[seul-edu] Pangée, LingoTeach, QVocab and data interchange

[Thomas Temp? <thomas@E604-b.resE.insa-lyon.fr>]

Hello...

I read with great interest the web pages about Qvocab, LingoTeach, Zope, and EduML. The main idea that seams to arise is that of trying to cooperate. The main idea for that cooperation is to use XML. At the risk of raising an old problem once more, let me introduce my view on that problem.

As I tried to figure how to exchange data among projects, I realized that even formalizing the data used only within Pangee in a way that allows straightforward interchange is no easy thing. Each program could well write its own DTD for its own needs, but this would be only a bit of solution, and would still require us to write specific filters for each app to share data with the others.

Note that this view embraces the scope of a spectrum of apps even wider than the 3 ones I cited here.

Obviously, EduML is about handling students records and such, not content. And even then, would a single DTD be sufficient to easily manage data both for Qvocab and LingoTeach, let alone the rest?
And if we built one, would it be any use in a project such as Pangée? We could always export our vocabulary lists, but it would be of little use to plain import them, as they would not fit our specific requirements (sorry, but this is only shown in the database RFC, RFC 03, which is not translated into english yet. But I give an introduction of vocabulary representation in the second half of the project desciption I translated into english. See http://prope.insa-lyon.fr/~ttempe).

What we need, in order to go beyond the one-to-one import filters, is a data formalization that embraces all of the projects around. This would allow to build a single DTD, say EduResML, for educational resources. And we won't get any further than pathching filters upon existing programs until we've built that EduResML. Or am I mistaken?

Then my question goes "is it possible to build such as DTD?". Because formalism is nothing but a way of _restricting_ the way information is put so it can be used in a more straightforward manner. Either we get too rescrictive, and we can't put the data into the mold without loosing some of its interest, or we get too lose, and (we've got HTML for that purpose already!) you just can't use the data anymore. This is why, for example, Pangée focuses on language learning, instead of trying to become the ultimate universal ultimate learning program. Or the ultimate universal do-anything program.

Is there an existing working group on something like EduResML, or EduLangResML? Or at least an ongoing all-encompassing (or much-encompassing) project?
Such a project could lead to a much wider-scope DTD, which would allow for an all-in-one interchange format, and in the same time trace the way to _the_ ultimate language-learning project.

Maybe I'll try to translate the RFC about data representation in the database ito english soon. I feel it could give some raw material to stress my point.

Now, has this point already beed raised a hundred times, is it outside the scope of this mailing-list, or is it of any interest if we want to move toward efficiently-designed XML?

Oh, and regarding Pangée, I looked into the Libwww docs today (www.w3c.org). I believe it is faisible to switch the program into using HTTP and XML once the data representation has matured... and once the school project is over ;).

Well, I guess I have had my load of computer usage for tonight;) I wish you all well,

Thomas Tempe.