[XML-SIG] How to proceed

[Lars Marius Garshol]

* Lars Marius Garshol
|
| saxlib
|   I plan for this package to contain lots of SAX 2.0-related
|   utilities, like DOM2SAX walkers, XInclude and XBase filters, more
|   advanced parser instantiation tools, more drivers etc.

* Martin v. Loewis
| 
| It sounds like a good idea to offer more functionality in
| saxlib. However, we have to be careful that PyXML continues to be a
| strict superset of Python 2.0. 

My idea for saxlib is that it should be a toolkit with SAX 2.0-related
add-ons. I didn't really intend for it to contain SAX 2.0 itself, just
useful drivers, filters and similar kinds of utilities.

| To achieve that, I'd like to see the 2.0-provided functionality be
| split-off before adding more stuff. 

Agreed.

| Then, it should be possible to use saxlib in a line-by-line
| identical form with 2.0 (e.g. by having xml.sax.saxlib20), and still
| provide extra functionality in xml.sax.saxlib.

Well, whether this belongs in the xml.sax package or not is unclear.
It's not part of SAX as such, just utilities built on top of SAX.
 
| I'd also like to hear opinions on *where* this functionality should
| be located - if it is not clearly specific to SAX, xml.sax may not
| be the right place.

It would be clearly SAX-specific.  However, much of it would also be
usable with the DOM as well.  For example, you might use the filters
to transparently perform XInclude processing as the DOM tree is built.

xml.saxlib is probably a better location for it.  Or xmlplus.saxlib.
Or whatever.

--Lars M.