Steven Ihde wrote:
Long term, is there any plan to implement StAX support in Xerces
(either as I've described, on top of burst SAX, or at a lower level)?
Anybody have any opinions in general about StAX?
I contributed to the Expert Group for JSR-173 (StAX) for
the Apache Foundation. However, I joined the group a while
after they had already set down the basic design. So there
was little room to redesign certain choices by that point.
I'm not that impressed with the StAX design. The problem
was that the group members had very different ideas on how
to design a pull API. So instead of settling on one they
decided to include both. The defense of the all-in-one
approach was that it was targeting J2ME which is reasonable
but I don't think it was reasonable enough to justify the
division and extra complication to users. Anyway...
I started designing my own pull API about a year and a
half ago but I never finished it. The design was completely
different than my NekoPull parser, by the way. Whereas
NekoPull was designed to be an XNI-level API, my ideas for
a pull API were more user-level in nature, like StAX.
Perhaps I'll dust it off and take another look.
As to your question, though, I'm not involved with the
day-to-day implementation of Xerces2 anymore but I have
not heard of any plans to implement StAX in Xerces2. But
maybe someone involved with the actual implementation of
the project can comment on that.
That being said, I'm sure the project would appreciate any
donation of code built on Xerces2 that implements the StAX
API. Regardless of my personal opinion, if it adds value
for the user, then it should be included. There's nothing
wrong with having a number of APIs sitting side-by-side --
DOM and SAX already do.
--
Andy Clark * [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
