Joerg,

I've agreed before with the importance of incremental improvements to 
the existing base.  However, Keiron has been the strongest supporter 
here of the principle of working from the existing base.  His work on 
the redesign is based, I believe, on his understanding of the existing 
design, his (and others') efforts to fix the problems over a long 
period, and his (and others') reluctant conclusion that some new design 
paradigms were necessary.

Do you believe Keiron's understanding to be false?  Do you believe that 
the same result can be achieved by incremental improvements?  If so, you 
may ignore the redesign efforts in good conscience.  If your conclusions 
are not so radical, then it seems to me that, when making major changes 
to production, you ought to make all reasonable attempts to keep changes 
in the production branch in sync with the redesign.  Your comment about 
the worrisome things that are still in the redesign underlines this.

Evidently, that means more work.  But before the redesign goes live, 
those changes are going to have to be made anyway, and that work will be 
particularly onerous if it has to happen all at once, rather than 
piecemeal by the person who writes the changes.

Given all of the above, "stealing" code from the unfinished redesign for 
the maintenance branch, to quote Jeremias, would be essential to make 
synchronisation in the other direction feasible.

Peter

P.S. As it happens, I agree that HEAD should represent the production 
code, and that development should take place on a branch.  In this case, 
though, it would probably just facilitate the isolation of the redesign.


J.Pietschmann wrote:
> Jeremias Maerki wrote:
> 
>> I'm not so happy about this one. I know I'm also guilty of doing more
>> work for the maintenance branch than for the redesign, but "stealing"
>> code from the unfinished redesign for the maintenance branch seems to me
>> like starting to take its breath away.
> 
> 
> The problem is that HEAD does not work, and even after
> Keiron gets som block and layout out of the door, it will
> be far from ready to release.
> 
> One thing which particularly bothers me is that much of
> the odd stuff is still in the redesign. Especially it
> does not all that much to reduce memory usage, in fact
> I have the feeling it will substantially *increase* it.
> I wish this Mark Lillywhite character were still around.
> 
> While there are other projects fighting with a HEAD
> representing "new idea" stuff and a branch with "tried
> technology" for the normal users to check out, IMO it
> should be the other way around: have always a working
> version in HEAD, make branches to develop new ideas,
> and integrate them into HEAD as they mature.

-- 
Peter B. West  [EMAIL PROTECTED]  http://powerup.com.au/~pbwest
"Lord, to whom shall we go?"


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

Reply via email to