[om] About new XML encoding methods for OpenMath

[Michael Kohlhase]

Dear all,


Actually, did not suggest anything, I was summarizing the discussion in the
OM2 committee. Other than this, I did suggest in OMDoc 1.1 to use 

<OMS cd="omstd:set1" name="in xmlns:omstd="http:www.mathweb.org/cd"/>

the namespace declaration could be anywhere above this. This is the form
Paul is probably referring to. But I have since reconsidered this
suggestion and it was not in my talk. The nice thing about misappropriating
namespace mechanism for cataloging purposes is that it is relatively
lightweigt. This particular use (qnames in attribute values) is not
terribly well supported. If I remember correctly, this case is not even
considered in W3C specifications like XML fragment (not that many people
use that anyway).

With respect to the other (radical-namespaces) suggestion, which I did
mention in my talk I am going back and forth like David. Maybe the way to
go is to develop both XML encodings, to a point, where we can see where
they are leading, and then let people vote with their feet.

     Michael



>   Namely, Michael suggested to use instead of:
> 	<OMS name="in" cd="set1"/>
>   to use, with a namespace declaration like
>   xmlns:lset="http://www.openmath.org/cds/set1", the following
>   	<lset:in/>
> 
> This has been suggested many times (since even before namespaces was a
> REC, if I recall correctly), the main disadvantage is that given such an
> approach one would essentially have to replace OMS by the ANY (xsd:any)
> content model in any DTD (or schema) for openmath.
> 
> This means that
> 
> <OMA>
>   <html:p>Some strange bit of xhtml
>     <html:a  href="...">zzz</html:a>
>  </html:p>
>  <OMI>1</OMI>
> </OMA>
> 
> would (assuming appropriate namespace declarations) validate against the
> OM schema.
> 
> in otherwords there would be virtually no checks done at all by schema
> validation, and more or less any random bits of well formed XML could be
> included and it would still validate. "OM Validation" would have to be
> specified as a layer above DTD/schema validation.
> 
> One could define special purpose schema which replaced xsd:any by
> exactly the schema corresponding to the CDs that you want to support,
> this would give a finer level of control than we have now, as then
> schema validation would check that symbols used were in the CDs
> specified, but it would require modifying the schema for each set of CDs
> supported. 
> 
> Some days I think the benefits of this change outweigh the costs, and
> some days I don't (I'm not sure which kind of day it is today)
> 
> David
> 
> ________________________________________________________________________
> This e-mail has been scanned for all viruses by Star Internet. The
> service is powered by MessageLabs. For more information on a proactive
> anti-virus service working around the clock, around the globe, visit:
> http://www.star.net.uk
> ________________________________________________________________________
> --
> om at openmath.org  -  general discussion on OpenMath
> Post public announcements to om-announce at openmath.org
> Automatic list maintenance software at majordomo at openmath.org
> Mail om-owner at openmath.org for assistance with any problems
--
om at openmath.org  -  general discussion on OpenMath
Post public announcements to om-announce at openmath.org
Automatic list maintenance software at majordomo at openmath.org
Mail om-owner at openmath.org for assistance with any problems