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] Re: determining ID-ness in XML

Isn't the point to use a means the XML processor 
isn't free to ignore per specification?  That is 
why the concept of "reliability" was introduced 
although one could say "efficiency" and mention 
the cases of XPointers and serialization. 

One extends the system vocabulary precisely because 
it extends the requirement for the XML processor.
If all you need is a convention, a PI or an 
alternative namespace are equally ignorable. 
Otherwise, we could just go on as is:  "if you 
need an ID, spec a DTD and cite it in the 
contract for the communication when using 
well-formed files.  This is only as reliable 
as your partners are diligent."

Independent parties should not be extending 
the rules for the system vocabulary.  That 
is precisely what so many here beat MS up for. 

There is no QED here.  There is nothing to 
be proved other than that a solution meets a 
requirement and that there is a rough consensus 
on the necessity for the requirement and the 
effectiveness of the solution.
Then it is a sales job to the Web Architecture 
group.  This is why a precisely stated requirement 
means so much to this kind of process: traceability.


-----Original Message-----
From: Elliotte Rusty Harold [mailto:elharo@metalab.unc.edu]
At 10:51 AM -0500 11/5/01, Champion, Mike wrote:

>We need to figure out whether this
>discussion is aimed at persuading the W3C to change the "system vocabulary"
>to support IDs without requiring external DTDs/Schemas, or whether it is --
>in the tradition of SAX and RDDL -- aimed at producing a convention that is
>orthogonal to the W3C's Recommendions.  The former strongly suggests
>attributes, but the latter more or less demands PIs, as well as some
>"political" work to make PIs less of an endangered species.

The latter solution doesn't demand PIs. It just prevents us from 
using the xml: prefix. If we wanted to be orthogonal to the W3C, then 
it's easy enough to define an xid:id prefix with xid mapped to 
http://xml.org/id/ or something like that. I think this has already 
been suggested as a way of adding multiple IDs to things. There's no 
need to get PIs involved.