Possible to restrict top-level xs:element names?

[kennardconsulting]

Hi guys,

I'm generating a bunch of XML Schemas that look roughly like:

<xs:schema ...>
   <xs:element name="foo"/>
   <xs:element name="bar"/>
</xs:schema>

You can see the actual ones here. My question is, this schema allows me to declare 'foo' and 'bar' in my XML document, but doesn't restrict me declaring, say, 'baz'. I was expecting something like...

<xs:schema final="#all">

...so that I could 'close' the set of possible xs:element names. Or maybe that the set was closed by default and I'd need...

<xs:schema ...>
   <xs:element name="foo"/>
   <xs:element name="bar"/>
   <xs:any />
</xs:schema>

...to make it open. But instead the set of top-level xs:element names appears 'open' by default and I can't find a way to close it? I have looked all over the specs and in this forum, sorry if I'm being dumb.

Regards,

Richard.

[Kevin Braun]

I'm not quite sure what you are asking.  Are you trying to restrict how
your schema can be extended, or are you trying to restrict what is
allowed in an instance document?  

Perhaps you have a conceptual misunderstanding?  Note that

<xs:schema ...>
   <xs:element name="foo"/>
   <xs:element name="bar"/>
</xs:schema>

is not the same as

<xs:schema ...>
   <xs:choice>
      <xs:element name="foo"/>
      <xs:element name="bar"/>
   </xs:choice>
</xs:schema>

(which isn't legal btw).

That is, the children of your <schema> element do not represent a model
for your instance documents, which is perhaps how you are thinking of
it?  Rather, they declare the elements, attributes, and types according
to which an instance document may be validated.  (Note by the way, that
there are different approaches to validating documents against schemas,
see
http://www.w3.org/TR/2004/REC-xmlschema-1-20041028/structures.html#validation_outcome).

I don't believe you can prevent someone from including your schema
document into theirs while at the same time adding more global elements,
if that's what you are trying to avoid.

Kevin

--
Objective Systems, Inc.
REAL WORLD ASN.1 AND XML SOLUTIONS
Tel: [phone]
Fax: [phone]
Toll-free: [phone] (USA only)
http://www.obj-sys.com



On 11/6/2009 1:58 AM, kennardconsulting wrote:

> Hi guys,
>
> I'm generating a bunch of XML Schemas that look roughly like:
>
> <xs:schema ...>
>    <xs:element name="foo"/>
>    <xs:element name="bar"/>
> </xs:schema>
>
> You can see the actual ones  http://metawidget.org/xsd here . My question
> is, this schema allows me to declare 'foo' and 'bar' in my XML document, but
> doesn't restrict me declaring, say, 'baz'. I was expecting something like...
>
> <xs:schema final="#all">
>
> ....so that I could 'close' the set of possible xs:element names. Or maybe
> that the set was closed by default and I'd need...
>
> <xs:schema ...>
>    <xs:element name="foo"/>
>    <xs:element name="bar"/>
>    <xs:any />
> </xs:schema>
>
> ....to make it open. But instead the set of top-level xs:element names
> appears 'open' by default and I can't find a way to close it? I have looked
> all over the specs and in this forum, sorry if I'm being dumb.
>
> Regards,
>
> Richard.
>  

[Henry S. Thompson]

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

The spec doesn't provide a mechanism by which a schema can constrain
the document element of schema-valid documents.

It _does_ provide license for conformant processors to do
schema-validity assessment with such a constraint in place [1].

Different processors allow users to invoke this kind of processing in
different ways -- what validator are you using?

ht

[1] http://www.w3.org/TR/xmlschema-1/#c-res
- --
       Henry S. Thompson, School of Informatics, University of Edinburgh
                         Half-time member of W3C Team
      10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
                Fax: [phone], e-mail: ht@...
                       URL: http://www.ltg.ed.ac.uk/~ht/
[mail really from me _always_ has this .sig -- mail without it is forged spam]
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.6 (GNU/Linux)

iD8DBQFK9Hp/kjnJixAXWBoRAkmzAJ4q7YYJ84l6UmdhZ+gdbrpNdEr9BwCeIxS/
EM49iXoq0GTVWxYb0Vv+bQI=
=muE/
-----END PGP SIGNATURE-----

[C. M. Sperberg-McQueen-2]

On 5 Nov 2009, at 23:58 , kennardconsulting wrote:

>
> I'm generating a bunch of XML Schemas that look roughly like:
>
> <xs:schema ...>
>   <xs:element name="foo"/>
>   <xs:element name="bar"/>
> </xs:schema>
>
> You can see the actual ones  http://metawidget.org/xsd here . My  
> question
> is, this schema allows me to declare 'foo' and 'bar' in my XML  
> document, but
> doesn't restrict me declaring, say, 'baz'.

I'm not quite sure what you're saying.  Where I come from, it is
natural to speak of declaring 'foo' and 'bar' elements in a schema
document (as is done in the skeleton you give), but odd to speak
of 'declaring' elements in the XML instance being validated.

If you mean "how do I prevent someone else writing a schema
document for my target namespace and including a top-level baz
element in it?", the answer is that you can't prevent other
people from writing what they like in documents on their own
hard disks.  The world is just structured that way.

On the other hand, you don't have to use the schema documents on
their hard disks when you validate.  If you want to ensure that
in a particular validation episode 'foo' and 'bar' will be
global elements in the target namespace, and 'baz' won't be,
you just need to arrange with your schema validator, at invocation
time, that it should read your schema documents, and not the
ones on other people's disks that include declaration for
'baz'.  How you arrange with a validator to read certain schema
documents and not read others is out of scope of the schema
language itself; if a particular validator does not allow you
to control things with the degree of precision you need, it's
a quality issue and you should look for a new validator.

If on the other hand you mean "how do I keep someone from
using a 'baz' element in my namespace?", again, the universe
is so structured that nothing you or XSD or any other schema
language can do could possibly prevent that.

The most you can do is try to arrange that such a usage
counts as invalid when the document is validated against your
schema.  So consider the question "how do I ensure that when
a user uses a 'baz' element in my namespace, it's marked
invalid?" The answer is, roughly:

   (a) tell the validator to read your schema document and
       not some bogus one which defines 'baz' (see above);
   (b) make sure your processor distinguishes between elements
       with [validity]=invalid and those with [validity]=notKnown
       (again, how you do this is between you and your processor,
       and out of scope for the XSD spec); and
   (c) make sure your own code that invokes the validator rejects
       the cases where [validity]=notKnown.

> I was expecting something like...
>
> <xs:schema final="#all">
>
> ....so that I could 'close' the set of possible xs:element names. Or  
> maybe
> that the set was closed by default and I'd need...
>
> <xs:schema ...>
>   <xs:element name="foo"/>
>   <xs:element name="bar"/>
>   <xs:any />
> </xs:schema>
>
> ....to make it open. But instead the set of top-level xs:element names
> appears 'open' by default and I can't find a way to close it? I have  
> looked
> all over the specs and in this forum, sorry if I'm being dumb.

I'm not sure what you mean by open and closed sets of elements
(or element names) here.  I'm also not sure whether by "top-level"
you mean "can be used as the root element of a document" or
"defined on the top level of a schema document; not local to a
complex type".  But in a way, I guess, it doesn't really matter:
XSD's design says, essentially, that the way you specify "these
declarations, and no others" is to say just that, but at validation
time, not at schema specification time.

I hope this helps.

--
****************************************************************
* C. M. Sperberg-McQueen, Black Mesa Technologies LLC
* http://www.blackmesatech.com
* http://cmsmcq.com/mib
* http://balisage.net
****************************************************************

[Michael Kay]

>    (b) make sure your processor distinguishes between elements
>        with [validity]=invalid and those with [validity]=notKnown
>        (again, how you do this is between you and your processor,
>        and out of scope for the XSD spec); and

You might like to note that Saxon's schema processor, and perhaps some other
products, expose this distinction by allowing you to select "strict" or
"lax" validation. Strict validation fails if the outcome is
[validity]=notKnown, whereas lax validation succeeds.

Regards,

Michael Kay
http://www.saxonica.com/
http://twitter.com/michaelhkay 

[kennardconsulting]

Thank you all for your valuable feedback. I think Henry put it most succintly:

"The spec doesn't provide a mechanism by which a schema can constrain the document element of schema-valid documents."

That's what I was wanting to know. I apologise if I used confusing terminology in trying to explain my problem!

Regards,

Richard.