Re: [dev-biblio] ODF and xsd schema
a) ODF is authored in RELAX NG, which has far more power to validate this sort of stuff than XSD. The Dublin Core is authored in XSD or RDF, and both namespaces are controlled by the W3C. I don't want to produce RELAX NG schema for Dublin Core as a prerequsite for further work. I assume in this that OpenOffice produces valid ODF output, and that ODF uses the dublin core properly when the dublin core is used directly. And there's debate about whether we want to be validating the XML vocabulary terms at all. So why are you using XSD to try to do just that? It doesn't make any technical sense According to those team assignments, (we/you) How Dare I ? (particularly the bit about sticking the user-defined content in the dc or dcterms namespace!). No, I never said anything about content. The name tag, 'Info 1' etc. of the user-defined fields if a qualified name (QName) should be validated. For user-defined meta data to be sharable, the author is responsible for correct tagging and valid content. In the case where the ODF specifies a tag (e.g. dc:creator) then the author is at most responsible only for content. b) the ODF metadata SC (and me in particular) have been working on and thinking about the use cases, requirements, and technical details of the enhanced metadata support for the past 12 months. I understand. I read your PDF. I am a user, whatever you do is fine by me. I'll cope. Notwithstanding the administrative issues with you making proposals outside the OASIS process, do you really think that you are going to tell us anything that we haven't already thought about? I'll ignore the absurd presumption of high-priesthood rank for your cartel, but will say that I haven't asked you to modify ODF in any way. You should feel free to summarily dismiss any proposal that was never made. What to see my schema ?, I asked. I'll take that as a no. As I said, I'm not trying to be rude, but I am pressing you to clarify what you are trying to do, and maybe rethink whether it's the best approach. I assuming ODF correct, which should explain the XSD schema. As I see it you have three types of output 1) meta namespace 2) dc namespace and 3) intermediate (user-defined). To move data from the meta or dc namespaces to the RDF namespace is fairly mechanical. Moving the intermediate to the RDF namespace can be done in two ways: 1) give meta:user-defined an xsi:type attribute as is the recommendation of DCMI for their dc and dcterms. or, 2) Use a GRDDL transformation ala the W3C. This is fairly new and not too well defined, but Chemists and Physicists familiar with Wave Mechanics will appreciate: A QName is 2/3 of a RDF tupple, ***to a high degree of probability***. Having said that, RDF is not the destination, it's the common measure. If you want to know how *you* would set up a workflow (maybe XSD based?) in this new world that I have been talking about, then I can probably help you think that through. It wouldn't be that hard. But it's just not clear from what you've said. It is my belief that the subject of meta data is one for which Open Source is peerless. Of the three types of output above, only the dublin core namespace can ever be trustworthy in Non-Open Source. e.g. msft:whiteblack/msft:white msft:blackwhite/msft:black My aim is to develop practical guidelines for applications which use ODF as a starting point. Access over 1 million songs - Yahoo! Music Unlimited (http://music.yahoo.com/unlimited) - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [dev-biblio] ODF and xsd schema
On Oct 31, 2006, at 6:24 PM, Gannon Dick wrote: a) ODF is authored in RELAX NG, which has far more power to validate this sort of stuff than XSD. The Dublin Core is authored in XSD or RDF, and both namespaces are controlled by the W3C. I don't want to produce RELAX NG schema for Dublin Core as a prerequsite for further work. I assume in this that OpenOffice produces valid ODF output, and that ODF uses the dublin core properly when the dublin core is used directly. Dublin Core is just a set of property terms really. It's technology-agnostic. But in ODF, the existing DC support is defined in RELAX NG, as is everything else in ODF. ... I'll ignore the absurd presumption of high-priesthood rank for your cartel, but will say that I haven't asked you to modify ODF in any way. Oh please; don't shoot the messenger. I'm just telling you a) you're not being clear, and b) the scope of these discussions are typically not for this list, but rather for the OASIS lists. And anyone can join the ODF Metadata SC, you included. You should feel free to summarily dismiss any proposal that was never made. What to see my schema ?, I asked. I'll take that as a no. Answering the question want to see my schema? presumes I understand what you are attempting to achieve with the schema, something that you have not made clear. As I said, I'm not trying to be rude, but I am pressing you to clarify what you are trying to do, and maybe rethink whether it's the best approach. I assuming ODF correct, which should explain the XSD schema. As I see it you have three types of output 1) meta namespace 2) dc namespace and 3) intermediate (user-defined). Now user-defined properties are not namespaced. To move data from the meta or dc namespaces to the RDF namespace is fairly mechanical. You wouldn't be moving metadata between these namespaces. You would be placing the metadata within the RDF framework/model; that's it. E.g. this: rdf:Description rdf:about=http://ex.net/1; dc:titleSome Title/dc:title /rdf:Description ... is just an RDF resource description with a dc:title property. The property is represented exactly the same in ODF 1.0; in the same namespace. Moving the intermediate to the RDF namespace can be done in two ways: 1) give meta:user-defined an xsi:type attribute as is the recommendation of DCMI for their dc and dcterms. or, 2) Use a GRDDL transformation ala the W3C. This is fairly new and not too well defined, but Chemists and Physicists familiar with Wave Mechanics will appreciate: A QName is 2/3 of a RDF tupple, OK, you need to change the language here. You are talking about transforming non-RDF XML to RDF. Yes, you can can use GRDDL (e.g. XSLT), or perhaps schema annotations in whatever language (XSD, etc.). Having said that, RDF is not the destination, it's the common measure. What does common measure mean? ... My aim is to develop practical guidelines for applications which use ODF as a starting point. Practical guidelines for what and whom? And to return to the subject I raised above, isn't this the job of the ODF TC? They develop the standard and write the documentation after all. Bruce - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]