Re: [Axiom-developer] Re: Fricas

[C Y]

> If for whatever reason we can't work together with Waldek in the
> Axiom project then at least Axiom still has access to his skills
> and the work of some other people through FriCAS. It is not the ideal
> situation but it is not a net loss. (I think at least Clifford Yapp
> agrees with me on this.)

Yes.  That is the situation so long as everyone maintains compatible
licensing arrangements, which I see no reason to change.  I don't think
either project should constrain itself in order to maintain
compatibility with the other, however - that defeats most of the point.
 There is a good chance that the paths will diverge to an incompatible
point, depending on the directions taken by each project.  The code
base is in a very early state right now, and will probably be
extensively re-worked.  (Axiom may go with Steve's new SPAD and FriCAS
might not, for example - who knows?)

> Devoting a little of the Axiom project's
> resources to track what is happening in FriCAS just seems very
> sensible to me.

Can't we just follow the FriCAS development list and logs?  I can see
adding a link to FriCAS as part of a general listing of CAS's,
something along these lines: 
http://maxima.sourceforge.net/compalg.shtml
That's pretty typical of most projects, in my experience.  But I do
agree that FriCAS should have its own independent identity.

> > If someone writes algebra that only works in FriCAS and not in
> > Axiom but says that it is "Axiom algebra", could someone be
> > confused?
> 
> No. This can be easily explained to any beginning user of Axiom and
> FriCAS - especially if no one closes their eyes about what is
> happening in the FriCAS project.

I have to say Bill in such a case as an end user I might get confused
when I tried to run "Axiom Algebra" in Axiom and it didn't work.  Isn't
it simple enough to just call it FriCAS algebra?

FriCAS is (or will almost certainly become) different from Axiom much
as Macsyma Inc. became different from DOE Macsyma, Vaxima, etc., except
in this case both projects are still active.  Making sure Algebra from
either project will work in the other is a separate QA effort and may
require porting.

> Are you confused about algebra that
> only works in Aldor but not Aldor in Axiom? Right now FriCAS is
> identical to Axiom in the algebra code that it runs. But FriCAS has
> solved some problems with the algebra bootstrap that have not been
> solved yet in Axiom. So the potential certainly exists for new
> algebra code to be written that can not (yet) be compiled in Axiom.

In that case (working in FriCAS but not Axiom) it is probably not a
good idea to confuse it with Axiom's algebra, just from an end-user
standpoint.  As an end user I just want to know which CAS to run the
Algebra in, not that it might work in another CAS.

> (We have been discussing that in another thread on MonoidRing.) As
> far as I know it is the intention of everyone involved that Axiom
> will eventually be able to compile this code.

I doubt the FriCAS project will want to spend effort on that though -
they will probably incorporate the fix and move on.  Eventually Axiom
may be able to use the new code, but it could be a while.  If we move
Axiom to building off of full fleged Category Theory everything may
need to be re-thought.

> > If a paper appears at ISSAC next year stating it was done on Axiom
> > but was done on FriCAS and won't work on Axiom what can we say?
> > Could someone be confused?
> 
> I don't think so. All it takes is one sentence of explanation.

I would think it would be better to specify FriCAS.  If I'm a user of
FriCAS and want my code to work there I'm not worried about Axiom - I
just need to know to download FriCAS.

> Well, I think *most* people would like to continue to use the name
> Axiom generically to refer to computer algebra systems that have
> evolved from the IBM ScratchPad project. I don't think there is much
> room for confusion here - certainly not more than already exists. It
> would be very easy to include both FriCAS and Aldor as categories in
> the IssueTracker system on the Axiom Wiki.

>From an end user perspective I can see that getting confusing,
personally.  I know you don't like the FriCAS name Bill but in the end
does it make much difference?

> > If someone demonstrates FriCAS and it has abilities that don't
> > work in Axiom could someone be confused?
> 
> That is easily explained and one very good reason why I think FriCAS
> should continued to be discussed in the Axiom email lists and on the
> Axiom wiki. To do otherwise is just to invite the type of confusion
> to which you refer.

Um - isn't that an argument for the projects to be separate?  I would
have thought the closer they were tied together the more confusion
there would be.

When Xorg forked from XFree86, the projects essentially went their
separate ways despite starting from a virtually identical code base. 
(Granted there were licensing issues there.)  There is no doubt Xorg is
a separate product from XFree86.

> > FriCAS is capable of developing its own resources.
> > That's what a fork is.
> 
> A fork is a failure to communicate (which also happens to be the
> cause of most wars :-(

I would have thought this fork could be characterized as a desire to
pursue different project goals and methods.  Either way, better a fork
and ongoing work in forks (particularly with compatible licensing being
maintained) than endless arguing and no work in one project.

Cheers,
CY


      
____________________________________________________________________________________
Park yourself in front of a world of choices in alternative vehicles. Visit the 
Yahoo! Auto Green Center.
http://autos.yahoo.com/green_center/