RE: [xml-dev] When to Validate XML?

[Leigh Dodds <ldodds@ingenta.com>]

> -----Original Message-----
> From: Ajay K sanghi [mailto:sanghi@giasdl01.vsnl.net.in]
> Sent: 01 November 2001 06:12
> To: Max Dunn; 'Magick, Brian'; xml-dev@lists.xml.org
> Subject: RE: [xml-dev] When to Validate XML?
>
> I think an important question to ask before deciding to turn on validation
> at DTD/Schema level is "How can I report back the appropriate
> non-conforming error in the XML file to the sender?".
>
> If my application has better reporting capability then I would turn off
> DTD/Schema validation. If in conjunction, I can use DTD/Schema
> validation as "invariant", nothing like it. And of course, if I am able to
report
> non-conforming error using DTD/Schema validation then I would turn it on.

This is an interesting point. Schema languages are increasingly
powerful in terms of the validation rules that they can express, but
are their error reporting capabilities increasing in step? One could
argue that this isn't a function of the schema language, but of
the processor.

However, it seems like a declarative way to associate readable text
messages for a user, or pointers to additional documentation would be a
useful feature. (A user here might be another application, and the pointer
may
be to a series of data cleansing routines).

So, do any of the existing schema languages support this?

Well Schematron does (one of the reasons I like it so much), and RELAX NG
and W3C XML Schema both have an annotation mechanism. So error messages
could be used to annotate the schema. I wonder if there's scope
for standardising on this annotation syntax? Could be useful seeing as
multi-schema processors are now appearing.

Anyone else thought about this?

Cheers,

L.


--
Leigh Dodds, Research Group, Ingenta | "Pluralitas non est ponenda
http://weblogs.userland.com/eclectic |    sine necessitate"
http://www.xml.com/pub/xmldeviant    |     -- William of Ockham