Hi Vuk, Thank you very much for RSS updates! Looks marvellous.
Because it's getting so nice, I can't help asking to fine tune the model a little bit more. May I ? (1) dc:subject Now the items denote bookmarks, not target documents, it seems somewhat strange that an item (a book mark) has dc:subjects (because they are the subjects of the target document). If you don't mind adding one more namespace declaration, it would be better to use Tag ontology [1], and replace dc:subject with t:associatedTag, e.g.: xmlns:t="http://www.holygoat.co.uk/owl/redwood/0.1/tags/"; ... <item rdf:about="(tag_uri)"> <link>(document_uri)</link> <t:associatedTag rdf:resource="http://dbpedia.org/resource/Uniform_Resource_Identifier"/> ... etc. (It would be still better if you include <t:taggedResource rdf:resource="(document_uri)"/> to generate a triple to connect the tag=item and the document, but I'm afraid this might be too much asking you ?) (2) value of dc:creator It's good idea to have user (tagger) URI as value of dc:creator. It would be much better if you use <dc:creator rdf:resource="http..."/> instead of literal value (<dc:creator>http...</dc:creator>). (3) non declared XML entities For some reasons, this RSS includes several XHTML entity references such as â, ¦, ã, etc. These are declared in XHTML DTD, but not predefined XML entities. This makes the RSS ill-formed, and would cause fatal errors in XML applications. (This is not an RDF issue, but RSS/XML in general). I'd suggest use numerical reference such as ģ (or %HH encoding for URI) if necessary. > I am not sure what should I do in the case when bookmarks are sorted by > popularity (e.g. http://www.faviki.com/?sort=popular). I > use http://www.faviki.com/?u=[id] for item uris, and left dc:creator and > dc:date blank. I'm not sure if this is ok. You can check it on > http://www.faviki.com/rss.php?sort=popular. What do you > think? OK, in this case, the items are the target documents, rather than tags assigned by Faviki users. So, you can safely use target document URIs for rdf:about attributes on each item (and rdf:li in items). You can still use http://www.faviki.com/?u=[id], i.e. this denotes the same thing as document uri does. The values of dc:creator and dc:date for this item should be those of the target document, which are usually not known in bookmarking system. Hence, you can omit these properties (better than blank properties). In turn, dc:subject is fine in this case. If you are interested in Tag ontology, you might want to use t:taggedWithTag in place of dc:subject (this would look somewhat tricky, but works). I appreciate very much your effort to make Faviki RDF output better. thank you and best regards, [1] http://www.holygoat.co.uk/projects/tags/ -- @prefix : <http://www.kanzaki.com/ns/sig#> . <> :from [:name "KANZAKI Masahide"; :nick "masaka"; :email "[EMAIL PROTECTED]"]. ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Dbpedia-discussion mailing list Dbpedia-discussion@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dbpedia-discussion
