W3C home > Mailing lists > Public > xmlschema-dev@w3.org > March 2007

Re: Schema 1.1: xs:anyEnumeration considered?

From: Pete Cordell <petexmldev@tech-know-ware.com>
Date: Mon, 19 Mar 2007 10:30:55 -0000
Message-ID: <004701c76a11$b16f05e0$c900a8c0@Codalogic>
To: "Bryan Rasmussen" <BRS@itst.dk>, <xmlschema-dev@w3.org>

----- Original Message From: "Bryan Rasmussen" <BRS@

> Whats the use of that? Wouldn't you just, in real life usage,
> want to say well this can hold any 3,4 length enumeration
> therefore I will use the <xs:pattern value="[a-zA-Z0-9]{3,4}"/>?

It's about the schema author specifying what they want.  Saying that
something is an enumeration that has to fit a particular pattern is very
different to saying that something is a string that has to match a pattern.

The presence of the pattern is after all a special case.  It would be quite 
legitimate to do:

    <xs:simpleType name="foo">
        <xs:restriction base="xs:string">
            <xs:enumeration value="ABC"/>
            <xs:enumeration value="DEFG"/>
            <xs:anyEnumeration/>  <!-- New -->
        </xs:restriction>
    </xs:simpleType>

which expresses a very different message to the reader than:

    <xs:simpleType name="foo" type="xs:string"/>

Also, databinding solutions are able to return the actual enumerated value 
(as an  integer type constant) or declare the various strings as const/final 
thus allowing for compile time checking rather than running the risk that 
the user enters a typo.

Cheers,

Pete.
--
=============================================
Pete Cordell
Tech-Know-Ware Ltd
for XML to C++ data binding visit
http://www.tech-know-ware.com/lmx/
http://www.codalogic.com/lmx/
=============================================

-----Oprindelig meddelelse-----
Fra: xmlschema-dev-request@w3.org
[mailto:xmlschema-dev-request@w3.org]På vegne af Pete Cordell
Sendt: 16. marts 2007 16:44
Til: xmlschema-dev@w3.org
Emne: Schema 1.1: xs:anyEnumeration considered?



Continuing my exploration of extensibility...

One thing I often see are sets of enumerations that are not extensible.  I
know that there is a trick with xs:union that you can do with this, but many
people don't know about it and it is ugly.  Simply being able to mark a set
of enumerations as extensible seems a lot cleaner to me.  For example,
something along the lines of:

    <xs:simpleType name="foo">
        <xs:restriction base="xs:string">
            <xs:pattern value="[a-zA-Z0-0]{3,4}"/>
            <xs:enumeration value="ABC"/>
            <xs:enumeration value="DEFG"/>
            <xs:anyEnumeration/>  <!-- New -->
        </xs:restriction>
    </xs:simpleType>

Here the pattern restricts provides the bounds of what the anyEnumeration
can permit.  (I thought about having a pattern attribute within the
xs:anyEnumeration element, but the schema snippet above better reuses what's
defined already.)

This very much more allows the developer to ask for what they want and
doesn't require them to work around the limitations of the language with
various insider tricks.

Going further, if named wildcards were allowed (as per my earlier topic in
the week), the anyEnumeration facet could be:

            <xs:anyEnumeration socket="foo"/>

and in another schema you could have:

    <xs:plugin socket="core:foo">
        <xs:enumeration value="HIJ"/>
        <xs:enumeration value="KLMN"/>
    </xs:plugin>

>From what I understand, this sort of notation could go a long way to
addressing the problems that Jon Bosak(sp?) described that UBL had with
enumerations.

Anyway, the question is, was such a thing discussed?

Thanks,

Pete.
--
=============================================
Pete Cordell
Tech-Know-Ware Ltd
for XML to C++ data binding visit
http://www.tech-know-ware.com/lmx/
http://www.codalogic.com/lmx/
=============================================
Received on Monday, 19 March 2007 10:31:14 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 23:15:41 UTC