Thanks for all of the kind replies. Thanks Guarav for the links, I wish there was a clear explanation for all the elements in Aubrey’s layers. As for your example of annotation, it is gorgeous and is obviously the result of a lot of dedicated work.

Birgitte and Lars, maybe an example would be the best way to explain what I am trying to ask. I apologize in advance that the following example is contrived, because I am hard put to find a true example of the issues I’m talking about in actual English texts on Wikisource. So I’ll try to take you through my imaginary example step-by-step.

1. Imagine an English-language encyclopedia that was incredibly popular, so popular that it was published in both American and British editions. It’s the same encyclopedia in both editions, but the spelling is different. In a digital version of this encyclopedia that included database functions you would be able to tag words so as to allow the reader to choose which version s/he wants in terms of spelling with something like this:

{{spelling|A=color|B=colour}} (I understand that American versus British spelling is something so simplistic that an automatic function could probably deal with it even without tags or templates, but bear with me by considering that there might be other valid variations that are far more complicated, and which need to be tagged and documented in order to provide the user with options.)

2. Now further imagine that this encyclopedia was so popular that it was republished many times in the *same* edition. Each time the typesetting was manually reset, which allowed for small corrections to be made (e.g. typos but sometimes even greater variations) but at the very same time allowed new errors to creep in. So when you edit the text, you have several good editions of the same encyclopedia that cast light upon one another, but none of which is perfect. The best way to digitally republish such an encyclopedia would be to fully document the variations using a function something like this (where a,b,c,d are various reprints of the text):

{{variant|select=Wikisource is the Free Library and invites you to contribute!|=abd|c=Wikisource is the Free Libraries and invites you to contribute!|note=c is often sloppy about singular and plural nouns}}

For those who are familiar with the “critical apparatus” that often accompanies classical texts in scientific editions, this is a way to take that kind of apparatus and embed it within the text itself on the edit page. But a database function would further allow the user to show one particular version as s/he chooses. It would also allow the user to have a function making indications of variant readings and notes on them appear or disappear by turning the function on or off.

3. Now further imagine that what we are talking about is not an encyclopedia, but rather a legalistic type of literature that is organized by numbered sections and subsections. Furthermore, this literature cites itself avidly, and certain subsections of this book might be cited elsewhere or appear in other contexts tens of thousands of times (literally). Because of the need for convenient citation (often through transclusion) along with the fact of numerous similar editions with different pagination, the page-based “Proofread Page” is no longer the optimal tool for creating digital editions of this literature, and actually makes things more difficult for contributors. Instead, wiki-pages based on the natural division of the text allow for easy citation while keeping things as simple as possible, plus links to various scanned editions can be provided for verification and further improvement of the text.

What I have described here in #3 is the main reason why “Proofread Page” is not heavily used in Hebrew Wikisource. It is installed but not well-supported with infrastructure. I emphasize that in my opinion it is an incredible and important tool, and certainly should be used where appropriate for huge numbers of texts. At Hebrew Wikisource there is certainly no policy against it, and of course we would love it if someone came and started to use it on appropriate texts and improved the Hebrew infrastructure for it. But that still wouldn’t make it appropriate for all texts.

However, in terms of database functions within the text itself I don’t think there is really any issue with Proofread Page. Because when all is said and done, the proofread text of a page is still wikitext. And the question is whether wikitext in general (not PP in particular) could be made to support the kinds of database functions described above.

I hope all of this is clearer than my original inquiry. Was anything discussed at Wikimedia (including Aubrey’s various layers) that might make solutions possible for functions like these?

Dovi