Wikidata talk:WP EMEW/Wikisource

Latest comment: 3 years ago by Jheald in topic Some first thoughts

Wikisource

 

Discuss

 

Progress

 


Discussion page for potential Wikisource uploads

Please sign your comments with four tildes (~~~~).

Some first thoughts edit

Wikisource is wikitext-based, so annotations would need mark-up based on templates.

A couple of options as to what to do with the original text here:

  • EITHER put it inside the template, as the template's first parameter (cf s:Template:taxon which does this, see in use in the wikitext of here.
  • OR sandwich the original texts between two templates: {{An-start}} original text here {{An-end}}

The first makes it easier to be sure that every annotation opened is then closed; but for all that I think I prefer the second, because it makes it more obvious when editing what is the original text, and how to easily strip annotations if desired.

ISSUE: Might need to be careful about line breaks and page breaks, with annotations that might still be open at the break, but this could be addressed.
Also to note that WS wikitext needs to support several different output modes -- eg page-by-page side-by-side with an image; page-by-page without the image; and running 'galley' text. The solution would need to be able to support all of these.

I would see the opening annotation template as changing a CSS style, eg perhaps introducing a pastel 'highlighter' background, or alternatively a box around the text, which the second template would then reverse back to normal. Elsewhere there would be a master control, changing whether the CSS style was active or not. This might be a javascript gadget. Without the JS gadget enabled and set to active no annotations would appear.

QUESTION: Would annotation presence identified by a coloured backgrounds conflict with highlighter backgrounds used by wikisource itself, esp in text-proofreading mode?

I would also see the annotation end-template as emitting a reference, which would contain the identified relevant wikidata Q-item, and the text of the annotation.

QUESTION: Where would the annotation be displayed? Side-notes are used by the field notebooks (eg: here), and are nice and near what they refer to; and they are all visible on the page. But they already look cramped. The Leland notes would appear to need to be much more voluminous and nuances. So, as a wikitext solution, we maybe should consider first a wikitext references/notes approach. These would not appear (that) near the text; but mousing over the typical [113] would mean the annotation could be read as a pop-up tool-tip.

There appear to be about 20 annotations a page, and 135 pages of Leland text (in this volume) -- so about 2700 annotations total? That is too many to put at the end of the text.

However the text appears to be broken into units, of about 1 to 2 pages each, so at the end of each of these units would be a natural place for the annotation wikitext notes to be presented, when viewed in the continuous galley-mode display.

A further question is where to put the wikitext of the annotations. The 'notes' syntax (as used in wikipedia articles) allows the text of the notes to be placed separately from where they occur in the running text of the article, and I would be minded to do this. Together with the {{An-start}} original text here {{An-end}} design for mark-up, it would make the wikitext of the running text very similar to regular wikisource wikitext, with minimal extraneous material at that point. That may be a very good thing for editability and acceptance. Internal 'names' (ie identifying alphanumeric codes) for each annotation could be chosen to match those on the VR system.

Possible wikitext might look like {{an-start|<type>|<id>}} ... {{an-end|<id>}}
Here the second <id> would be used to generate the named reference. The first <id> would not be strictly speaking necessary, but could be helpful to include to debug open/close matching glitches. The parameter <type> could be used to identify whether the note identified a place or a person or something else, which could be used as a selector between different available CSS styles - eg a different pastel background highlighter colour.
As to where to put the wikitext of the actual text of the annotation, that might best live at the bottom of each page, after the "printed text" for that page. I think (?) the media-wiki software may automatically then auto-group and present those annotations, if the text is viewed in wikisource 'single page' mode.

Community acceptance edit

The WS community is fierce that "Interpretive annotations are never allowed" (s:Wikisource:Annotations#Objectivity). However it also says that this "only applies to the annotations added by users."

The way forward here may be to regard the WikiSource text not primarily as a digital version of the 1906 edition by Toulmin Smith, but as a hosting of digital version of a 2021 annotated edition by Viae Regiae (that would also be available in other forms, eg a quite distinct IIIF manifest). The WS version would therefore be faithfully representing annotations from the distinct Viae Regiae edition, not "annotations added by users".


Above first thoughts by Jheald (talk) 16:44, 28 February 2021 (UTC)Reply

Discussion ? edit

In the above, I have somewhat assumed a completed proofread clean transcription of the text, that the annotations then apply to. As EncycloPetey points out, this in itself would be an enormous task -- (particularly in view of Leland's archaic and variable spelling, preserved in the 1906 edition, which will make accurate proof-reading a major undertaking). User:PKM is approaching the Internet Archive to see whether they can re-run the page images for Leland through the most up-to-date iteration of its OCR system; but this too may have difficulty with the relentlessly non-modern spelling. In fact it may well be that proof-reading may be as hard or harder as the annotation process that the Viae Regiae (Q105547906) team are about to commence.
So: is it possible that Wikisource would accept text with the VR annotations, but that was not necessarily particularly thoroughly proof-read? Jheald (talk) 12:51, 1 March 2021 (UTC)Reply
Return to the project page "WP EMEW/Wikisource".