Welcome to the Xerces developer's list and in particular to the
Xerces C++ parser project.
Here is a quick overview of what is included with the code, what
current state is, and what I see as the next things most needing
to be done.
Xerces-C is a validating XML parser, written in C++, and
including both DOM and SAX APIs. The code is based on IBM's
XML4C version 3, which is a newer version than has been available
from IBM's website.
Here are some of the new things that have been added since
the last release from IBM:
o Plugable validator. In preparation for supporting XML schemas,
the XML validator has been made a plugable, replaceable module.
o Better conformance with the XML specifications.
o Incremental parsing. Allows application code to request just
the next item from the parser, rather than initiating a
complete parse of an XML document.
o Error message texts made common with the Java parser.
Error message handling structured for convenience in translation
to different languages, and for easy support of different
message loader facilities on different OSs.
o Abstract character set transcoding interface, allows
different platforms to more easily port to their preferred
transcoding implementation.
o XML name space support, both in the parser and the DOM
What's next, short term
o Stability and performance. Much of the code is very new.
The existing new features makes an interesting set,
and I would like to polish them into a good stable release.
Longer term (my short list)
o XML schema validation. Timing of this depends somewhat on
how the schema proposal from the W3C develops.
o Implement the rest of DOM level 2 and SAX 2.
o DOM memory footprint. The DOM is intrinsically an in-memory
representation of an XML document, so storage requirements
for large documents are always going to be a problem. Even
so, the current implementation uses more than it really
needs to.
Your ideas here.
If you have an interest or opinion on where the project is
going, or, even better, wish to contribute to it, this is the
place to be.
-- Andy Heninger
[EMAIL PROTECTED]
