Template talk:DNB errata

Development

Pardon (I could not resist). The recent "incorporation" of Template:Page break into DNB Errata may not be optimal for what you are trying to accomplish in the end nor in light of the move to the preferred <pages_index /> for the main namespace over the "older" methods of transclusion. In short, we've squeezed everything into what was merely intended to toggle the output's short preset prefix via applying/denying/assigning the label= parameter. Doing so resulted in.....

• label=blank [making the first un-named parameter the only resulting output, if present]
• removal of label= completely [making the default prefix page + the first un-named parameter, if present, the output]
• override both of the above with label=some_other_text [making some other text + the first un-named param., if present, the output]

In all three cases, it is the first unamed parameter that is actually concerned with reflecting the transcluded page or sequence of page(s) in order to mirror the listing laid out in the Index's pagelist or in order to circumvent the list.

Either way, your DNB errata's {{{page}}}&/or{{{page2}}} should be replacing that first unnamed parameter at the end of our new default prefix, Dictionary of National Biography, Errata (1904), p. both of which sitting on top of a new default suffix on a new line titled {{{note}}} with a default value of N.B.— f.e. stands for from end and l.l. for last line

By "using" the Page break template in it's current out-dated application, we are generating pagebreaks with spans that may look and function just fine but make a mess of the id pointers back to the Page: namespace that "real" transcluded pages (& sections?) rely on -- with outputs such as this 'underneath':

<span class="center pagebreak" id=".27.27Dictionary_of_National_Biography.27.27.2C_Errata_.281904.29.2C_p..5B.5BPage:Dictionary_of_National_Biography._Errata_.281904.29.djvu.2F60.7C50.5D.5D.E2.80.93.5B.5BPage:Dictionary_of_National_Biography._Errata_.281904.29.djvu.2F61.7C51.5D.5D" style="clear: both; display: block; font-size: smaller; border-top:1px dashed; border-color:#999999">Dictionary of National Biography, Errata (1904), p.<a href="/wiki/Page:Dictionary_of_National_Biography._Errata_(1904).djvu/60" title="Page:Dictionary of National Biography. Errata (1904).djvu/60">50</a>–<a href="/wiki/Page:Dictionary_of_National_Biography._Errata_(1904).djvu/61" title="Page:Dictionary of National Biography. Errata (1904).djvu/61">51</a>
N.B.— f.e. stands for from end and l.l. for last line</span>

Instead of an output something like this

<span class="center pagebreak" id="50" style="clear:both; display:block; font-size:smaller; border-top:1px dashed; border-color:#999">Dictionary of National Biography, Errata (1904), p.<a href="/wiki/Page:Dictionary_of_National_Biography._Errata_(1904).djvu/60" title="Page:Dictionary of National Biography. Errata (1904).djvu/60">50</a>–<a href="/wiki/Page:Dictionary_of_National_Biography._Errata_(1904).djvu/61" title="Page:Dictionary of National Biography. Errata (1904).djvu/61">51</a>
N.B.— f.e. stands for from end and l.l. for last line</span>

I've ripped the needed tidbit from Template:Page break and incorporated it into Sandbox2 with the same basic function and output as before but the span id is now set to {{{article}}} to insure no conflict with the pagenum class will occur.

{{DNB errata/sandbox2|article=Campbell, Frederick W.|page=60|page2=61}}

Dictionary of National Biography, Errata (1904), p.50–51
N.B.— f.e. stands for from end and l.l. for last line

Page

Col.

Line

I can "gamble" a bit further and have the id defined "properly" with the actual page of the 1904 errata note ( {{{page}}} - 10 ) but that runs the very, very, very slim risk of conflicting with pagenum spans if by chance the same page number is being transcluded from any particular DNB Volume at the same time to the main-space. — George Orwell III (talk) 00:25, 7 February 2011 (UTC)

I’m not that familiar with this aspect of the coding. So let me try and build a bridge to where I believe you are coming from…I observed last night there were difficulties with some recently formatted pages making use of {{DNB errata}}. When I recognized I might be unintentionally interfering, I redirected to other tasks. You are familiar with the conversation from last night. When I re-entered the conversation, it appeared that all was resolved. From the reader's point of view, this template functions as desired (almost). I’ll come back to that in a moment…

At this time, If I understand you correctly, you have resolved the programming side to stop generating harmful interference (page breaks elsewhere). If that is the case…thank-you…if not what can I do to help resolve this for you…

No worries - Application of this template will be fixed in the sense that the needed part(s) of the Page break template have been incorporated directly into DNB errata rather than calling Page break everytime a mainspace DNB article happens to also have errata info without a.) all that junk from happening as described above; b.) the mis-use of what was merely intended for actual page breaks within a series contiguous pages; not as a "footer" of sorts here; and c.) allowing Page break to be restored to its previous simplistic state without the recently added parameters that only serve the DNB project pages and nothing else.

All it additionally needs is if you want to link directly back to the <section>'s errata info line(s) much like using an anchor would, leave things alone and just link back to the page in the Page: namespace itself or some other function/assignment. Not much more effort needed in short - plus of course swaping in the fix for the current template once Billinghurst OK's my work. — George Orwell III (talk) 03:08, 7 February 2011 (UTC)

Back to the glitches…

I am aware of only two issues requiring additional work on this template. For some unknown reason this template is not able to transclude from Errata djvu page 171 for Theophilus Howard page (all other Errata containing articles in volumes 23 & 28 it functions as intended). It also turns out that the page numbering in the 1904 Errata volume is not always fixed at a specific offset of ten, it begins to drop lower starting just after volume 28. I have a direct link from Theophilus back to that errata page as an interim fix during troubleshooting. JamAKiska (talk) 02:04, 7 February 2011 (UTC)

Well the first issue was nothing more than typo in <pages index> (& you're welcome - I got your note) but it does illustrate the overall point being made far above (though the specific fix was totally unrelated to all this).

To summarize the point: to always strive to remain consistent and uniform "underneath" in the resulting coding while at the same time achieving whatever it is you want to accomplish by applying one template or another. Not really all that big of a problem on WS (yet) but your typo did show how one thing being applied can affect another thing also being applied's functionality if improperly applied at the same time :-)

The second issue with the coming offsets shrinking shouldn't be too hard to solve - but I'd rather get past this hurdle first before even approaching the specifics on that problem if it's OK with you folks for now — George Orwell III (talk) 02:58, 7 February 2011 (UTC)

Good points. When you get to the page numbering…let me know so I can learn a little more in this direction. One Quick question please…I have been using the line break on occasion when out of other options. How would I discover if that is a source of other problems in other locations ? Thanks again… :^) JamAKiska (talk) 03:18, 7 February 2011 (UTC)

The short answer is: your guess is as good as anybody's. Template:Page break itself is nothing more than than a <span> with an assigned class (the self-explanatory 'pagebreak' & 'center' classes actually) and an id= assigned by whatever it is that one does or does not type into the 1st(un-named) &/or label= parameter(s) of the template. Think of this id= value much like the assigning of an #anchor to some text -- in reality that is what is happening 'underneath' it all when we use the {{anchor}} or {{section}} templates in case you didn't know. So unless the "finished" id=" " value conflicts with the id=" " of some other span on that page or one being transcluded into that page from the Page: namespace, the likelihood Template:Page break is causing the problem seems very low (well at least to me).

What happened on that DNB page above with ommitted "slash" was that the 'sectionalized' text was indeed normally transfered to the mainpage but leaving <pages index> "open-ended" like that caused everything else we may have tried to type after that point, never mind just our little DNB errata template, not to appear either when saved! (programing wise; everything else was "mistaken" for a part of a new <noinclude> section "on" the same Page: as the one transclusion was taking place "from" rather than totally separate text, typed-in [or added by some other template] in the main namespace only). — George Orwell III (talk) 05:20, 7 February 2011 (UTC)

Part 2

Wait a second - think I'm expierencing another quirk and this template's use in it's current state. On any mainspace page with the DNB article transcluded in from DNB Vol. 28 (any Volume?) that also has a DNB errata template in use, giving the errata info at the bottom of the page - where does clicking on the source tab along the top of the view frame actually take you to???

I get sent to the 1904 Errata's Index: instead of the original Volume's (Vol. 28's) Index:? — George Orwell III (talk) 05:47, 7 February 2011 (UTC)

Can I start with that I had started this approach from a conceptual point of view, and the fact that it being a template, then we could do things under the hood to generate output, and if it changed, then having the parameters was what was important. It was put out there for comment and development about alternate means to present data, it wasn't a finished product, and it isn't/wasn't something set as guidance to how we do the data. The progression to a standard implementation occurred quicker than I expected as I

To the id you are correct, and it wasn't until I was digging through it yesterday that I saw/recalled that it had an id attribute (non-documented). Happy to shuffle the text around, and omit it, have label = and put all the text into the notes.

Not sure about this so don't take it the wrong way if I'm wrong but.... it seems like you are still missing the rest of the point on "using" or "calling" Template:Page break rather than just mimicking its output by direct coding. No matter how you slice it you are 'breaking' that template by introducing another undocumented parameter, irregarless if it is called by an #if statement or not, while introducing the pagebreak class under DNB errata's usage at the same time in the mainspace article. It's not a "real" page break meant for the mainspace - it's just a references-like 'footer' that happens to look nifty as a typical pagebreak does in it's layout. Unless you Want the span class overlap with spans using pagenum (used in transclusions), the current approach is not the way go about getting the DNB template to work across the board IMHO. — George Orwell III (talk) 12:06, 7 February 2011 (UTC)

With regard to the #lst for the fragments, that is used as the use of <pages> in this situation seemed less relevant, less pull on resources and I could also see that the addition of page numbers could have added a confusing element or the potential for it. <grumble> The source tab used to rely on <pages> and that seems to have changed, I am now on ThomasV watch to get an explanation. — billinghurst sDrewth 10:51, 7 February 2011 (UTC)

??? The #LST sectionalized table row transclusions seemed fine to me - its the ultimate spans produced in the finished product & their classes/id=/title= interactions that is not up to par the way I see it.

Don't know if they are playing a role in the source tab thing but it happens when using sandbox2 as well & that doesn't use any defined class beside center so it isn't either template at the core causing this behaviour. — George Orwell III (talk) 12:06, 7 February 2011 (UTC)

┌─────────────┘
I went ahead and swapped in Sandbox2 for the Main DNB errata template and all displays exactly the same as before minus all that span id= stuff going on underneath. Turns out we need to keep the pagebreak class or the alignment of all the left-sided page links back to the Page: namespace abover the errata info becomes a bit too much to the right. The prior version before the swap-out remains in the regular Sandbox for now. — George Orwell III (talk) 13:43, 7 February 2011 (UTC)

Offsets

Kind of hard to build a range of offsets when the Index: pagelist itself doesn't reflect the actual offsets either! :-ο

Anyway, I swapped in the "fixed" template to the main position already - hopefully nobody even noticed that change - and have since managed to "redo" offset 10, didn't know where exactly offsets lower than 8 started/ended so I just jumped to the end where djvu page 293 & higher have an offset of 4 now. — George Orwell III (talk) 17:04, 7 February 2011 (UTC)

That figures…this template is still holding together…Offset drops to 6 on Errata p.195 — JamAKiska (talk) 17:06, 7 February 2011 (UTC)

Offset of 6 is properly ranged now too (~3 pages it seems) — George Orwell III (talk) 17:51, 7 February 2011 (UTC)

Offset of 10 is good until image page 179, where it drops by 2. Offset remains at 8 until image page 199, and again drops by 2 to an offset of 6 for two pages and drops for a final 2 to offset of 4 beginning on image page 202 where it holds steady till the end. Not sure when we’ll replace those 6 missing pages. JamAKiska (talk) 18:19, 7 February 2011 (UTC)

Check it again whenever you get the chance - it should cover that map exactly. Of course, much of those ranges haven't been created yet but the link part should open the djvu page in edit mode for the ones still missing at least. — George Orwell III (talk) 18:25, 7 February 2011 (UTC)

Image pages (djvu) 181-199 - offset of 8, image pages of 200-201 - offset of 6, 202+ offset of 4. JamAKiska (talk) 18:43, 7 February 2011 (UTC)

Please - my brain is fried enough as it is... is it set right or is there something that still needs to be adjusted???

(re: the underlined above) I believe djvu 179 is still offset by 10 and djvu 180 has no-text so djvu 181 was the logical place to switch to offset by 8 — George Orwell III (talk) 18:55, 7 February 2011 (UTC)

Your interpretation is correct… (we agree !), wanted to compare notes prior to finalizing template. And thanks…JamAKiska (talk) 19:04, 7 February 2011 (UTC)

Nicely done…and Thank-you JamAKiska (talk) 21:31, 7 February 2011 (UTC)

_/‾Source‾\_ tab

I check with the developer and it is the act of transcluding page namespace components that inserts the /Sources\ tab. Though it may be something that can be a future development.<shrug> I will do some checking around order of insertion, then I will try some cheating, and then if that is unsuccessful, we will need to determine if this kills the experiment or just a painful artefact. — billinghurst sDrewth 22:31, 21 February 2011 (UTC)

That sucks. Makes me wonder what the purpose of creating the &lgt;pages> tag was in the first place if it can't handle multiple inputs elegantly. — George Orwell III (talk) 01:35, 22 February 2011 (UTC)