities of attaching text to plates. I have copied from example. So, on The Improvisatrice.pdf/4 I have the artist and engraver below the plate using smaller block, float right, etc and this works fine.

The problem is that there is a larger plate on The Improvisatrice.pdf/5 and if I put these pages together in wikisource, the plate texts on page 4 are widened to fit this wider image and go beyond the plate, which is not what is required.

I have therefore been forced to split the opening title pages into two to prevent this happening. I cannot find anything about this in the help pages, so some guidance would be appreciated.

@Esme Shepherd: I have done some little tinkering. Please check the first title page now, whether this is ok. Hrishikes (talk) 11:53, 5 June 2018 (UTC)

Esme Shepherd (talk) 14:46, 5 June 2018 (UTC) Thank you. That seems to be fine. I have adjusted the links and the second title page is now superfluous.

Page aware floats[edit]

I appreciate that we have had some issues, but I wanted your view on two very quick and probably horrendusly broken Page: aware versions of the float template I created earlier being {{float LR}} and {{float RL}} respectively. If the function can be implemented in the main templates more cleanly it would of course be desirable.

However, given the complexity of coding Page: aware templates , I'm wondering if it's just easier to assume one layout regardless any scans notwithstanding. ShakespeareFan00 (talk) 01:01, 6 June 2018 (UTC)

See also my other posting in the main Scriptorum on a related issue. ShakespeareFan00 (talk) 01:01, 6 June 2018 (UTC)
Why? How about you use one of the two the existing sidenotes systems rather than create a third f***ed system that we will need to repair. We haven't managed to get it right for those two, so why do you think that any of us can build something better on your system? — billinghurst sDrewth 05:56, 6 June 2018 (UTC)
What that reads to me is, "because I am unwilling or unable to debug or consider certain esoeteric template interactions that might sometimes happen, I am going to insist that existing templates (which may be broken in other ways) be used, because it's simpler for me to understand only those known issues and interactions ." I'm sorry to be blunt and call you on that, but that's the way it's reading to me. ShakespeareFan00 (talk) 06:54, 6 June 2018 (UTC)
Which on re-considering is a actually good approach. So a template that even with the best of intentions is never goin to work should be ditched. I'll get started soon. ShakespeareFan00 (talk) 19:28, 6 June 2018 (UTC)
For what it's worth - User:ShakespeareFan00/Sidenotes testing contains test cases for various forms of an interaction I was trying to resolve.
Also which ONE of "existing" sidenote systems should be used to the exclusion of others be used then? So that all uses can be conformed and the f***d up (your words) systems can be removed? In the absence of any kind of reasonable response on this, as I've said elsewhere I am inclined to use one form of sidenote (i.e {{Outside L}} or {{Outside R}} universally in any given work) and ignore the paged layout of the scans entirely. ShakespeareFan00 (talk) 07:00, 6 June 2018 (UTC)

[Unindnented] All of which is a distraction from the other primary reason these were created, to have a Page aware version of the float templates more generally, which is a SEPERATE concern from the side-titles issue. In any event I've userfied them as they need more extensive testing.ShakespeareFan00 (talk) 09:47, 6 June 2018 (UTC)

Is there a lag on the db version used for insource searches?[edit]

I'd been updating the sidenotes to use one nominally consistent layout, but seem to be seeing an older version of the search result which is NOT reflecting what's actually in the pages when I check whats ACTUALLY present in the edit form. Either that or I need my eyes testing as I was sure on some of the pages continuing to show up I had updated them.07:27, 6 June 2018 (UTC)

Partially you are asking the wrong question. Yes there is an effective delay on repeated queries of this nature but the reason has nothing to do with insource searches. Assuming the change you wish to examine was made to an included item (typically a template; but depending upon the search namespace selected could be all of Page: space in the worst case) the update has to propagate via mediawiki's internal job queue and that can be S-L-O-W as it can depend on the edit activity going on anywhere else on the system!
The bad aspect of using insource is the high load it places on mediawiki. Thus it is good practice to perform an initial search only to locate items of interest to any given case and then examine the results ideally using another method. 08:24, 6 June 2018 (UTC)
Thank you for the technical explanation. ShakespeareFan00 (talk) 09:24, 6 June 2018 (UTC)

differences between templates[edit]

What are the specific differences technically and functionally between templates {{block center}} and {{center block}}? Particularly with regard to use with formatting poetry. Layman's terms please :) Thanks, Londonjackbooks (talk) 04:18, 7 June 2018 (UTC)

Tricky, custom footnotes[edit]

Looking for creative footnote solutions for a page that uses double daggers and pilcrows instead of sequential footnote numbering. {{ref}}+{{note}} format lets me specify the footnote icon but won't work well when compiled into chapter view. Ideally there would be some sort of template that took {{ref}} as it is but pulled from a list-defined {{reflist}}... Czar (talk) 02:41, 8 June 2018 (UTC)

The house style here is to use incremental numbers instead of those special symbols. There are not two parallel sets of footnotes there, one using numbers and the other symbols; only one symbol-using set is present. So there is no reason to deviate from the house style. Hrishikes (talk) 02:51, 8 June 2018 (UTC)
@Czar: Symbols don't work when we transclude pages together and become endnotes for a chapter, they lose their uniqueness. See Help:Footnotes and endnotes. — billinghurst sDrewth 08:33, 8 June 2018 (UTC)
@Hrishikes, thanks—is this stated somewhere, perhaps in the Wikisource:Style guide? Adding clarification to Help:Footnotes and endnotes, which I had read, and would like to link the proper section for more details Czar (talk) 09:54, 8 June 2018 (UTC)
@Czar: Please see the last line of the first section of Help:Footnotes and endnotes --- If the work only uses special symbols in one symbol set, the house style is to instead use incremental numbers. -- Hrishikes (talk) 10:19, 8 June 2018 (UTC)
@Hrishikes, that was me. I was hoping to link it to the actual style advice. Czar (talk) 10:27, 8 June 2018 (UTC)
@Czar: Thanks, but the help page concentrates on one mode only, that of the number format. The symbol method is given in the Ref and Note templates subsection, with the preceding line: The preceding instructions (= number method) are best practice for producing footnotes on Wikisource. However, in some rare cases it may be more appropriate to use one of the following methods instead. You can see an example of symbolical and numerical footnotes at the end of this chapter in a sister site. I did it because there were two sets of footnotes. Hrishikes (talk) 10:51, 8 June 2018 (UTC)
Sounds good. Just wanted to check. Thanks! Czar (talk) 13:43, 8 June 2018 (UTC)

Help adapting to new file?[edit]

This scan was missing a couple of pages. I found a new version and uploaded it, and I've been trying to move then handful of pages that have already been transcribed to the new pagination scheme...but I'm only making the mess worse. I'm thinking that not only the index, but also each individual page, probably needs to have its cache purged before it will display the proper page image on each page...but even with manual purging, I'm getting inconsistent results. Any suggestions on how to fix it? At worst, maybe an admin could just delete everything so I can start's not like a ton of work has been done. But if there's a way to preserve what work has been done, obviously that would be highly preferred. -Pete (talk) 04:49, 9 June 2018 (UTC)

@Peteforsyth: The best way would be to download the djvu, add the missing pages, reupload it, and move the downstream pages to new locations. That's the way I do it, instead of uploading a new file in a new location. Hrishikes (talk) 07:34, 9 June 2018 (UTC)
@Hrishikes: Thank you, that makes sense, and I find guidance on how to do it here: Help:DjVu_files#Inserting_a_new_pages_(e.g._a_placeholder) I have found the DjVuLibre package for Linux, so I think I'm able to add pages as suggested. I will also need to remove pages like this one. I am somewhat daunted by the project, but will see what I can do! -Pete (talk) 18:12, 9 June 2018 (UTC)
Update...I have figured out how to use djvm to delete pages. I practiced on a simpler file -- in this one, every even page was blank (presumably, it was printed single-sided, and scanned as though it was double-sided). As an added bonus, it forced me to finally learn to write a simple shell script (since djvm doesn't natively support deleting more than one page at a time). Thanks for pointing me in this direction, I've learned some useful stuff today. -Pete (talk) 20:46, 9 June 2018 (UTC)
All set now, and if anybody wants help in a similar situation, feel free to ask...I think I'm reasonably conversant in manipulating DjVu files now. -Pete (talk) 22:13, 12 June 2018 (UTC)

Yes check.svg Done

Newline between pages[edit]

Can anyone identify why the page Page:EB1911 - Volume 04.djvu/131 is starting a paragraph when transcluded into the main namespace at 1911 Encyclopædia Britannica/Boetius, Anicius Manlius Severinus? No {{nop}}, no div templates beginning or ending either page. Prosody (talk) 01:15, 10 June 2018 (UTC)

It looks like {{EB1911 fine print/s}} does contain div tags, for what that's worth. -Pete (talk) 01:25, 10 June 2018 (UTC)
Because the starts and ends are in the header and footer respectively, I was under the impression that it would produce a single div tag when transcluded, which seems to be the case. That was in fact what had brought this page to my attention, I made the change from {{EB1911 fine print}} to the /s and /e equivalents since earlier this year {{EB1911 fine print}} was changed to div based. But still there's an extra p tag. Prosody (talk) 01:38, 10 June 2018 (UTC)
Hmm. Sorry, that's about all I've's all pretty mysterious to me. -01:52, 10 June 2018 (UTC)
{{EB1911 fine print/s}} fixed.— Mpaa (talk) 08:20, 10 June 2018 (UTC)

Weird formatting in original[edit]

I do believe this author anticipated Wikisource by more than 100 years, and adopted a formatting style with the specific intent of confounding me personally. Any ideas on how to solve this riddle?

Essays in Historical Criticism/The Legend of Marcus Whitman/Part 1

On pages 9-15 of this work, the author has chosen to quote the narratives of Spalding and Gray in parallel: each narrative runs from one page to the next, with a horizontal rule separating them. On the final few pages, Gray's narrative has run its course, and only Spalding's continues, and then the author's own narrative resumes. (I imagine this was done to highlight similarities and/or contrasting points in each, but I haven't yet read it closely enough to confirm.)

I think perhaps the best thing to do is to simply preserve it as-is, with an additional rule between "includeonly" tags to indicate the separation between the narratives at the page breaks. But, I worry that this will be confusing to a reader looking at the text in a web browser, as pages aren't really perceived the same way on a screen.

The other alternative I see would be to reunite each narrative using section tags...but perhaps that would override an intentional act of putting certain parts of each narrative side-by-side.

Maybe something fancy with columns...? I'm at a loss. -Pete (talk) 18:59, 12 June 2018 (UTC)

I would use sections to present one narrative, then the other. As long as sections are marked on the pages, it will be easy to transclude them separately. It doesn't look as though side-by-side is done very well in the original, as I merely find it confusing. --EncycloPetey (talk) 20:21, 12 June 2018 (UTC)
I'm eventually going to run into the same issue with some volumes of Ruffhead, albiet having to transclude from another Wikisource as well. The split narrative idea works for me. ShakespeareFan00 (talk) 22:06, 12 June 2018 (UTC)
OK, thanks for the tip. I think I got it: Essays in Historical Criticism/The Legend of Marcus Whitman/Part 1
I noticed that when the main text resumes, it refers to "this narrative." Since it immediately follows the Spalding narrative in the original (but the Gray narrative in the transcription), I added a Wikisource user annotation. Any feedback, or changes I should make? (Should I find a way to have the footnotes for the narratives occur within the narrative sections, for instance?) -Pete (talk) 14:27, 13 June 2018 (UTC)

Converting Jp2 files for upload to Commons[edit]

I have recently been converting JP2 files by hand (previously, I relied exclusively on @Hesperian:'s bot to pull them from IA and convert them). My process is workable, but cumbersome, and I'm hoping for tips on how to improve it.

I prefer to work in Linux, though I'm happy to fire up Windows (or possibly even an old Mac) if it's worthwhile for a task like this. Here's what I've found:

  1. I browse through IA and download all relevant JP2s in a work
  2. One by one, I use the command line tool "img2pdf" to convert the JP2's to PDF files
  3. I open the PDFs, and one by one I right-click and save the PNG contained within
  4. I upload the individual files to Commons.

It feels like a pretty inefficient process, but I haven't been able to streamline it better than what I described. Do you have any suggestions? (If memory serves, I believe I've seen @Ineuw: commenting about converting JP2's in the past...pinging in case you have ideas.) -Pete (talk) 20:39, 13 June 2018 (UTC)

Jpeg2ktopam, from NetPBM, can convert them to PNM files, as can opj_decompress from libopenjp2. Convert from imagemagick can turn those files (or least the PNGs they're converted to) into a large PDF file. I'm not sure what you need.--Prosfilaes (talk) 20:51, 13 June 2018 (UTC)
Thanks. Sorry if my overall goal wasn't clear -- I'm trying to losslessly convert files like those listed here (for instance, this one) into PNG files for upload to Commons.
I haven't heard of PNM files, and from a quick look, I'm not sure whether or how it helps to convert into PNM files. (Perhaps it would help if I could batch convert PNM's into PNG's more easily than batch extracting PNG's from PDF's, though?) -Pete (talk) 21:23, 13 June 2018 (UTC)
I would definitely look at imagemagick. It can do just about any kind of command-line image processing you can imagine wanting to do. Your format conversion task is as simple as "magick image.jp2 image.png". Hesperian 02:17, 14 June 2018 (UTC)
@Peteforsyth: Like you, I also download jp2 from IA in case of images where HD version is preferable. In that case I download the jp2 folder. I open the files with Irfanview and save as jpg or png. If I want good djvu, I convert the jp2 folder to djvu. Hope this helps. Hrishikes (talk) 02:34, 14 June 2018 (UTC)
As others have stated, the imagemagick software would be the easiest via the command: mogrify -format png *.jp2 However it seems Ubuntu's version of imagemagick isn't compiled with JP2 support. So alternatively you could install libopenjp2-tools and use the following command to convert a whole folder: opj_decompress -ImgDir . -OutFor PNG --Riley AJ (talk) 04:12, 14 June 2018 (UTC)
If you are Windows,nothing beats Irfanview for batch conversion from one format to another. Also, you can upload images to the Commons in batches of ~50 images at a time if the source is the same. In Linux I haven't done any batch image preparation yet, and my experience with Wine was less than successful. In general they crash. — Ineuw talk 15:01, 14 June 2018 (UTC)

What a great collection of suggestions -- thank you all. I'll experiment with them this weekend. @Riley AJ: your observation is especially helpful, I was wondering why some of the imagemagick-bsed suggestions I was finding online were not working on my Ubuntu Maté-based system. -Pete (talk) 16:33, 14 June 2018 (UTC)

Update: @Riley AJ:'s suggestion is exactly what I was after. I'll try out Irfanview later, but for now I'm happy with this. Thanks so much! As a minor followup, I'm curious if anybody understands why IA has two or more directories of JP2 files for each work? I've generally just gone to the biggest one, on the assumption that it's probably the highest-quality collection; but I wonder if I'm missing something. -Pete (talk) 01:49, 16 June 2018 (UTC)
According to this IA blog post, the file <identifier>_files.xml contains a little bit of information about what each file is within the directory. Including, whether the file comes from the original uploader (marked with source="original") or if it's a processed file (source="derivative"). So for "riverofwestlifea00vict", the original scans (or at least original untouched upload) would be the "riverofwestlifea00vict_orig_jp2.tar" file. Which I assume is why it has "orig" in the file name. At least that's all I could find about it. --Riley AJ (talk) 05:56, 16 June 2018 (UTC)

Adding a poem[edit]

I have been advised to use Wikimedia to be a reference for a poem about an historic mulberry tree. I am trying to create a page on Wikimedia about the tree:

Wikimedia is daunting. But I am sure the process will be simple. Can someone experienced please assist?

David Wilson

Hello. May I suggest adding the poem here to Wikisource—in a manner similar to this poem—also noting the addition of source information on the corresponding Talk page. Then you can link to the poem here at Wikisource from the Wikipedia article rather than place the poem over there in its entirety. Feel free to wait for other feedback on this; other editors may have different approaches. Londonjackbooks (talk) 07:48, 17 June 2018 (UTC)
Even better, if you are able to do so: you can follow our Beginner's guide to adding texts and make a copy directly from the original newspaper. I have already done the first two steps. The text of the poem is here. You don't have to worry about the other articles on that page. —Beleg Tâl (talk) 12:25, 17 June 2018 (UTC)

An Anglo-Chinese Vocabulary of the Ningpo Dialect (PDF file) is not properly displayed ?[edit]

Hi there,

I was just wondering whether it was my computer or my network : Index:An Anglo-Chinese Vocabulary of the Ningpo Dialect.pdf scans are not displayed on the Page: namespace. I can't even see the preview of the cover on the index page. Is it related to the PDF type of the file and should we upload a DJVU file of this work instead ? Assassas77 (talk) 08:13, 18 June 2018 (UTC)

Not just you. Thumbnails have been failing to generate for me also, for a few days now. —Beleg Tâl (talk) 10:19, 18 June 2018 (UTC)
@Assassas77, @Beleg Tâl: Please report your issues at phab:T196961 -- Hrishikes (talk) 11:50, 18 June 2018 (UTC)
@Assassas77: File corrected, please check. Hrishikes (talk) 17:40, 18 June 2018 (UTC)
It works for me but the pictures seems a bit rough (Page 12) compared to the file on the Internet Archive [1] I tried using the ia-upload tool to re-upload it but the script says « Upload failed ». Assassas77 (talk) 19:15, 18 June 2018 (UTC)
@Assassas77:, yes, the IA file has very high resolution (and compression), that is the problem. I reduced both, that's why the file is visible on display now. You have to decide between the two evils: low-resolution display or no display at all. If you want the IA version of the file, just revert my upload. Hrishikes (talk) 01:17, 19 June 2018 (UTC)
@Hrishikes: I reuploaded the DJVU file directly from IA using ia-upload. The new index is here : Index:An Anglo-Chinese vocabulary of the Ningpo dialect.djvu. I'll ask an administrator to do the transfer (diff). Assassas77 (talk) 20:33, 19 June 2018 (UTC)

Hyphen Problem[edit]

I have a doubt regarding proofreading. Consider the example of the word disambiguation. In the book, it is given as disamb-iguation where the first word before hyphen is in line 3 and the word after hyphen is in line 4. While proofreading, do I need to remove the hyphen symbol? Adithyak1997 (talk) 15:04, 18 June 2018 (UTC)

Yes. Hrishikes (talk) 15:35, 18 June 2018 (UTC)
Yes, you can remove the hyphen symbol if the separation is between lines in the same page.

If the hyphen is at the end of the page and the rest of the word is on the next page, you'll have to use {{hws}} and {{hwe}}.
  • On the first page, you would write {{hws|disamb|disambiguation}}.
  • On the second page, you would write {{hwe|iguation|disambiguation}}
For more informations, read Help:Formatting conventions. Assassas77 (talk) 15:40, 18 June 2018 (UTC)

Aligned table - nofooter param on it's own is useless...[edit]

See: User:ShakespeareFan00/Aligned table test, where I was testing this in anticipation of using it more widely..

By comparison some minimal examples:-




The resultant output being (from Special:Expand Templates) .

<div class="mw-parser-output"><table>



I fail to see why these minimal examples would work ( the issue of suprious row generation aside), when the use of a table end syntax with the aligned table template (and a nofooter=yes) doesn't. Given this I also opened task T197712. When I follow the nominal documentation there's a reasonable expectation that things will behave in a consistent manner. The fact that simple things like this don't always seem to behave as the documentation would suggested is an increasing frustration, especially given the lack of time, expertise or patience amongst Wikisource contributors to continually work around Mediawiki limitations. Sometimes there are things that needed to be fixed at a deeper level. ShakespeareFan00 (talk) 20:13, 19 June 2018 (UTC)

I think you need to make peace with parser and templates and leave them alone ...— Mpaa (talk) 22:01, 19 June 2018 (UTC)
To me your response is understandable but it would be nice sometimes to see progress on some long-standing issues being fixed, some of the ones I encounter have been there for over a decade :( 22:14, 19 June 2018 (UTC)
@ShakespeareFan00: If in doubt, try not to mix your syntax styles. Mediawiki is not (despite propaganda implying otherwise) a comprehensive analyser of its "language." When there is no introducing {| present the parser never even recognises any of the other "table" constructs |, |+, |-, or |} and they are passed through unchanged with results as you have already experienced.
As for mixing styles? Try this substitution in your Aligned table test trial and I think you will find it works as expected:
1 {{aligned table|cols=3|style=border-collapse: collapse; width: 100%;
2 |row1header=yes
3 |nofooter=yes
4 |Session and Chapter.
5 |Title.
6 |Short Title.}}
7 {{aligned table|noheader=yes|nobody=yes}}
In other words: if you choose to use a LUA-based script to build a portion of the table, then it is just common sense (if you trust the programmer!) to use the same code to complete the structure… 11:03, 20 June 2018 (UTC)
Thank you.. , Having an {{aligned table/suspend}} {{aligned table/resume}} and {{aligned table/ribbon}} (or failing that a rowXribbon=yes option in the module.) would also be helpful. The purose of the first two should be obvious from the names. The last one was intended to add content that would only be applicable in Page namespace (such as column headers) , in effect a wrapper so as to avoid an ugly construction of <noinclude>... content ...</noinclude> logic.. Is there an issue tracker for suggested template improvements?ShakespeareFan00 (talk) 11:07, 20 June 2018 (UTC)
{{aligned table/suspend}} ≡ {{aligned table|nofooter=yes}}; {{aligned table/resume}} ≡ {{aligned table|noheader=yes}}: inelegant but functional. {{aligned table}} does not have any kind of support for the "ribbon" concept (presumably you are referring to {{TOCstyle}}, or has the topic wandered off-track? 05:47, 21 June 2018 (UTC)
I might ask about 'ribbon' support on the template's talk page, see Page:Short Titles Act 1896(ukpga 18960014 en).pdf/2 for an example of where an aligned table needs a 'ribbon' row option, The current approach works, but isn't as clean as it could be.ShakespeareFan00 (talk) 08:46, 21 June 2018 (UTC)
Note to the IP contributor, Can you also leave your related comments in the phab ticket, which I am about to close? ShakespeareFan00 (talk) 11:16, 20 June 2018 (UTC)
I was too slow. It appears you've closed the ticket already. I'm not too clear what you wanted to add in any case? 05:47, 21 June 2018 (UTC)
Only to note the workaround. ShakespeareFan00 (talk) 08:46, 21 June 2018 (UTC)
A different issue here - Page:DfT-circular-01-2016.pdf/81, The header can't use aligned table as it's got a rowspan, on Page:DfT-circular-01-2016.pdf/80 where the header/footer handling doesn't affect the flow the two style can be mixed. Here it generated stripped tag errors.

If aligned table is incompatible with multipaging in this way it's documentation ought to say so. And once again I've encountered a limitation that nobody seems to have the patience or expertise to implement a long term repair for, as opposed to "clever-quick fixes." :( ShakespeareFan00 (talk) 14:40, 21 June 2018 (UTC)

Okay time to take some deep breaths and re-code.. ShakespeareFan00 (talk) 18:55, 21 June 2018 (UTC)
During one of those deep breaths please recall the ancestry of {{aligned table}} was that it was copied from a version developed in the wikipedia environment with but minimal modification to handle a limited form of multipaging. The wikipedia version has evolved in another direction in the nearly three years since the split and may have improvements (but most likely is not remotely usable in the wikisource environment.) In other words {{aligned table}} is not a brilliant choice for use here if anything else locally supported can do a given job.

None of this stuff is "official" mediawiki software so raising a phabricator ticket will gain exactly no sympathy and will be a pretty much wasted effort only costing the requester reputation without reward. 07:51, 22 June 2018 (UTC)

In other words the "the template/module is not (yet) compatible with the Proofread page environment". The next question I was going to ask would be if it was possible to have some kind of 'localised' issue tracker? I was going to put a note on the talk page for the Module concerning the headers issue. 08:20, 22 June 2018 (UTC)

Arbitary break and rant about lack of flexibility in certain markup and non-Page: awareness of some templates.[edit]

Currently {{aligned table}} isn't widely used but on a number of these, changing already validated pages back to having "one-style" would incur a performance hit (as the use of multiple calls to {{ts}} obviously has more overhead than a single call to {{aligned table}}, perhaps you'd be able to review and come up with solutions to specific instances?.
Some of the table with mixed-styles are because of colspans or rowspans in the header rows, something which is a known (and now documented) limitation with {{aligned table.}}. This is not easily solvable (given that the current mixed style approach isn't advised as you explained earlier) other than manually placing hundereds of calls directly into the table, which as I've said would have a notable performance hit. {{ts}} as currently coded is literately a call to massive SWITCH statement, which for several hundred calls will eventually hit parser limits.
Other uses of {{aligned table}} are to generate ToC or INDEX pages, Whilst some of this could reasonably be re-coded to use TOCstyle, or lists, both of those have their own alternative limitations. In respect of It wasn't for example possible to use wiki-syntax for lists directly on this Page:Report of a Tour Through the Bengal Provinces of Patna, Gaya, Mongir and Bhagalpur; The Santal Parganas, Manbhum, Singhbhum and Birbhum; Bankura, Raniganj, Bardwan and Hughli in 1872-73.djvu/4 as there was no (apparent) means in direct wiki-syntax to change the style of list index shown. Here it was fortunate that only a single page had to be re-coded as {{ordered list}} is another template that as far I as I know doesn't (yet) have "paged" variants to generate appropriately different numbering via a start=<number> value to the opening <OL> tag in Page: namespace.
Good, {{ordered list}} DOES have support for differing start values, not necessarily for the split-list items mentioned below though. ShakespeareFan00 (talk) 09:49, 22 June 2018 (UTC)
(Aside) I've also sometimes run into instance where a table element, is directly split over a "page-break", these are not straightforward to code for and usally need the relevant portions of the text to be relocated which complicates proofreading.
(Aside) I've also come across 'list' elements that span over page-breaks, something which current wiki-syntax for lists can't easily cope with, whilst it is possible to code for this using HTML5 directly inline, This is at best a 'clever-fix' to workaround a limitation which should be fixed in Mediawiki itself.
Whilst it's helpful to know that there are limitations, It would be nice if sometimes there were efforts made to find long-term stable solutions, rather than having to figure out 'clever-fix' ones that rely on playing 'hunt the behaviour' when other things change (such as the parser migration).

It's disappointing that I get the impression of an apathy to resolve many many of these concerns and limitations. ShakespeareFan00 (talk) 09:30, 22 June 2018 (UTC)

DocBook XML[edit]

Are there any tools/scripts for converting a book in DocBook XML format for upload on Wikisource? Czar (talk) 23:36, 19 June 2018 (UTC)

@Czar: Pandoc can convert from Docbook to wikitext. It wouldn't upload, you'd have to do that manually. Sam Wilson 00:48, 20 June 2018 (UTC)
Pandoc left lots of artifacts but I suppose it did the job. I cleaned it up and uploaded this major reference work. Open to ideas on how best to format, given that the XML didn't include direction on layout. I opened a thread at Talk: Literary Research Guide. (not watching, please {{ping}}) czar 17:17, 24 June 2018 (UTC)

Formatting lists on Page:LostApocryphaOfTheOldTestamentMRJames.djvu/17[edit]

Here a template is used to create a list, with a concern , the list style type given doesn't actually exist as CSS style type! meaning that whilst the intent is good, the formatting intended doesn't actually appear to be generated. A suggestion on how to possibly do this in a way that is "compatible" long term would be appreciated. ShakespeareFan00 (talk) 11:31, 22 June 2018 (UTC)