CC'ing axkit-users too - sorry if you get this twice.

This is just an update instead of seeing all these random CVS checkins...

I'm working on two things at the moment, which I'll go through in turn.

1. Migrating AxKit into Apache.org.

My god this is a herculean task. They ignore me and ignore me and ignore
me. I guess it's because we're not Java, therefore we're not cool. Or
maybe people are just busy. Either way, I'm ignored until I get some
insider like Ask Bjorn Hansen on the task (who has been a godsend every
time I've asked him to help). But I'm getting there. I'm now waiting for
them to create a CVS repository and accounts for some initial committers.
If you feel like you should be or want to be an initial committer on the
AxKit project and I haven't contacted you already, then please contact me
personally - we certainly have a need for a few good men - one area we're
definitely missing on is a release manager (1.5 didn't update "Changes"
in case nobody noticed), and a web site manager (since everything will
have to change to the Apache look and feel).

The biggest thing I want to see changing in the move to Apache.org is I
don't want AxKit to just be "Matt Sergeant's project" any more. I don't
want to orphan it, I just want it to grow without needing me.

2. AxKit B2B

OK, so you've seen the CVS commits. What the hell is this thing?

It's basically just me hacking. Yes, it's really that dull. But it could
turn into something cool. It aims to solve a few problems that I think
AxKit as it stands doesn't solve terribly well:

 - it's still a bitch to install, despite my best efforts. I still think
the biggest obstacle is mod_perl.
 - it's not that great at web apps. XSP in AxKit is *much* better than
Cocoon's version for this, but its still a fair way off - you can't drag
and drop a new application into a system and expect it to "just work".
 - it's still overly complex to configure a transformation pipeline, and
managing these apache config directives is really hard (did you ever see
the merging code in axconfig.c? well it's not entirely complete/correct,
but you'd never know). Why don't we use XML?
 - I really wanted to play more with POE. I like the idea of a single
process providing threading via cooperation. I've now played with POE
quite a bit and I still like it.
 - AxKit doesn't solve B2B problems for people spending megabucks on B2B
integration systems. We can solve those with SAX generators and "stuff",
and provide a much cheaper alternative for Small/Med size companies.
 - AxKit's design is over complex. It special cases too much. Sometimes
it's best to go back to the drawing board. That doesn't mean that AxKit is
at the end of it's life though - I think both bits of software will have
separate lives for separate purposes. Besides, it's open source. I
couldn't really kill it if I tried.
 - Following on from the last point - I want to try doing a complete SAX
based system (yes, like Cocoon2). Not for any performance reasons, but
because I keep harping on about how great SAX is, and I need to put my
code where my mouth is.

Anyway, so that's what AxKit B2B is all about. I know people were curious
about the checkins. The code doesn't really run yet, so don't bother
trying to check it out unless you're really going to hack on it (and
likely I'd already know who "you" are, if you were). I'll let you know
when it can do something useful. Probably in the next couple of weeks.

-- 
<!-- Matt -->
<:->Get a smart net</:->


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to