[XML-SIG] Developer's Day position paper

[Thomas B. Passin]

Paul wrote:


> I agree with everything Skip says. We need something as fast and easy as
> xmllib. We also need something compliant to the Unicode and XML
> specifications. Does anyone disagree? Does anyone think that these goals
> are mutually exclusive?
>
> --

I agree - easy and reasonably fast.  There's lots of jobs out there that use
small xml files and want to be easy to get off and running. xmllib is good
for that.  Also agree on Unicode and the XML standards (I think we can wait
awhile longer for the dust to settle on xml-schemas, though).  I think that
an XPath processor would be important since it could be the basis of any
number of query processors.

Another thought, does anyone else think this should count for anything?
That is the subject of compatibility with JPython.  Right now, xmllib (minus
the c-based parsers) ought to work with JPython, but the 4thought suite
won't, since they need to compile C stuff with bison, etc.  I think we ought
to have a basic library that's easy to use and will work with both flavors
of Python on any machine.  (On the other hand, JPython should make it
relatively easy to work with all those nice Java products).  This leads me
to think that we shouldn't rely on the 4thought suite as the **sole**
processors in the library.  Anyone want to add some thoughts to this?

Tom Passin