Le Thursday, March 15, 2007 8:32 PM [GMT+1=CET],
Sylvain Loiseau <[log in to unmask]> a écrit :

> I don't disagree with that. And it does imply that the TEI is
> different from the ODT format, since ODT does not imply
> interpretation at all.

If ODT wasn't designed with interpretation in mind why should
it be an XML based format with a particular schema ?
Why couldn't OpenOffice use a closed "binary blackbox" format ?
The fact is that if you have to document a format, which is a strong
community need for an interchange standard, it has to rely on a
comprehensive descriptive underlying format, like XML. If you
don't have to document a format, it doesn't have to rely on
something descriptive.
By the way, an indice of TEI and ODT closeness is the TEIOO
package ( described like :
"The XSLT stylesheets effectively provide a mapping
between the TEI DTD and the OpenOffice schema".

[from another mail]
>> I agree that the rationale of the TEI is to adress the notion of
>> interchange via "machine", to create "machine readable text". But you
>> should agree that adressing the point of being *machine*-readable in
>> general is different from adressing the point of being suited for
>> specific *program*.


> I would be interested in hearing your opinion about the alternative I
> proposed, between defining an encoding of application parameters
> common to all application, and defining a markup mechanism that each
> application used for expressing its parameters in its own way.

No program should claim a private markup for the same reason that
no person should claim a private markup. Maybe WSDL and UDDI
could help to build a proposition to "define an encoding of application


Serge Heiden, [log in to unmask],
ENS-LSH/CNRS - ICAR UMR5191, Institut de Linguistique Française
15, parvis René Descartes 69342 Lyon BP7000 Cedex, tél. +33(0)622003883