Thanks. I agree. 

It's just that Xerces is so close to being able to check everything
dynamically. There are basically three areas of checking:
- Whether a node has been inserted in a valid position in the structure
- Whether a simple type's value is correct (e.g. is it an integer
between 0 and 128)
- Are any identity constraints violated (duplicate IDs, etc.)

I have managed to do the first two quite easily using Xerces, but the
last one is difficult, because that part of the code in
XMLSchemaValidator was intended to be run just once. So, I'll need to
write something for this - probably using Xalan's XPath.

By the way, I am working on an open source XML editor - xerlin.org. 

Regards,

Evert


On Thu, 2002-10-10 at 15:06, Joseph Kesselman wrote:
> The DOM's architecture deliberately does not require continuous checking, 
> since it is usually _MASSIVE_ overkill -- most applications check the data 
> before they add it to the DOM, and it can be checked at serialization 
> time. DOM Level 3 plans to add the option to more intensively check a 
> subtree on demand.
> 
> That isn't to say a DOM _can't_ continuously check itself.  But that's 
> something of a special-purpose requirement, and probably doesn't belong in 
> a general-purpose implementation like the one in Xerces.
> 
> ______________________________________
> Joe Kesselman  / IBM Research
> 
> ---------------------------------------------------------------------
> 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]

Reply via email to