Re: The relentless march of abstraction (fwd)

[ht@cogsci.ed.ac.uk (Henry S. Thompson)]

John Cowan <jcowan@reutershealth.com> writes:

> > Just the first few paragraphs opened my eyes to where the process is going
> > with XML. I never understood what "infoset" was all about. Now that I do, 
> 
> Henry is a fine fellow who has done much good work.  His view of the
> Infoset is not the only view, nor is this potted summary (IMHO) a fair
> representation of even Henry's view, much less the total view.
> 
> (I speak as an editor of the Infoset, but not officially for the Core WG.)

Thanks for the gentle pat on the back, and the gentle reproof.

Makes slides in haste, repent at leisure -- probably a good motto for
us all.

To attempt a brief clarification, at the risk of just making things
more confused:

1) The draft W3C Infoset spec. is about providing a single vocabulary
for other W3C specifications to use in describing their input
requirements and/or output provision as regards XML 1.0 plus
Namespaces.  Another way to describe this is as an abstract data model 
for XML 1.0 plus Namespaces documents, or, as I said in my talk, the
definition of the type *XML Document*.

2) Without emphasising that I was moving from the realm of
specifications to the realm of applications, in my London XML DevCon
talk I went on to use the phrases "Infoset Pipeline" and "Post
Schema-Validation Infoset" in a more concrete way, as an abstraction
over implementations of parsed XML 1.0 plus Namespaces documents.  Not 
quite the same thing, as John correctly observes above.

I certainly hear (loud and clear :-) people on this list saying they
don't find it helpful to focus on the level of graphs of information
items, they just want SAX event streams, or maybe even just character
streams.  That's fine -- I was just trying to communicate _my_
excitement at what added value _I_ had found by concentrating on the
more abstract level.

Finally, from the W3C perspective, I stand behind what I said about it
being _impossible_ to coordinate effectively between e.g. XLink, XML
Signature, XML Query and XML Protocols without both Infoset and
Schema.

ht
-- 
  Henry S. Thompson, HCRC Language Technology Group, University of Edinburgh
          W3C Fellow 1999--2001, part-time member of W3C Team
     2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 131 650-4440
	    Fax: (44) 131 650-4587, e-mail: ht@cogsci.ed.ac.uk
		     URL: http://www.ltg.ed.ac.uk/~ht/