[Axiom-developer] Re: [NOT :-)] Time to say good bye to Tim.

[William Sit]

Dear Martin:

Thanks for your trust in me and I also value your deep
investment in the Axiom project and don't want to see you
(or anyone else) leave.

I am truly sorry that the community has split up like this,
all due to my naive believe that Tim was willing to really
do according to a "vote" on the merging issue. No doubt, Tim
has irritated almost everyone else save Stephen Wilson or CY
perhaps. He is, however, to be admired for his stubbornness
to hold onto his principles. Of course, I, like you and some
others, would much prefer him to be more flexible. (Tim, I
hope you are listening.)

Perhaps that's what Axiom is destined to be. I have far more
confidence in the more pragmatic approach that myself and
others have expressed (and there are other "silent" souls
out there too). As someone who only knows a small part of
the Axiom algebra world, I do not carry any "weight" whether
I stay or leave the community but I have not thought about
leaving (this is clear from my reply to Tim's "contribution"
message.)  I would very much like everyone to calm down and
arrive at a cooperating solution. But with egos flying, and
"principles" that are non-negotiable, it would be too much
to ask, but I sincerely hope that we write with less
threatening tones.

I feel sad that Tim has such a narrow mind that the only
thing that counts as "contribution" is "commit a patch" (or
"diff -Naur").  He has neglected the technical difficulties
when the patch is massive, unfinished, while the Gold branch
keeps shifting. He has been unfair to Waldek and Gaby.
Surely, Bill Page has contributed a lot by maintaining the
website as the de facto webmaster.   Bill has been very open
and generous to accept comments and helpful to explore new
ideas. How can these developers be accused of not
contributing to the project? Perhaps Tim can learn a bit
from Bill (as a "manager", Tim: time to get a raise! )?

We have to step back far enough to visualize what Axiom
would be like a few years down the road (not 30 years, I
know I won't be around). If Waldek, Gaby and Bill leave
entirely, what would happen? First they need to re-establish
the entire data-base for the website and create a new one.
Financially, it is not a real problem assuming people have
the will. But it has taken us a few years to build this and
it would take at least a year to recreate and build up
support again. Given so few are active, do you have
confidence? Is it worth someone, as an example, to build a
comparable website, say for FriCAS?  Would another split
happen with FriCAS? No one can tell.

You and Ralf (and Waldek) have painted yourselves into a
corner. Unlike Waldek, you and Ralf probably have not yet
gained the skills (nor would be interested in gaining them
as far as I can guess) to maintain the internals of Axiom.
Thus, like me, you will be dependent on people like Waldek,
Gaby, Stephen and Camm, and perhaps some newer developers. I
don't think Stephen and Gaby will " jump ship".  Gaby is
willing to put up with Tim by preparing patches. I have no
experience on such matter, but I do believe Tim is right on
insisting submission of patches. Tim is just too slow and
wants to be 100% sure committed changes will only fix
things, not break them (good luck, of course, but the goal
itself is admirable.)  I wish he has more trust in and
courtesy for the others.  But I digressed.  For you, Ralf
and me, and other algebra-level users/developers, all we can
do is to raise the algebra issues and hope that someone will
listen and help us. At this stage, those "someone" are all
working to improve the internals, with only marginal
interests (but certainly eventual interest) on the algebra
side. With this assumption, I believe "jumping ship" is not
to our best interest, at least not for the near future.

Consider one of two scenarios: (a) Waldek remains the only
guru on his branch. Everything you like to happen in
language improvement, bug fixes, new algebra code addition,
etc., has to depend on him. (b) Someone like a Waldek2 (or
Waldek3) comes along in a year or so, knows the ins and outs
well, but decides to do some other things, or do things some
other ways. Waldek2 does not have to split, but it is enough
that he would be interested, at least for the short (that's
two to three years) term to concentrate on things "of
interest" to him. We welcome such people, and we can't blame
anyone being interested only in "things of interest". An
example of that kind of skilled person is Stephen, who wants
to rewrite the compiler. I am grateful he discovered, and is
interested in fixing the Union problems and thanks his
willingness to consider about my proposals. But I shouldn't
ask (and certainly not threaten) him to take that as his
priority. Some one else may want to rewrite the interpreter
and others go about user interface projects. Independent of
which branch they will finally merge their efforts (perhaps
both), the fact remains, we, the algebraists, would still be
waiting. (A clarification is perhaps needed here: I do not
mean to separate the community into algebraist and
non-algebraists: I understand that Waldek, Gaby, and others
are also mathematicians (perhaps not algebraists in the
mathematical sense) and they are interested in using the
Axiom Computer *Algebra* System. They certainly wear two
hats, while I only wear one. But like we three, they also
have to wait.)

So my own solution to this situation is: if I ever develop
more Axiom code, I'll develop them not assuming the newest
language features that are still to be, but use what is old
fashioned well tested constructs. That is the only way to
produce working algebra code *right now*. It does not mean I
won't participate is discussions and experimentation of new
ideas (like the IndexedUnion or Unit and Dimension
subprojects), but just that I won't view that as my main
bread and butter.
For me, if I only work on algebra code under the current
language constraints, I don't have to worry about code
obsolescence. If my code is useful to others, someone will
use it and maintain and improve it (I hear Tim:
documentation!). If not, let it die a deserved death.

If I were to worry about the future of Axiom, there are
other issues that I did not emphasize but in my guts I feel
are the Achille's heels for Axiom. It took me this whole
week to write up and LEARN about *how to implement* this
very simple mathematical concept of Union and I am not a
newcomer. There is just so much more computer science
involved.  I am sure, as one of the few Algebra developers
around, you know what these issues are. When will a stable,
usable Axiom be available (under whatever name)? And how
many more years before the younger generation will embrace
it enough to contribute more algebra code? Those are my
worries. They are the reasons I prefer a more pragmatic
approach, and I join others' plead for continued unity.

Would it be too much to ask that you and Ralf (and Waldek,
and Tim) to put your egos behind and withdraw your threats
(and be more flexible)? I have faith that Waldek will try to
merge his with Tim's once things become stable and Tim will
accept them.

Martin, I hope you don't get too upset with Tim. He is only
one in the community. Don't blame him for all the ills. You
have an otherwise very cooperative community and no reason
to leave it.

Enjoy your vacation, in any case. I wish I have one.

William