- From: Matthew Dovey <matthew.dovey@las.ox.ac.uk>
- Date: Thu, 8 Mar 2001 22:35:39 -0000
- To: <rden@loc.gov>, "ZIG" <www-zig@w3.org>
> It does if you're talking about using http URIs. But (hopefully you would argree) we > need a better long-term solution. Some URIs "address" (e.g. http) and others > "identify". Not sure I agree with this - all URI's "address" either by location (URL's such as ftp, http) others by name (e.g. URN). All URI's should resolve to retrieving a document. Typically a dtd is specified by giving an URI which allows you to retrieve that dtd (and could be an http as well as a urn:). An XML validating parser will retrieve the dtd via the specified dtd to validate the structure of the document. So really any URI could in theory be used to specify a dtd. Of course you could specify a URI which resolved to something that isn't a dtd (or xsd etc.) e.g. a web page but in many ways this would be the same sort of illegal operation as specifying an undefined OID. Matthew
Received on Thursday, 8 March 2001 17:35:17 UTC