xml-dev - Re: Who will maintain SAX?

[David Megginson <david@megginson.com>]

Paul Tchistopolskii writes:

 [kind words about SAX omitted]

 > SAX2 is also not as reliable as SAX1 was 
 > ( I'm using the version from SAXON.jar - no 
 > real problems there, just some small problems, 
 > when systemID gets 'lost' e  t.c. ). 

Bug reports are welcome -- there are a couple of known bugs in the
SAX2 support classes already, and I'm going to put out a bug-fix
(2.01) release this fall; there will be no API changes, just
corrections to bugs in JavaDoc and support classes.

For the most part, though, the reliability issue is probably just a
result of the fact that the various parsers' SAX1 drivers are much
more mature than SAX2 drivers; that will change in time.

 > I don't know what was the idea behind 'comments are optional' in
 > SAX1. I can't even *guess* what was the idea.  My guess is that
 > five ( six? two ? ) developers decided "ah - usually people don't
 > need access to comments".

We had a long discussion over this during the initial design, and
opinions were split.  The final rationale is that SAX1 was designed to
be a production API (XML to something else), so it needed to preserve
only the logical content of the document.  A lot of the purely lexical
stuff -- CDATA section boundaries, comments, whitespace inside tags,
original order of attributes, etc. -- was omitted to keep SAX simple.

We also wanted to help avoid the mess in the HTML world, where people
put significant information, such as programming code, in comments;
the fact that that is not a common practice in XML today may be partly
a result of our original SAX1 design decisions (XML-Dev'ers can all
pause here to pat themselves on the back).

Everyone has one pet lexical feature that was left out (I actually get
about three pieces of mail mentioning CDATA section boundaries for
every one mentioning comments), but on balance,


All the best,


David

-- 
David Megginson                 david@megginson.com
           http://www.megginson.com/