Arnaud Le Hors wrote:
Do we have any idea how many people actually implement components?
Granted, I'm a bad test case since I helped design
the framework, but I've used the parser configuration
a lot for various tasks. Examples include: an HTML
parser, a DTD to XML converter, and a configuration
that allows RelaxNG validation in the pipeline. So I
feel the pain everytime that XNI changes because I
have to change *everything*. And not only that, I
have to do it right away because old versions of my
code won't work with the new version of Xerces.
I have heard of a few people that are writing their
own configuration. But for the most part, you never
know how many there really are because there is no
requirement for them to tell us. I know I'm always
surprised when I type "NekoHTML" into Google because
it's being used all over the place.
In the real world, I think most people will code to
standards such as DOM or SAX for the very reason that
they don't want to tie their code to a specific impl.
But when they need more power and flexibility, then
they may code directly to XNI. And any breaking
change we make has serious consequences for them.
In short, I prefer non-breaking changes not only
for myself but also for all the known and unknown
users of XNI.
--
Andy Clark * [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
