W3C home > Mailing lists > Public > xml-dist-app@w3.org > July 2001

Re: Infoset based rewrite of SOAP Section 4

From: Mark Nottingham <mnot@mnot.net>
Date: Mon, 2 Jul 2001 11:11:00 -0700
To: Doug Davis <dug@us.ibm.com>
Cc: Noah_Mendelsohn@lotus.com, xml-dist-app@w3.org
Message-ID: <20010702111059.A27968@mnot.net>

Good question. If we went down this road, a means of communicating
this would have to be introduced; one suggestion is to make it an
integral part of the binding. Another would be to signal it somehow.

Just to be clear, I'm very much in favour of using Infoset to
communicate well in our documents, by using the terms defined. I'm
interested but somewhat wary of the use of Infoset to allow other
representations on the wire; my feeling now is that it's a good hook
to leave for future use, but we shouldn't explore it too much, as
that way complexity lies.

In other words, use Infoset, yes, but only define an XML
serialization, allowing 1.2 implementations to assume its use.




On Mon, Jul 02, 2001 at 01:22:19PM -0400, Doug Davis wrote:
> Does this "pared-down" XML mean that "technically" the soap
> message _might_ not be completely spec compliant anymore because
> the device we're sending it to knows to expect this "pared-
> down" version?  If so, can that device still say it is a SOAP
> node?
> -Dug
> 
> 
> Mark Nottingham <mnot@mnot.net>@w3.org on 07/02/2001 01:00:17 PM
> 
> Sent by:  xml-dist-app-request@w3.org
> 
> 
> To:   Noah_Mendelsohn@lotus.com
> cc:   xml-dist-app@w3.org
> Subject:  Re: Infoset based rewrite of SOAP Section 4
> 
> 
> 
> 
> I recall that one of the use cases we considered (don't remember if
> it got through or not) would be use of 'pared-down' XML for use by
> small devices (hope this won't start the debate over the meaning of
> 'devices' again ;)
> 
> It seems that using Infoset would enable this - those wishing to use
> such a subset would define its relationship to the Infoset, and use
> that simplified serialization instead.
> 
> 
> On Mon, Jul 02, 2001 at 11:38:06AM -0400, Noah_Mendelsohn@lotus.com wrote:
> > Rich Salz writes:
> >
> > >> In particular, the idea of "describe the
> > >> abstract data" and then "describe
> > >> particular syntax in detail" scares
> > >> me.
> >
> > Good, I think we have the right issues on the table now.  I think we both
> > agree on what is possible and why, and now it's a question of whether the
> > extra layer of abstraction clouds or clarifies the implementations that
> > people will be building.
> >
> > For me, it is a little hard to reason about certain forms of compressed
> or
> > otherwise optimized XML without reference to the Infoset.  It's quite
> > reasonable in that case to assume that at no place in the implementation,
> > from API (e.g. DOM) to bits on the wire, did the "<" exist.  For that
> > reason, I think I find it more straightforward to reason about the
> > Infoset.  I think I am reading you to say:  look, this Infoset stuff is
> > very abstract when an implementor is trying to figure out what belongs on
> > the wire, keep it simple and direct.
> >
> > If I have understood you correctly, I think those are two reasonable
> > positions to have on the table for comparison.  Although I lean toward
> > Infoset, I don't think the choice is entirely obvious.  The arguments
> > against definitely include the ones you give.  Thank you.
> >
> > ------------------------------------------------------------------------
> > Noah Mendelsohn                                    Voice: 1-617-693-4036
> > Lotus Development Corp.                            Fax: 1-617-693-8676
> > One Rogers Street
> > Cambridge, MA 02142
> > ------------------------------------------------------------------------
> >
> >
> >
> 
> --
> Mark Nottingham
> http://www.mnot.net/
> 
> 
> 

-- 
Mark Nottingham
http://www.mnot.net/
Received on Monday, 2 July 2001 14:11:01 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:59:02 GMT