I'd say make PSVI part of an expanded/extended API that is
optional. We don't have nearly enough experience with this
to be able to say what a good design for this will be.
Ted
----- Original Message -----
From: "Andy Clark" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Tuesday, July 03, 2001 6:45 PM
Subject: [Xerces2] PSVI
> PSVI: "Post Schema Validation Infoset"
>
> This is going to throw a big wrench into the works, I fear.
> And what's worse is that as new standards are created that
> augment the document's infoset, there's no standard way of
> layering (or representing) the accumulation of these
> augmentations. (It's interesting to see that other schema
> languages like "Relax", "Trex", and "RelaxNG" are avoiding
> infoset augmentations entirely -- even disallowing default
> attribute values.)
>
> Anyway, we need to think about how such augmentations are
> going to affect XNI. Is PSVI something that should be part
> of the API, part of an expanded API, or out-of-band data?
>
> Anybody with ideas?
>
> A very simple case that I'm thinking of is defaulting
> element values. You can determine if an attribute value
> was specified in the document but you can't do the same
> with element values. In the case of passing character
> data, we do have the option of expanding the contents of
> the XMLString structure to hold additional information.
> Perhaps by using bit flags, even.
>
> I'm sure there are a *lot* of other places where we want
> to pass additional information. For example: people would
> want to get the datatype and content model info (and
> validators) for elements and attributes that are passed
> through the pipeline. But right now we don't have direct
> access to this stuff via XNI. Is this even a problem,
> though?
>
> --
> Andy Clark * IBM, TRL - Japan * [EMAIL PROTECTED]
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
