Lists Home |
Date Index |
We do something very similar to this, and it is really useful in the context
of a processing app. Basically you parse in a document with its associated
schema and created a kind of "Extended DOM" (we call it ExDOM) in memory. We
built the interfaces so that standard DOM calls do what you would expect,
but additional extended calls give you access to schema information,
including more information about structure (we use intermediate nodes in the
DOM trees to distinguish between different content models than might
underlie the same concrete instance).
This all makes sense to me, but it should be part of the DOM effort (i.e. an
extension to a programming API). I don't understand why you would ever want
to serialize this as an XML document. I also feel that working on a PSVI
instance rather than an XML instance should be out of scope for XPath and
other XML-focused (as opposed to programming language-focused) standards...
there seems to be a lot of agreement on this.
Having a PSVI-like implementation in memory would enable an XPath processor
to do some potentially useful things (the only one I can think of is type
checking). But this should be *completely* invisible to everyone but the
author of the processor. Also, your comments about the need to apply
potentialy invalid XPaths are extremely convincing... I would be inclined to
agree that any notion of strong type checking should be layered on top, with
XQuery being the obvious place to do this.
> -----Original Message-----
> From: Rick Jelliffe [mailto:firstname.lastname@example.org]
> Sent: Friday, May 10, 2002 12:00 PM
> To: email@example.com
> Subject: Re: [xml-dev] PSVI formalization
> From: "Matthew Gertner" <firstname.lastname@example.org>
> > That said, you led me down the garden path with this PSVI stuff. My
> > understanding of the PSVI (very shaky, I admit) is that it
> is an abstract
> > formalism for representing an instance with its associated schema.
> A Post-Schema Validation Infoset (instance) is an XML Infoset
> that has been augmented with information as a result of being
> validated by a W3C XML Schema (schema).
> This information could include, among other things,
> - defaulted values
> - type information
> - whether validation was attempted, and whether it was successful
> A PSVI does not necessarily conform to a schema. The augmentations
> as a result of schema processing could be that some elements
> are invalid. (This is why optimizing away "impossible paths" should
> be an application-dependent thing, or a distinction between XQuery
> and XPath2. An application may be interested in part that are invalid
> in order to repair them, report them, or complete the document.)
> A PSVI can be constructed in several ways, not merely by running
> the same schema on the whole document at the same time.
> Rick Jelliffe
> The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
> initiative of OASIS <http://www.oasis-open.org>
> The list archives are at http://lists.xml.org/archives/xml-dev/
> To subscribe or unsubscribe from this list use the subscription
> manager: <http://lists.xml.org/ob/adm.pl>