Talk:TEI Web Publishing

How easy should it be?

If we're really serious about making publishing easy, surely the easiest is to insert the following at the top of the TEI?

<?xml-stylesheet href="path/to/stylehseet.xsl"  type="text/xsl"?>

(said Stuart at 04:06, 6 March 2010)

But that would work only if we want to publish a single TEI document rather than a collection of them. (xi:include will not always help, sometimes the relationship between documents should be expressed as reference rather than inclusion).

So maybe let's turn your question into a possible basic condition: how to publish a set of TEI documents without having to modify them manually with e.g. stylesheet PIs? Piotr 05:05, 6 March 2010 (EST)

The first sentence of the page talks about 'a single TEI document' Stuartyeates 15:00, 6 March 2010 (EST)

Maybe it shouldn't be steps but rather issues. Still, the 2nd sentence talks about a set of documents, so we can make the assumption that the 1st one talks about a single-membered set, as a particular case. (Or we can make a different assumption :-)) Piotr 15:59, 6 March 2010 (EST)

I think the issue here is that we're conflating documents and files, both of which are further conflated in the minds of many users with works. Stuartyeates 21:41, 6 March 2010 (EST)

That would make a fine introduction. Indeed, xinclude has nothing to do with documents (or has as much to do with them as inter-paragraph spacing, perhaps). Let me rephrase (still as an exercise, to see if we can get anything interesting out of it):

Given a document encoded in the TEI, what are the available publishing strategies?

• One is to add the PI (to the (main) file containing the document) and count on a client-side engine in a browser (if this is what you meant, Stuart). The XSL list suggests it may be a tough task, especially where cross-platform, cross-browser portability is intended. (expertise involved: XSLT, browser XSLT engine quirks)
• Another is to prepare a transformation scenario, possibly including a pipe, and run a processor on it. (expertise involved: XSLT + Saxon/Xalan/xsltproc -- these are at least well-documented)
• The next scenario would involve a native TEI db (which? let's say eXist for starters; expertise: +XQuery; and aren't we looking here at a slightly different range of applications?)
• Next is Thomas Crombez' Google Docs suggestion, which may involve an easy step (->HTML) using default stylesheets.
• Ready-made TEI publishing solutions (other than Sebastian's stylesheets) -- if there were any such full-fledged systems, would we bother doing this?
• Basic command of (X)HTML/CSS is a given. Somewhere on the way, one should possibly mention Cocoon, XProc... am I making any sense? If I am, it begins to look like a book... :-/ But something could be carved out of this as one of the possible examples to follow. There are lots of sources on XSLT, and Thomas has a tutorial for people who want the Google way; as Martin noted, there is little documentation on the database scenario, and this is where this might go.

At this point, my interests are in the use of a native XML database for the purpose of manipulating dictionaries and corpora, so I would be glad to help develop documentation on that on the understanding that there are going to be areas where I will be learning and experimenting, so not really contributing from an expert's position. Piotr 17:30, 7 March 2010 (EST)