Re: cons development

[Steven Knight]

Hello Dmitry, Dominique, et al.--

Sorry I didn't get a chance to reply sooner to this thread.

> I would be willing to commit some time to cons. I am more than full-time
> busy as it is, so it may not be wise for me to volunteer to be a project
> manager, but I might if no one else is willing.
> 
> What does Rajesh Vaidheeswarran think about it these days? Do you still
> stand by what you said in your August status update? Or has anything
> changed?

As far as I know, there is no change in status.

> How many participants are there in this project? Is there anyone who would
> want to be on the team? (I, for one, am volunteering.)  Anyone else who
> doesn't have time for doing work on cons, but is willing to help with
> suggestions?
> 
> I'd be particularly interested to hear from Steven Knight (if he is still
> on this list), since he seems to have been a big figure in cons, and is
> more recently a big figure in SCons. What's your opinion of cons, and of
> cons's future?

Right now, it's hard to be very positive about the future of Cons,
simply because no one has really stepped up to put in the time to get
the project back on track.  I think there is a decent amount of interest
out there in the Cons community that could be put to use turning Cons
around, but it needs someone to create a direction that people will
follow.

Here are the initial steps that I would take:

    --  Move the "official" web site from http://www.dsmit.com/cons/ to
        http://www.gnu.org/software/cons/ so updating the site isn't
        dependent on any one person or company in the future.

        (This should ideally involve having Rajesh put a redirector page
        at the dsmit.com site so people start going to gnu.org.)

    --  Update the new web site with the revised design that's been
        checked in for more than a year now.

    --  Start using the http://savannah.gnu.orf/projects/cons/ site
        to manage the project by translating the TODO list into the
        Project/Task Manager, and figure out the right way to integrate
        Cons with Gnats for FSF-blessed bug tracking.

        (Personally, I'd rather use the SourceForge-like bug tracker on
        Savannah--you can still see four old bugs that I entered back
        when they first set it up.  The FSF didn't like that because
        it's not Gnats, but as far as I can tell, they've never
        integrated Gnats into Savannah, which is a pain.)

    --  Fix the dependency bug in 2.3.0 and release 2.3.1 ASAP.

    --  Announce the release of 2.3.1 and the change in project
        management.

    --  Get people working on integrating Cons::Plus into the
        Cons proper; it's a very useful interface.

    --  Publish a list of to-do work and poll the user community
        on the next priorities.

As you can probably tell from the above, part of me *would* really like
to offer to take over as project manager, but between my day job, family
and SCons, I don't think it would be wise.  I'd probably fall into that
familiar pattern of an initial flurry of activity, followed by silence
when other things become higher priority.  The last thing Cons needs now
is a resurrection effort without any guaranteed follow-through.

I also don't have time for work on the Cons code itself; all of my
personal coding and documenting time goes to SCons now.  (If Cons came
back to life, though, I would volunteer to fix the 2.3.0 dependency bug.
It's somewhere in my code, anyway...)

I would still like to see Cons thrive, so the above notwithstanding, I'd
be very glad to help a new project manager in whatever other small ways
I can: advice, organizing, publicizing, etc.

Lastly, if Cons doesn't get back on track as a separate project, I
have a long-term plan to use the Perl Inline::Python module to put a
Cons interface on the SCons build engine.  We've architected a lot of
flexibility into the SCons build engine API, with the express purpose
of being able to support multiple interfaces in this way.  Among other
things, this would provide really *good* -j support right from day
one, because Anthony Roach wrote a very slick multi-threaded tasking
subsystem for the SCons build engine.

For starters, someone should volunteer to keep some kind of informal
list of people who might be willing to work on Cons, what they'd be
interested in doing, etc., so that whoever takes on active project
management has a roster to start with.

So how's about it?  Is there sufficient interest in resurrecting Cons?

        --SK