Re: Potential new issue: PSVI considered harmful

Hmm, I think Dare's notes are mostly addressing things I didn't say, 
which means that I wasn't clear enough.

> [Dare Obasanjo] I'd also add [to the PSVI]
>
> - addition of information regarding the success or lack thereof of validation of the elements and attributes information items in the infoset. 
> - addition of tables that show relationships bindings between identity constraints (key/keyref, ID/IDREF) for the element information items in the XML document (the W3C XML Schema REC suggests that these probably shouldn't be surfaced to applications) 

I see.  The first probably isn't interesting outside of the class of 
software that's concerned with validation.  As for the second, has any 
other application class proposed to use it?  Seems sensible enough in 
any case.  And my other comments apply to it... it's clearly orthogonal 
to validation.

> <Tim Bray> The problem is that we are making the old SGML error all over again. 
>  An
> SGML document can't be parsed at all without reading the schema (DTD),
> and the DTD conflated primitive typing, parsing support, entities,
> default values, and other stuff in a really messy way.
>
> [Dare Obasanjo] I'm not sure what this has to do with the PSVI.

The problem, once again, is the 3 letters "PSV".  It's reasonable to 
annotate a document with type information and possibly other stuff as 
you note above.  It's not reasonable to regard this as necessarily a 
consequence of validation.

> On the other hand XML 1.0 itself is tied very closely to DTDs and the attendant conflation of entities, default values, notations and weird primitive typing (NMToken, CDATA, IDREF???)  

Yep.  We should learn from our mistakes.  The notion of the PSVI is 
evidence that we're not.

> information, but the PSVI suffers from the following flaws:
>
> 1. the inclusion of default values.  These are sufficiently problematic
> that the IETF is about to recommend they not be used at all, and I for
> one think there is a good case that they should be deprecated for
> architectural reasons.
>
> [Dare Obasanjo] Can you expand on what problems default values in the PSVI cause and also how they differ from those caused by default values in an XML documents DTD? 

To the extent that an instance is self-describing and self-contained, 
that's a win.  Default values appear magically as a result of reading 
something else somewhere else.  That's bad.  Yes, defaults induced by 
DTDs are identical to, hence just as bad as, those from XML schemas.

> <TimBray>
> 2. the notion that annotation is necessarily linked to validation.  The
> problem here is with the "PSV" part of the name: there's nothing wrong
> with a Type-Augmented Infoset (TAI), but why link it to validation?

> [Dare Obasanjo] I can't tell what you are objecting to here. Are you arguing against the name "PSVI" and proposing an alternative mechanism for augmenting an XML infoset with type information to claim this name or...? 

Yes.  Actually, I'm not proposing any specific mechanism, just leaving 
the door open by decoupling the notion of an augmented infoset from that 
of validation.

> 1. Type-augmented XML is a good thing and a recommendation should be
> prepared describing it both at the infoset and syntax level. (I gather
> there is already some work along these lines in XML Schema?).  Serious
> consideration should be given to 80/20 points rather than simply
> re-using the plethora of primitive types from XML Schema.
>
> [Dare Obasanjo] Proposing a new working group to duplicate the work of the W3C XML Schema working group needs more justification than you've given. In fact I really haven't seen any.  

Er, I didn't propose a new working group.  I do propose decoupling the 
augmented infoset from the process of validation.

> <TimBray>2. Type-augmented XML has nothing to say about default values created 
> in
> any schema.
>
> [Dare Obasanjo] I agree that default values and type information are orthogonal but they both exist in the PSVI because it is the Post Schema Validation Infoset which means it includes everything that can be obtained from validating an XML document. 

Sigh.  I'm arguing that both value defaulting and type annotation are 
orthogonal to the process of validation.  Yes, this happens to be 
produced as a side-effect of validation - but I can imagine doing type 
annotation in a much more lightweight way than bringing a large complex 
declarative schema facility to bear.  In fact, why shouldn't I just be 
able to jam something into the instance or infoset saying "this 
attribute here is an integer"?

> <TimBray> 3. Any software can create and/or use type-augmented XML, whether or 
> not
> any validation is being performed.
>
> [Dare Obasanjo] How exactly will this new system you propose avoid conflicting with W3C XML Schema? 

Where's the conflict?

> <TimBray> 4. Work on XQuery and other things that require a Type-Augmented 
> Infoset
> must not depend on schema processing, and should not have normative
> linkages to any schema language specifications.
>
> [Dare Obasanjo] I suggest discussing this with the XML Query Working Group before drawing up such proclamations. 

Indeed. -Tim