You are viewing a plain text version of this content. The canonical link for it is here.
Posted to c-dev@xerces.apache.org by "Perry A. Caro" <ca...@Adobe.COM> on 2001/04/27 20:05:50 UTC

Re: PLEASE HELP: SAX parser errors on validation when ValScheme=Val_Never?????

Dean Roddey wrote:
> 
> As a practical matter, DTDs are not just about validation. They cause a lot
> other stuff to happen. In most cases, people who are not validating still
> very much want it to be processed and want to know if anything goes wrong.
> So don't get too tied up on DTDs being just a validation mechanism. Some
> people think thats a bad thing, that DTDs aren't just about structural
> validation, but that's the way it is. I know that, even if I don't need any
> structural validation, I'd like default and fixed attributes to be
> generated, for entity references to be recognized and expanded, notations
> that I've defined to be recognized, ID/IDREF stuff to be checked, the types
> of attributes to be checked against their definitions, etc... None of those
> things are related to how many X's can be inside a Y, but they are all
> driven by the DTD. And I don't necessarily want to have to do validation in
> order to get these things to happen.

All of that is true, particularly for XML users who come from the SGML
world, but since we're talking about practicalities, there are a large class
of XML documents in the real non-SGML world which have the following
characteristics:

1) The document entity contains no entity references, except for character
escape entities.

2) The document entity has no need for ID/IDREF enforcement, notations, or
other non-validation but declared/defined in doctype stuff.

3) The document entity is defined by a doctype with the following format:

<!DOCTYPE Name PUBLIC PubidLiteral SystemLiteral>

The DOCTYPE is intended primarily to declare by unique name the type of XML
document that this instance is. Validation is encouraged, but not required,
often because the XML is machine generated and presumed to be valid.  The
SystemLiteral may not resolve to an actual Internet resource, it's just
another form of unique id!

I can attest to the fact that lots of my users want to process XML documents
of this type with validation turned off.  How do I know?  Because this is
the number 2 most complained about problem with Xerces-C that I hear from
the users I support (number 1 is the lack of an XML output formatter for the
DOM).

Part of my job is to promote Xerces-C as a standard for our XML technology,
but users often put me on the spot when Xerces-C does not compare favorably
with other competing technologies, like James Clark's expat.

Perry

---------------------------------------------------------------------
To unsubscribe, e-mail: xerces-c-dev-unsubscribe@xml.apache.org
For additional commands, e-mail: xerces-c-dev-help@xml.apache.org