xml-dev - Re: SAX2 Lexical Handler Suggestions

[David Brownell <david-b@pacbell.net>]

> I can see that you come from the same camp as most XML programmers I've
> encountered, which I would sum up as "XML is for documents, why would you
> want it to do anything more". 

I'd put it differently:  SAX is a "Simple" API for XML ...


>     Our system is a web based authoring system

... and it's quite well known that editors require quite
a lot more than "Simple" functionality.  In fact they
usually demand all kinds of extra information which the
XML infoset explicitly discarded using the 80/20% rule.
(IMHO more lie 99/1 in this case.)

After seeing bits and pieces of editor requirements for
several years now, I think that if SAX is to support
that functionality it should be through specific APIs
(EditorHandler, maybe) addressing editor issues.  If your
itch involves having such features, I suggest you scratch
it in that way.  (I don't like how DOM cluttered up the
80% simple case with incomplate support for someone's 20%,
which btw is often labeled as "editor" stuff, and want to
see SAX stay uncluttered.)

It would need to expose at least part of a token level
event stream ... so that editors would not play with the
whitespace separating attributes (or outside of the root
element), rearrange attributes, change text encodings for
documents, and so on.


>     As
> far as knowing the actual attribute's raw text I need some way to match
> which attributes were entities

"contained" ??

>     that were expanded by the processor and which
> were entered by the user and having the information about which are default
> attributes and which were actually from the XML document is very handy to
> save space in the downloaded HTML.  I know that SAX parser absolutely can
> handle these changes, since I have a modified version of the AElfred2 parser
> that currently supports them.

Hmm ... which version is that?  My version ("original" AElfred2")
didn't do that, last I recall.  Please forward patches ("diff -u").

- Dave