Waldek Hebisch <hebi...@math.uni.wroc.pl> writes:
| Alfredo Portes wrote:
| >
| > On Sat, Nov 5, 2011 at 10:47 AM, Waldek Hebisch
| > <hebi...@math.uni.wroc.pl> wrote:
| >
| > >=A0Concerning OpenAxiom, I think that differences
| > > are much smaller. =A0But we would have to get consensus in
| > > tens (if not hundreds) issus like the current change to
| > > underscore handling. =A0And statements like 'this will be a
| > > forever incompatibility with OpenAxiom' does not make me
| > > optimistic that attempt to merge would succeed.
| >
| > I think you misunderstood Gaby's message. In previous
| > messages in both lists there were several comments as to keep
| > compatibility between the 2 projects. But the change you plan for Fricas
| > is probably something he is not planning to do and it has not been discusse=
| > d
| > or proposed in OpenAxiom, therefore his warning.
|
| Let me explain how I read his message: the change is easy to
| implement, so if he approves it he could do it in OpenAxiom
| and there would be no problem. So apparently he has something
| againt it. But he does not say so explicitely and gives no
| argument against, just raises issue of compatibility. Yet
| AFAICS impact of this change is much smaller than several
| of changes that were made to OpenAxiom (and FriCAS too, but
| my impression is that OpenAxiom has slight lead in introducing
| incompatibilities). It is true there we recent comment about
| compatibility. But Gaby is smart enough to know that various
| design/implementation decisions impact compatibility and
| that compatiblity is important for users. So it is hard
| for me to believe that he suddenly became more concerned
| about compatiblity and his comment was sincere.
Waldek --
Alfredo is exactly right.
I believe you are blowing this out of proportion, far more than it
deserves; and that is uncharacteristic.
First of all, FriCAS is your project and you should feel free to
implement whatever you have determined is in its best interest.
For OpenAxiom, I cannot implement everything. I cannot include all
good ideas in OpenAxiom. Rather, I have to triage and sometimes
elect not to implement an idea however good it is (in fact I have done
so for many of my own). When an idea is proposed for OpenAxiom, I try to
explain why I make a choice -- whether accepted or rejected. And in
this case, underscore is not even proposed for OpenAxiom, so there is
less to discuss.
You are right that I know better about the combustible
syntax+dialects-separated-by-common-core-language to have my mouth
shut. Over 15+ years working on C++ standards and compilers, I've seen
civil wars break out in the C/C++ community far more often on syntax and
incompatibility issues than on semantics incompatibilities. I should
not assume that the tiny AXIOM community is any different.
My expectation was that a measured and level-headed reader would give
appropriate weight to the caveats my message contained. Let me
reproduced it here for convenience:
# FWIW, my suspicion is that this will be a forever incompatibility with
# OpenAxiom -- just a note for those library writers who are aiming all
# AXIOM flavours.
It started with "FWIW", standing for "for what it worths"; then
continued with "my suspicion", and finished with "just a note for those
writers who are aiming all AXIOM flavours".
That you managed to extract from that more than what was said is, ahem,
an achievement. Since, I suspect you know better, something else must
be at work and I do not know what, so I will refrain from further speculation.
As Alfredo pointed out, some people have voiced criticisms on these
mailing lists about incompatibilities and the fractured community
(some I have heard voiced to me directly by gentlemen). So my
message was just a note (as it said) that my suspicion is that
particular change would never be implemented in OpenAxiom. That should
not deter you from doing what you believe is in the best interest of
FriCAS. Ideas flow between the two projects freely and openly (and I
consider that heathly), but as far as I know you and me do not dictate
to each other what other should do. And unless a dramatic event
happens, it will continue to be so.
| > Regarding merging, I honestly think both projects would benefit from
| > it, and I am
| > sure a large group of the people reading both lists want this. However only
| > you and Gaby have the last word on this matter. If there is no
| > intention of merging,
| > then of course feel free to make the changes you think are best for Fricas.
|
| I believe that technicaly merging sources requires substantial work
| but is possible. As I wrote the tricky part would be to get
| consensus in matters that were done differently in both projects.
This is the very first time that I see you bringing up this issue. I
you are serious about it, then it should be discussed in a level-headed
environment, unencumbered by side distractions like this one that are
blown out of proportion. And probably with not too many cooks in the room.
| Both projects would have to engage into merge. I would be willing
| to sincerely try, but first would have to be convinced that the
| other side is willing too.
Let me show you how this kind of stance ensures nothing happens:
imagine I responded by "I am willing to sincerely try, but first I would
have to be convinced that the other side is willing too, is serious
enough not to take every occasion to blow fairly benign issues out of
proportion." How you would react? The pattern in both statements is
that one side self-proclaims itself being on the right side and demands
that the other side first passes some ill-defined test before it
considers trying to do something.
| Coming back to underscores: I spent
| some time thinking about the issue and I believe that the
| change is significant improvement with almost no drawbacks.
| Yet it is not so big issue to risk unity of project -- in the
| past I have few times taken what I consider technically inferior
| solution but appeared to be majority view. However, when
| opposition to what I consider pretty clear technical decision
| is voiced in such a way (see above my interpretation of Gaby
| message), then I have doubts if more tricky (== when technical
| aspects are less clear and impact is bigger) decisions will
| be resolved in constructive way. Of course, that my view
| and Gaby may see it in different way.
1. I am/was not opposing anything you implement in FriCAS. As I said,
you should free to implement whatever you think is in its best
interest, and you should not let compatibility comments from me
distract you from what you believe is best for FriCAS.
2. From my perspective (and this reflects my philosophy when
adding something new to OpenAxiom), I do not implement something in
a language just because I can or just because it is technically
easy to implement. I first consider the idea, how it enhances or
benefits the language or solves some problems or solves some common
programming issues. This point is usually non-trivial to appreciate.
This is not aimed at you; it is a reflection of what I have seen in
the wide world over 15 years.
3. Once I determined that something would be good for OpenAxiom, I
first look around for how similar things have been done for AXIOM
(or Aldor) before. I weight variations and finally settle on something.
Then I write a specification. And only after I consider
implementation. Furthermore, I try to do something in the "spirit
of AXIOM" (a hard notion to pinpoint, granted).
4. I do not think underscores in names is in the "spirit" of AXIOM.
It is not forbidden, but one would have to take extra effort to
have them. Furthermore, nobody has convinced me of its value (see
point 2. above) and that is probably OK because nobody has proposed
it for OpenAxiom.
It is not that I do not know what underscore in names look like --
I have been working on C++ standard, compilers and library for over
15+ years where underscore as separator is the norm.
I do not miss it in OpenAxiom not did any OpenAxiom user bring it
as one of the top 5 issues they have with OpenAxiom.
5. Points 2 through 4 explain my "suspicion" of something not getting
into OpenAxiom "forever". Of course, the Chairman of the
Adjustment Bureau has latitude to rewrite that "forever" into
something else; after all, it was only a suspicion.
I hope this canard is put to bed for good.
-- Gaby
------------------------------------------------------------------------------
RSA(R) Conference 2012
Save $700 by Nov 18
Register now
http://p.sf.net/sfu/rsa-sfdev2dev1
_______________________________________________
open-axiom-devel mailing list
open-axiom-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/open-axiom-devel
