OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [xml-dev] IDs considered harmful or why keys might be better than IDs...



On Monday 12 November 2001 03:48 pm, Jonathan Borden wrote:
> Well the DTD is data also, and I was assuming the context of my earlier
> post when I suggested that IDs can be provided in XML documents via
> internal subsets. 

I think DTD's and documents, even with internal subsets, to be different 
things.

> The point is that XML IDs are identifiers which are
> defined in XML 1.0 and do not depend at all on application specific
> semantics.

They always depend on application semantics. The only thing XML 1.0 does is 
says that they nust be unique in a valid document.


> Actually one cannot simply use whatever syntax one pleases for a fragment
> id. To be conformant to RFC 2396, i.e. to be a "URI" as defined in 2396,
> the fragment id must be constructed according to RFC 2396 and the
> appropriate media type. _Assuming_ XPointer is accepted as the relevent
> syntax for application/xml, text/xml, then a Raw Name is one of the 3
> syntaxes specified in XPointer, the others being ChildSeq and Full
> XPointer.

Right. Given that XPointer has not *yet* been accepted, and that RFC 2396 
explicitly states that no syntax has yet been decided upon, it is still 
perfectly possible to define a syntax.

FWIW. The RawName syntax in XLink is there for XHTML/HTML compatability only. 
As I've said, I think that's a bogus rationale for having something that can 
cause serious interoperability issues, and which is forcing people to examine 
other mechanisms for grandfathering the semantics because the underlying 
mechanism is unreliable.