- From: Al Gilman <Alfred.S.Gilman@IEEE.org>
- Date: Fri, 7 Oct 2005 11:32:39 -0400
- To: "Weibel,Stu" <weibel@oclc.org>
- Cc: wai-xtech@w3.org
Stu, I don't know how familiar you are with the long-running debates in the Web community about how to characterize, in a manner something like the role filled by DOCTYPEs and MIME types to date, the general content characteristics of 'documents' containing a mix of for example, XHTML, SVG, and MathML markup vocabularies. The issue is reviewed in the following appendix which should be treated as a Request for Comments: http://www.w3.org/TR/WICD/#media-type-argument I'm under an action item from the Hypertext CG to try to get the Compound Document Formats WG some external help in resolving the issues raised in this appendix. The URC Standards, with my encouragement, have used a <dcterms:conformsTo> element to assert what specs the current XML object asserts allegiance to. See for example http://www.freestandards.org/pipermail/accessibility/2005-February/001100.html Is there anything you know, or contacts you can find in the Dublin Core 'Usage' activity can tell us, that would mitigate against our continuing to advocate for [syntax suitably bound to this DC term] in compound documents? There are pragmatic considerations mitigating against a total reliance on external metadata, as they tend to be less reliable than the utterances within the data object itself. In addition, relying on external metadata in HTTP header fields does not address large webs distributed on media, for example digital talking books. http://www.niso.org/standards/resources/Z39-86-2002.html I'm not asking you to resolve the internal/external issue for us, although Dublin Core experience that would bear on this issue could be helpful. I merely mention the internal/external issue as something that means we should be looking at what are good internal-metadata candidates and I'm not aware of why dcterms:conformsTo is not a leading candidate. Al
Received on Friday, 7 October 2005 15:32:58 UTC