avalon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Nicola Ken Barozzi <nicola...@apache.org>
Subject Re: Here is where we are divided
Date Tue, 03 Dec 2002 08:08:30 GMT

Berin Loritsch wrote:
> It appears that we are split down the middle on whether we want a
> clean slate or build on what we have.  Part of the problem is because
> we are afraid of what it would take to come to consensus.  As a result
> there are some potentially really cool things that we might be able
> to do with a clean slate that we might not see.  There is also the
> fear that we are abandoning our current users.
> 
> First, the reassurances.  None of us want to abandon our current users.
> We aren't all masochistic/sadistic/whateveristic.  That is why any
> brand new development absolutely requires a compatibility layer.
> 
> Next, the realities.  We have an existing framework that works.  It
> has a couple rough edges that we can't clean up because there is
> current software dependent on it.  They are not show-stoppers, but
> things that stick in some of our craws.  Any time we have new
> contracts that were not there before, we have a new version.  It is
> a question of degrees, but it is a new version nonetheless.
> 
> Framework Version 4.2 means we keep everything we have already.  It
> means we don't smooth the rough edges.  It means we don't do anything
> radical.
> 
> Framework Version 5.0 might be a smoking gun.  Emotions are high right
> now, which means that we might have to delay any hopes or dreams for 5.0
> even longer.  Keep in mind it will always be an emotional prospect.
> However our attention can't be devided when we are discussing it.
> 
> We either shoot for the moon, or we play it safe.

I have not yet voted on this, but it seems that there is a deadlock, so 
here is my opinion.

I think that complete rewrites are very dangerous, and must be done only 
as a last resort. They make code loose much of the information that's 
there, in bugfixes, patches and testing.

On the other hand, things must continue to go forward, and things that 
obviously don't work must be rectified and changed.

Avalon has still in minds of many a bad name about the heavy 
refactorings it did years ago. Avalon is stable, works well, but still 
developers remember what happened back then. I don't want this to repeat.

So this is my opinion: let's go for *AValon* (V==5).

                           *BUT*

Let's build on 4.1. That means *not* change packagenames, *not* change 
class names just for the sake of it, *not* create new classes when the 
existing ones can do with proper deprecation.

Let's bring out our issues and visions, understand what users need and 
we want; then take 4.1, package per package, discuss them, see what can 
be taken - and I'm sure it's most of it - and make AValon.

+1

-- 
Nicola Ken Barozzi                   nicolaken@apache.org
             - verba volant, scripta manent -
    (discussions get forgotten, just code remains)
---------------------------------------------------------------------


--
To unsubscribe, e-mail:   <mailto:avalon-dev-unsubscribe@jakarta.apache.org>
For additional commands, e-mail: <mailto:avalon-dev-help@jakarta.apache.org>


Mime
View raw message