Re: ANN: SAX Filters for Namespace Processing

["Simon St.Laurent" <simonstl@simonstl.com>]

On 31 Jul 2001 12:14:00 -0700, Ronald Bourret wrote:
> Even as I was writing my "you shouldn't do this" comment, I knew it
> wasn't completely valid. Telling somebody they can or can't use a
> product in a certain way is always a debatable thing to do. I think what
> upsets me is that:

In this case, I wish this usage had been considered during the
development of the namespaces spec and dealt with explicitly.  The
problem seems to be that a debateable practice was permitted (not
forbidden) by that spec, and that subsequent specs have provided
explicit support for that practice. It's very difficult now to change
that.
 
> 1) Unqualified child elements strikes me as a bad practice, period, so I
> don't want to do anything to encourage it.

I agree that it's bad practice.  I'm considering this bit of code as an
exercise in coping, not an exciting venture into new terrain.

> 2) Simon's solution needs to be applied carefully, since it has a very
> real potential for name collisions.

I agree completely.  That's a large part of why I went to the trouble of
creating rules functionality that allows developers to apply it
selectively to particular namespace contexts.

I worry quite strongly about developers who mix qualified and
unqualified elements with the same local name in a single document type.
That seems like playing with fire, at best, and it's something that it's
very difficult to control.  I haven't seen any examples of that yet, and
I hope never to see it.