[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: The subsetting has begun

• To: 'XML Dev' <xml-dev@l...>
• Subject: RE: The subsetting has begun
• From: "Cavnar-Johnson, John" <JCavnar-Johnson@s...>
• Date: Fri, 28 Feb 2003 10:22:08 -0600
• Importance: Normal
• In-reply-to: <15725CF6AFE2F34DB8A5B4770B7334EEEACFE0@h...>

Inline

> -----Original Message-----
> From: Bullard, Claude L (Len) [mailto:clbullar@i...]
> Sent: Friday, February 28, 2003 8:27 AM
> To: 'Gavin Thomas Nicol'; XML Dev
> Subject: RE:  The subsetting has begun
> 
> Sure and we did SGML before that.  I think the
> SW is well-thought through.  Bray et al stated
> a position well.  Let's give credit where credit
> is due.  I also think the Common XML is well
> stated.  Applause all round.  There isn't a
> consensus but there are serious efforts to
> figure out a different layering strategy to
> XML 1.0.

I think there is a great deal more consensus than the discussion on this
list would lead you to believe. Norm Walsh stated it pretty well on the
TAG mailing list.  First, you remove DTDs, the internal subset, and
entities.  Essentially everyone who wants to subset XML 1.0 wants that.
If you define a parser conformance standard for simple well-formed
documents (Rick Jelliffe called them headless), you are essentially
done. You could remove notations, comments, and processing instructions,
but I don't think these are all that important. The functionality can be
handled by elements and attributes if they are removed or applications
can choose to ignore them if they are there.

Second, you have to decide how other XML recommendations apply to this
class of parser. I think this is pretty straightforward in most cases.
Common XML describes a pretty sound set of best practices, but these
parsers should conform to other XML standards just as current parsers
should. The real work would be in identifying any standards that can't
be supported by these parsers.  Does that boil down to what do these
parsers do when they see a DOCTYPE declaration, or is there more to it
than that?

> 
> We could probably make a pretty good list of
> the logical layers at this point.  That is the
> documentation side of the job.  But a subset
> or a new definition for the core comes down
> to essentially a "next release version of
> the software" that several vendors all have
> to agree to support.  What I don't see here
> on XML-Dev are vendors clamoring to do that.

I see lots of vendors going this direction, at least implicitly.
Microsoft, IBM, and several other vendors are pushing SOAP pretty hard,
so I think that tells us where they stand.  Sun's actions started this
thread.  

SNIPPED rest of message.

• References:
  • RE: The subsetting has begun
    • From: "Bullard, Claude L (Len)" <clbullar@i...>

• Prev by Date: RE: The subsetting has begun
• Next by Date: Having to programmically create XML Schema, is there a better way that just using the Document Object Model
• Previous by thread: RE: The subsetting has begun
• Next by thread: RE: The subsetting has begun
• Index(es):
  • Date
  • Thread