On Monday 18 February 2002 11:00, Gianugo Rabellino wrote:
>. . .
> 1. XML:DB metadata. . . .
> 2. DB specific metadata.. .
> 3. User provided metadata. . .
>. . .
I'm jumping in without much knowledge of XIndice internals, but I've
been thinking about this lately the context of CMS/information
management.
Maybe these thoughts help:
Speaking of metadata, the "resource fork/data fork" concept of the
Mac filesystem comes to mind here:
-Staging/release forks, the same data at several stages of completion
("release" is published, "staging" for internal use only, later
promoted to "release").
-Various metadata forks as you mentioned, different facets of the same
data
Maybe implementing the concept of "forks" internally in an open way
wouldn't be too hard? I think it would be more useful that specifiying
a fixed number of metadata types.
>. . .
> This flexible metadata structure IMO asks for an XML metadata format.
> RDF/DC seems to be the most obvious choice. . .
>...
IMHO the user metadata format should be left open for maximum
flexibility. Or, if this "forks" concept makes sense, schemas could be
assigned to forks based on fork names (i.e. "dc-meta-fork" maps to DC
schema).
--
-- Bertrand Delacr�taz, www.codeconsult.ch
-- web technologies consultant - OO, Java, XML, C++
