Patrice Bonhomme writes:
 > 2/ You have to distribute the alternative DTD with your documents. And
 > if we are dealing with a large set of different type ressources (this is
 > our case), it could become difficult to handle all those different DTD.

Don't forget that you can very often distribute the document without
the DTD (heresy!). If you, as author/publisher, have validated it,
does the end user need to do so as well?

 > 3/ In an editorial point of view, each time you want to modify your tag
 > set, you have to go on the pizza website and generate your brand new
 > DTD.

The pizza website _per se_ is unsatisfactory; it needs packaging as a
stande-alone application which you can run on the command-line as part
of your production process. I was going to look at how to do this
yesterday but got distracted :-}

 > Another solution, more XML oriented, could be to re-think the DTD and
 > its architecture. It is not simple, but maybe XML Namespaces associated
 > with XML Schemas could help. I will try when my phd will be finished ;-)

It is entirely possible (likely) that the TEI needs a recasting in the
world of schemas and namespaces. One just wonders who will do the

 > Another remark concerns some XSL stylesheets to display TEI ressources
 > or fragments. I know you've done a great job with XSL but it is a huge
 > task to cover the whole TEI.

and its also almost impossible, in the TEI cultural context of the "rend" and
"type" attributes. my TEI stylesheets are riddled with assumptions
about these, despit my best efforts. it needs a whole extra
customization layer added.

 > I was wondering if it could be possible to
 > unify our efforts (i did quite the same for the CES some weeks ago) to
 > provide an XSL architecture for TEI.
yes please....

 >  Who is using XSL with TEI
 > ressources? What are the requirements, the type of output? Do we need
 > FO? etc.

We are a good candidate for XSL FO, I think, as TEI tends to be used
for relatively large, complex, texts which are suited to batch
formatting. I think we should work at it. Speaking as a relatively
`real' user of TEI + XSL FO, I think its a viable system.

 > This type of initiative should also be launched and why not supported by
 > the TEI consortium that seems to be ... rather asleep at the moment.
 > Even if it is not the same context, i am always amazed to see the
 > dynamic Apache community and their different working groups.

Isn't it perhaps a little early to judge? The Apache XML stuff looks
all very inchoate to me.

Who knows what the TEI consortium is up to. I just hope they dont turn
into a stifling ISO.