W3C home > Mailing lists > Public > www-tag@w3.org > January 2003

Re: On subsetting XML...

From: Norman Walsh <Norman.Walsh@Sun.COM>
Date: Fri, 17 Jan 2003 10:02:33 -0500
To: www-tag@w3.org
Message-ID: <87hec72512.fsf@nwalsh.com>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

/ noah_mendelsohn@us.ibm.com was heard to say:
[...]
| I don't think it's quite fair to characterize this as saying they are not 
| forbidden but only deprecated.

My error. Apologies.

| Also: this is obviously controversial, but I don't see where it says in 
| the XML recommendation that every application of XML must use every 
| feature.  Surely there are applications that use only elements, and will 
| fault at the application level if confronted with an attribute.  SOAP is 
| such an application.  It makes use of some attributes, but not of PIs. 
| SOAP applications MUST NOT put PIs in messages, just as certain 
| applications MUST NOT put attributes in their XML.   SOAP is an 
| application of XML, not a redefinition of it. 

True. So we could define a onion-like set of subsets:

  elements
  elements+pcdata
  elements+attributes
  elements+pcdata+attributes

  {e+p,e+a,e+p+a}+{comments,pis}

  Each (or some) with or without doctype declarations.

I hope my observation that that would be a nightmare is uncontroversial.

So, the goal, it seems to me, is to find the 80/20 point for a subset that
can be detected syntactically. I think forbidding the doctype declaration is
that 80/20 point.

I have no problems with SOAP forbidding PIs or anything else, but I
think that what I proposed is the maximally useful subset that
addresses the needs of SOAP at a syntactic level.

| to the SOAP message as a whole?  PI's aren't well scoped to the XML 
| tree...they just sort of float in the document.

I'm sorry, Noah, but I have to disagree. They're exactly as well
scoped as empty elements. It happens that different users have given
them different semantics, but the same could be done with elements.

<body>
  <content>
    <ndw:hcf/>
  </content>
</body>

I could define the semantics of "ndw:hcf" as applying to the whole
document or only to the <content> element or only to everything that
follows it in document order or only to everything that precedes it or
any number of other semantics. And there's nothing you can do to stop
me.

                                        Be seeing you,
                                          norm

- -- 
Norman.Walsh@Sun.COM    | All our foes are mortal.--Valéry
XML Standards Architect |
Web Tech. and Standards |
Sun Microsystems, Inc.  | 
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: Processed by Mailcrypt 3.5.7 <http://mailcrypt.sourceforge.net/>

iD8DBQE+KBsJOyltUcwYWjsRAl2lAJ45g4wV1oATcFAaWo9MiS/JqKeNCQCfTY50
KQKsXH21kdugvACXJgFiPBY=
=lulU
-----END PGP SIGNATURE-----
Received on Friday, 17 January 2003 10:02:56 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:15 GMT