Re: XPackage anyone? (RDF newbie question)

Hi Daniel,

I understand your argument for keeping the metadata in a database; however, even for a database you need a "vocabulary" (in the sense that names for tables and columns form a vocabulary).  But Chaals is right: it doesn't matter whether that vocabulary exists in the form of RDF data or an RDF schema - any format (such as a list of terms and their use or semantics as I found on the Fink site) would still be a vocabulary.

While this is the first time for me to try to create RDF (and an RDF schema for it) so I'm struggling with RDF concepts and RDF/RDFS syntax at the same time as trying to create a vocabulary for my application, I've worked on XML and XML schema before - which still involves creating a vocabulary and that part of the process is not so different. I've found that when designing a vocabulary it can be very helpful to look at vocabularies (in whatever form) for similar or related subjects: it's not just words used that may be better than the words you'd thought of - it's also that some words signify (or merely suggest) concepts that are helpful to have (even if you end up using different vocabulary for those concepts).

So ... I did follow your "go directly to database" (or similarly-worded) link on the rpmfind site but even there I did not see any clue to a vocabulary. Is there a DDL or other schema for your database available somewhere? (Or would this be regarded to constitute a security risk)? Or just a summing up of the terms you use might also be helpful.

Thanks,

At 23:36 2003-01-15, Daniel Veillard wrote:
>On Wed, Jan 15, 2003 at 05:28:28PM -0500, Charles McCathieNevile wrote:
>> Hi Marjolein,
>> 
>> Some things trigger my "instant response" mechansims ;-) You could possibly
>> find out more about rpmfind from Daniel Veillard, who put it together. If I
>> recall correctly he used a draft version of RDF to encode the stuff, but I
>
>  Right,
>
>> couldn't find a pointer easily on rpmfind.net either. I guess I will have to
>> get an RPM system and dig inside it a bit.
>
>  Well, only metadata extracted were using RDF (and an early version)
>But I'm using a past tense, because nowadays the metadata are kept in
>a database instead of XML/RDF files. Managing 50,000 - 100,000 files
>and their relationship was a real nightmare, that didn't scale and was
>impossible to keep clean.

[snip]
-- 
Marjolein Katsma      webmaster@javawoman.com
Java Woman - http://javawoman.com/

Received on Sunday, 19 January 2003 07:24:13 UTC