Re: [Axiom-developer] Re: Fork time

[C Y]

I think one way to look at this fork (which as I think someone observed
is already in some ways a de-facto fork) is as a proving ground for
ideas which can eventually fold back into Axiom.

As I understand it right now, the motivations most in conflict with the
current direction of the project are the ones wishing to attract more
developers and get the system more fully functional quickly.  I can
certainly understand those motivations and even agree with them to some
extent, but I think a fork around wh-sandbox might be the most logical
place to pursue them.

Steve and I appear to have some ideas which will result in something of
a re-think of the interp structure, and we also seem to agree that the
fully "literate as we go" approach makes the most sense - hence the
focus on making this as easy as possible.  Personally I'm not so
concerned with having "something working" in the sense of the fully
working CAS - I am more concerned with having a system that offers
something new enough and unique enough to draw people.  I think the
only thing that stands a serious chance of doing this for the
mathematical community is a system that has a chance of producing
formal correctness proofs of any of its answers upon request.  For a
system to be trusted to that level, every part of it must be subject to
straightforward audit by as many eyes as possible.  That means making
understanding all parts of the system as simple as possible, AS A
DESIGN GOAL.

I don't dispute the enormous usefulness of Waldek's work or argue with
those who label it as the best currently running Axiom system.  Indeed,
I'll likely use it myself.  I am tempted to vote yes to the proposal
with the caveat that only files which are not already truly literate be
updated, as I don't think there is going to be much practical
difference in untangling non-literate files with Waldek's fixes and
non-literate files without them. However, I personally like best about
Axiom it's focus on "pie in the sky" goals that have the potential to
change the way people think about computational mathematics and its
role in the wider mathematical and scientific communities.  Sort of a
"QED Manifesto" type project for CAS.  This seems to collide with the
goal of getting a system "up and running" to attract developers, as the
mechanisms I envision being most attractive to developers involve what
are probably fairly major work even to get started.  Hence my
conclusion that at least a temporary fork to pursue the "gathering of
developers to the current Axiom CAS" goal is a good idea - I support
that goal but it's a bit orthogonal to some of the other directions I
personally would like to see pursued.

I would like to re-iterate again that I am very grateful to the effort
put in by ALL of the developers, and I personally at least feel no
hostility to ANYONE for putting their time and effort into such a
wonderful volunteer project.  Waldek has done amazing work and I am
very, very grateful to him for doing so - I have no doubt his efforts
will continue to be outstandingly successful.  What I don't want to see
is two different (and worthy) goals slugging it out endlessly to the
detriment of both.  I don't see it as a fragmenting of the community,
just the community working on different aspects of what is in the end
the same problem - a strongly typed, robust mathematical CAS.  If we
don't become too incompatible on the Algebra level, I don't think a
fragmentation needs to occur.  Waldek's fork can be to Axiom like
Ubuntu is to Debian.

Cheers,
CY



 
____________________________________________________________________________________
TV dinner still cooling? 
Check out "Tonight's Picks" on Yahoo! TV.
http://tv.yahoo.com/