OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   RE: [xml-dev] Question regarding PSVI..

[ Lists Home | Date Index | Thread Index ]
  • To: "Murali Mani" <mani@CS.UCLA.EDU>
  • Subject: RE: [xml-dev] Question regarding PSVI..
  • From: "Dare Obasanjo" <dareo@microsoft.com>
  • Date: Wed, 29 May 2002 18:21:50 -0700
  • Cc: <xml-dev@lists.xml.org>
  • Thread-index: AcIHU5aV0K9bYTnNRiOtztJWip8B4wAHQrKg
  • Thread-topic: [xml-dev] Question regarding PSVI..

> -----Original Message-----
> From: Murali Mani [mailto:mani@CS.UCLA.EDU] 
> Sent: Wednesday, May 29, 2002 1:58 PM
> To: Dare Obasanjo
> Cc: xml-dev@lists.xml.org
> Subject: RE: [xml-dev] Question regarding PSVI..
> Dare, agreed with what you say about Schema Infoset, actually 
> I also said that it is pretty much useless w.r.to what we 
> were discussing.
> I have some questions regarding PSVI:
> a) specific question: In section 3.3.5 -- there is a 
> statement which says something like "The first (item 
> isomorphic) alternative above is provided for applications 
> such as query processors"... hmm.. actually what is meant by 
> the first alternative -- is it going against the "intuitive" 
> determinism which one might expect???
> Anyways, my actual question is this --
> Is there any application for the type definition associated 
> with an element information item in the PSVI?? My concern is 
> again querying. Now if we assume that every query always 
> tries to start from the root and select a set of nodes based 
> on a path expression, then what we need is type inferencing. 
> The type associated with an element information item in PSVI 
> seems useless here..
> Also, consider element declaration -- that might be needed 
> for checking integrity constraints like how XML Schema 
> defines them. The following may be slightly vague -- there 
> are two other ways of specifying integrity constraints -- 
> using path expressions, or using type names. If path 
> expressions are used, then we do not need element 
> declarations, or type names for element information items. If 
> we use type names, then we need type names for element 
> information items, but I do not think we need to know the 
> element declaration for an element information item.
> Anyways, my question is -- where will the type names and 
> element declarations associated with an element information 
> item in PSVI be used?

I've seen user scenarios where both are needed. 

1.) The type names for element declarations are useful for type based
queries like those that will exist in Xpath 2.0 and XQuery. 

 EXAMPLE: /myco:company-roster/myco:employee[. instance of element of
type myco:managerType] 

where myco:managerType is a complexType defined in the schema. 

2.) Currently MSXML ships with the ability to obtain the Schema Object
Model (SOM) object that represents an element or attribute declaration
from a DOM instance. Our users love it and we are inundiated with
requests to provide similar functionality for the .NET frameworks
version of the DOM. One user scenario involved creating a schema-aware
XML text editor and using the element decls for the particular element
being edited to determine whether the edits were valid or not which was
way more performant than validating on each edit. 

If two wrongs don't make a right, try three. 

This posting is provided "AS IS" with no warranties, and confers no


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS