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

RE: Will the next version of XML Schema have a schema-for-sche

• From: noah_mendelsohn@u...
• To: Rick Jelliffe <rjelliffe@a...>
• Date: Thu, 18 Oct 2007 14:09:39 -0400

Rick Jelliffe writes:

> In XSD, for example, the use of elements to describe complexContent and
> simpleContent is jarringly odd and unpleasant, and it comes from the
> fact that XSD cannot use attributes to select a content model. Elements
> is all it has in this situation.

Yes, indeed!  That was exactly the reason for that clumsy syntax, and some 
of us argued that having a more convenient syntax was more important than 
having the schema language be able to put a tight bound on its own 
validation.  Of all the truly questionnable design decisions that I 
contributed to, this is the one on which I wish that I had more strongly 
non-concurred.  I did advocate a simpler syntax, but when the more complex 
one got more support in the workging group, I went along.  In retrospect, 
I think that was a big mistake.  While there are many reasons, often 
discussed, why some aspects of schema are truly and structuarlly (too) 
complex, but users' experiences are greatly complicated by the fact that 
the syntax is so unnecessarily clusmy.  For example, there's nothing 
structural in the language that would prevent a syntax like:

        <element name='e' type="xsd:integer" maxInclusive="10"/>

This would map just fine to the components we have, and could have been 
offered instead of:

        <element name="e">
          <xsd:simpleType>
            <xsd:restriction base="xsd:integer">
            <xsd:maxInclusive value="10"/>
            </xsd:restriction>
          </xsd:simpleType>
        </element>

The latter is arguably better markup and is certainly more explicit.  To 
build components for the former, you'd need to follow a runtime type graph 
to find out that xsd:integer is a simpletype.  That's not something you're 
at all likely to be able to talk about in a Schema for Schemas.  Still, I 
bet a lot of people who hand author schemas would prefer the shorter 
syntax.

Noah

--------------------------------------
Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------

• References:
  • RE: Will the next version of XML Schema have aschema-for-schemas that is standalone (no English prose needed to describeconstraints in the language)?
    • From: Rick Jelliffe <rjelliffe@a...>