ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Peter Donald <dona...@apache.org>
Subject Re: AW: So, The Show Must Go On
Date Wed, 17 Jan 2001 14:34:16 GMT
At 12:34  17/1/01 +0100, Christoph Wilhelms wrote:
>I hope this is a joke (for crying out loud)!
>
>
>Unfortunately it seems to be no joke :-(.

It is unfortunate when anyone leaves, especially in this kind of situation
- even more so when he was original developer of Ant but as he said break
point was reached. I think he will be missed in long term (maybe not the
short ;=]) so I hope he still sticks about ;)

>Can't we just do a good job on refactoring Ant 1.x?

yes and no - up to a point. But then we start breaking backwards
compatability. It is painful to track a moving target. With Ant2.0 wehave
an opportunity to start fresh and not put our users through hell as we
evolve - it will hopefully be more stable (once finished) and won't cause
any difficulties in migration till at least 3.0 ;)

>When this is done (or even parallel) can we discuss
>visions on Ants future as "grown up" OO-Developers?

It will be done in parrallel and if you recall few of the technical
discussions were lacking - it was just the other stuff ;)

>It came clear to me, that code-based dicussion is
>worth NOTHING - You can punch me for that, but it's
>my opinion! 

code-based discussion can be worthwhile. Not everyone has formal training
or knows how to share their ideas with others. I know some damn brilliant
programmers who were self-taught and think "Design Patterns" are things you
see on bathroom tiles. However they are also brilliant at designing
structures - they have actually used many of the patterns (ie
factories/singletons, chain of command, decorators, IOC, SOC whatever)
without even knowing that they were doing it ;) Code is just an easier way
to express things for some people and being opensource rather than company
based we should have to be slaves to the design gurus (even thou I like
that ;-])

>If I could I would delete ALL proposals
>from the CVS until we have a rather complete Requirenment
>Spec and Design Document which ALL of us agree to...

Well the basic "features" were largely agreed to - I actually listed quite
a few in an email a while ago (damned if I can find it in the archives thou
;( ).  We may not have formalized it but I think there is enough shared
mindspace for it to continue - thou it would be good to formalize.

Strangely enough if we hadn't of had source based proposals (specifically
AntFarm) we would have completely missed out on some brilliant ideas -
mainly because we would think - pft too difficult/complex/whatever -
without giving it real thought.

Notice that AntEater and Myrmidon (both developed by ant-dev members) were
basically incremental advancements over ant1.x with more structured models.
It wasn't until AntFarm and Frantic came along that we started to think
outside the box so to speak. I like AntFarm and will strongly +1 it's
approach thou not sure about frantic yet.

Regardless of whether or not the source is actually used it was of benefit.
The "discussion" may have been triggered by the addition of other proposals
but it would have happened anyway given the personalities/attitudes of
different people. In a way it is better that it happens at the start rather
than mid-stream.

>I think priority, for now, should be Ant 1.x, not only
>because Ant2 discussion was on cost of Ant1 developement.

Well the two main contributors to Ant are dedicated to Ant1.x so I don't
think we have anything to fear from that ;)
Cheers,

Pete

*-----------------------------------------------------*
| "Faced with the choice between changing one's mind, |
| and proving that there is no need to do so - almost |
| everyone gets busy on the proof."                   |
|              - John Kenneth Galbraith               |
*-----------------------------------------------------*


Mime
View raw message