nestable C/C++ XML parser?

Paul Miller stele at
Tue Dec 7 18:19:21 GMT 1999

> Sure, I too see a need for this, and I've even implemented it.
> However, this is something completely different from doing parsing on
> behalf of the parser. Parsing is turning a stream of bytes (or
> characters) into something higher-level, but this is not what you are
> talking about.
> As far as I understood him, the original poster wanted to do the
> parsing (that is, the reading and interpretation of bytes/chars) on
> behalf of expat.
This is more or less correct. I want to use XML as an application data
file format. Why? Two primary reasons:
1. I don't need/want to invent a new syntax - I like XML just fine and
it handles object-oriented nesting of data quite nicely
2. I can publish a DTD and make it easier for my end-users to use my
application data in their own applications (I work in special effects
applications, and certain high-end customers like to use my data in
their own custom tools) without doing a lot of hand-holding

Whether this constitutes a "good enough" reason to use XML I don't know.
The primary use of XML seems to be web-oriented e-commerse stuff, of
which I don't give a hoot about (I'll leave that stuff to the web

Given my needs, I know the data in the XML file, and I know what to do
with it once I get to it. But I *do not* want to go with the huge
complexity of DOM. I've indicated in a previous thread the kind of API
I'd like to access the data. I was hoping expat would let me do nested
parsing, but it doesn't.

Frankly, for this kind of application file format stuff, validation and
namespaces probably aren't really necessary, but I want to use the XML
syntax mostly because it's well defined. This means I'll probably have
to implement my own restartable low-level "parser" which just deals with
the syntax and the basics. I was hoping to layer on expat, for no other
reason than to gain free validation once expat gets it, but considering
my needs this probably isn't necessary, and it's just a bit more work
for me.

>From this (and other discussions) it looks like this type of XML parser
for application data would be generally useful (in the C/C++ community),
so I'll be sure to make my efforts available.

Paul Miller - stele at

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at
Archived as: and on CD-ROM/ISBN 981-02-3594-1
To unsubscribe, mailto:majordomo at the following message;
unsubscribe xml-dev
To subscribe to the digests, mailto:majordomo at the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at

More information about the Xml-dev mailing list