RE: Hmmm.... future of cons?

[Nadim Khemir]

> -----Original Message-----
> From: Gary Oberbrunner [mailto:address@hidden
> Sent: Tuesday, May 21, 2002 10:25 PM
> [snip]
> Sounds like what we need are:
> 
> - all recent patches
> - any doc improvements
> - any fixes people have been hoarding
> - Cons::Plus?  Someone will need to figure the best way to integrate
> this.  Should we just shovel it into the release package as is, or do
> something tighter?

Then minutes later.

> -----Original Message-----
> From: Gary Oberbrunner [mailto:address@hidden
> Sent: Tuesday, May 21, 2002 11:48 PM
> 
> OK, Doug's plan sounds good to me.  No new features, just a few fixes if
> anyone has them.  It sounded like David Sanderson had a couple; David,
> do you have fixes for those bugs or just the bug reports?
> 
> Perhaps we could include Cons::Plus, but only in a contrib/ dir, to make
> sure people get it, but we don't have to do any more work?
> 
> -- Gary O

Gary (and all the other involved), I definitively understand that you have
other important things in your life and please do not take no offence in the
following (it is as much for me as it is for you).

When I wrote the mail that started this discussion I also had in mind
branching or starting a new project. If you read the mail chronologically,
you'll see that we go from: a new Con, consider Scons, fix Cons (your first
mail above) to fix as little as possible and continue as we did before. We
use an efficient damping filter here.

Do not misunderstand me, fixing a new release integrating the minimal
patches and update the web site is very good and certainly the thing to do
first but it's the wrong answer to this thread "future of cons".

As I said before cons is sick of "open sourceness", plus many think that it
is good as it is and those persons are key persons, they know the code, the
know the possibilities and the limitations and they do fine with them. Funny
this reminds me of my favorite project, I have my own editor, it work fine,
there are some bugs lurking around but I like it the way it is and I can
always dig a bit and fix the small problems. Still when I think about it
might be time to re-start it or leave it to someone else. I feel the same
with Cons.

Cons is good and the job you guys did is great but we must think future so
here are my question again:

- Who leads?
- Keep old code or start new?
- Scons, what do we do with it?
- Where are the requirements?
- Where is the design?

I can help with some points, I am in and will invest time in this project if
we get things sorted out.

One more time this thread is about the future of Cons not it's present.

After the borring stuff I'd like to answer my own questions (some of them at
least)

Language: I'd like to keep perl as dev. Language, I'm fluent in perl and
Python is just too boring (that is MY opinion) and there is quite a lot of
construct files out there. Now the good new is that nothing stop us from
having both languages (perl can call python and I'd guess the opposite works
as well) even though I'd rather have a perl only solution.

Money: there have been quite a lot of discussion about financing; some say
they will finance other that things would happen if there is money in the
path. I really dislike both of them. My energy level drop to
"watch-tv-instead" level as soon as money is involved.

Requirements: Generalize Cons (Make shines here), parallel cons (I mean
multi machine build not single machine multiple processors)

Code: I am for a complete rewrite (possibly using some of Cons and SCons
code). Shorter, easier code (so even new comers can help quickly). No 6000
lines blob and certainly no user documentation included in the code. I have
nothing against "generating" the final perl script with documentation
included. Documentation for the strategy used in the code is a must. Get our
heads out of the sand and see if there are modules that can do some of the
job. Cons is quite self contained and that is good, up to the point where it
is uselessly self contained.

Backward compatibility: If a re-write is to be done I'd like to have the old
scripts work if they do not force any structural thinking in the new design

My 0.20 Sek (that's 0.02$)

Nadim..