W3C home > Mailing lists > Public > xmlschema-dev@w3.org > June 2001

Re: ASN.1 => XML Schema questions

From: Dan Connolly <connolly@w3.org>
Date: Tue, 26 Jun 2001 19:15:32 -0500
Message-ID: <3B3925A4.1B80563E@w3.org>
To: Kohsuke KAWAGUCHI <kohsukekawaguchi@yahoo.com>
CC: Geoff Elgey <elgey@dstc.qut.edu.au>, xmlschema-dev@w3.org
Kohsuke KAWAGUCHI wrote:
> 
> > I'm pretty sure the reason is that in W3C XML Schema validation,
> > the result includes not just a "yes, this is valid"/"no, not valid"
> > but also "and this part of the input matched this part of the
> > schema" i.e. "it has this type, is associated with this annotation"
> > etc.
> 
> I don't think so.
> 
> Yes, W3C XML Schema can tell the application that "this
> element/attribute has this type", but almost any schema language can do
> this.

Really? Which ones?

Assuming you meant RELAX-NG,
I just spent an hour or so looking at the RELAX-NG
stuff, and I don't see any indication that it can.

On the contrary, it seems to adhere to the design
principle that Murata Makoto suggested for W3C XML Schema
that was not adopted:

[[[
RELAX NG does not support features of XML DTDs that involve changing the
infoset of an
      XML document. In particular, RELAX NG

          does not allow defaults for attributes to be specified
]]]

--        RELAX NG Tutorial
http://www.oasis-open.org/committees/relax-ng/tutorial.html
Tue, 12 Jun 2001 12:10:31 GMT

The "no changing the infoset" design appeals to me because
it seems simpler, but I gather it doesn't meet the needs
of folks that want to do XML Protocols, Java bindings,
and the like.


> And W3C XML Schema still cannot tell things like "these three elements
> matched this <sequence> particle", can it.
> 
> For example, consider the following schema and the following instance.
> 
> <sequence maxOccurs="unbounded">
>   <element ref="foo" maxOccurs="unbounded"/>
> </sequence>
> 
> <foo/><foo/><foo/>
> 
> I think this is OK wrt the UPA constraint.
> 
> Now would you tell me how many <sequence> blocks are there, in the above
> instance?
> 
> Is it
> 
> sequence{<foo/>}, sequence{<foo/><foo/>}
> 
> or is it
> 
> sequence{<foo/>,<foo/>,<foo/>}
> 
> or what?

Good question; I don't know; I don't know if it's specified
at all.

> > Keep in mind that the other languages won't give you type/annotation
> > info as a result of checking.
> 
> Please keep in mind that there are only one or two validators that give
> you type/annotation info for XML Schema at this moment, and there is
> still no interface for that.

Well, the tools are somewhat immature for XML Schema,
but I think you can get at the PSVI via XSV; maybe
not from the web form. I think Henry showed me how
one time, but I can't seem to find the details.

Does anybody else have tools that give access to the
PSVI info?

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Tuesday, 26 June 2001 20:15:39 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 11 January 2011 00:14:21 GMT