Re: SAX2: When to resolve System Ids

Michael Glavassevich <[email protected]> Wed, 31 Mar 2004 11:34:53 -0500
Newsgroups gmane.text.xml.sax.devel
Message-ID <OF582731A7.4BB84AAC-ON85256E68.0058D718-85256E68.005B15DA@ca.ibm.com>
"Karl Waclawek" <[email protected]> wrote on 03/31/2004 11:06:52 AM:

> > It's not a relative URI if it's already resolved. What if your 
application 
> > wants to use a different base URI instead of the one the parser would 
use? 
> > If the URI has already been absolutized you cannot do this.
> 
> True, but how often does that happen? And if one needs that feature,
> why should one need it only for externalEntityDecl(), but not for
> notationDecl() or unparsedEntityDecl()? I am not doubting your
> arguments, just trying to get a handle on the real world impact.

True, but what percentage of DTDs outside of conformance test suites 
actually contain notation declarations and unparsed entity declarations. I 
know of a few real world DTDs which contain notation decls, Docbook for 
instance. There are probably many more I'm not aware of. Just because 
something is rare doesn't mean it's not of importance to someone. It's 
hard to get a handle of real world impact because happy developers don't 
often share how they're using the technology with the community. You do 
however tend to discover who's using a particular function and why it's 
important to them when it becomes broken, but even if there's no response 
to the change it doesn't necessarily mean everything is okay.

Michael Glavassevich
XML Parser Development
IBM Toronto Lab
E-mail: [email protected]
E-mail: [email protected]