Monday, January 27, 2003 9:19 AM
> Le classement en genres des textes que j'encode en TEI est nécessaire à
> cause de la demande des utilisateurs de la base textuelle (FRANTEXT).
Splendid to see that FRANTEXT is moving to TEI encoding. Bringing this huge
and hugely important resource into the TEI fold has important implications
for the interoperatiblity of textbases.
> 1-A quel endroit du TeiHeader cette information peut-elle se mettre?
> et sous quelle forme?
Basically, this belongs in the profileDesc component of the teiHeader
(see http://www.tei-c.org/P4X/HD.html#HD4 )
using the text textClass element (P4 sec 5.4.3) and some suitable taxonomy.
The precise form depends on which of the possiblities envisaged in that
section you adopt.
> 2-J'ai distingué 4 grandes catégories: le théâtre, la poésie, le
> roman et l'essai (ou littérature d'idées)
> Ces distinctions vous-semblent-elles valables (en ce qui concerne les
> genres littéraires)?.
> 3-Pouvez-vous me conseiller pour ce qui concerne les textes
> non-littéraires? (liens exemples, nomenclature de genres.)
For both these things I would strongly advise you to use an established
taxonomy for the generic descriptions rather than inventing your own. In the
English-speaking world, examples are the Library of Congress and the British
Library schemes (and do I mean the schemes per se, rather than the actual
classifications given to specific works under those schemes, see below). You
will probably want to choose a Francophone equivalent, but choosing an
existing published standard taxonomy would ensure that its terminology could
be systematically mapped to other "national" taxonomic schemes. This is
important if interoperability (particularly searchability of distributed
textbases by remote clients) is to be achieved. Once you are TEI-encoded,
your user constituency can be extended far beyond your current one, and
FRANTEXT can become part of a much larger distributed textbase.
For example, I am currently involved in encoding an Anglo-Norman version of
Roger Frugard's Chirurgia, which will go into the AND textbase. I would hope
that at some future date a remote client examining usage in medieval
Francophone medical texts could fire off a query at both the AND and
FRANTEXT and be able to retrieve results on that text type from both
textbases in a single operation. A key prerequisite of that is that our
respective teiHeaders use taxonomies with an established mapping between
them. We will need on our side to improve somewhat on LoC descriptors for
this text such as "Anglo-Saxons - Medicine", "12 cent. English" etc. since
these conceal the not insignificant fact that this particular text is in
French with some Middle English glosses and has little to do with
"Anglo-Saxons". But something, albeit more accurate, at that level of
detail, and using a standardized taxonomy is essential if a textbase is to
exploit the full potential (including things unimagined by current users) of
TEI markup and the WWW. And that would apply also to literary texts: your
proposed "grandes catégories" would need considerably greater (but
standards-based) granularity if your users are to have the same scope for
generic differentiation in their searches that you already give them in
terms of period, author and internal context in your current FRANTEXT
The Anglo-Norman Dictionary http://www.anglo-norman.net/