--- "Nikolaos I. Ioakimidis" <ioakimidis@otenet.gr> wrote:
> In general I agree, but I prefer this code to accompany Maxima in a
> different directory than SHARE, such as ADDITIONAL_SHARE_CODE
> or NOT_FULLY_TESTED_SHARE_CODE. This may be safer for the
> user of Maxima and will also reduce Jim's objections concerning bugs.
> Thanks to Wolfgang we can go on with the files Wolfgand suggestes in
> Maxima. Of course permissions by the authors of these and similar
> files are required.
We may have to do something like that in the future, but I think you
misunderstand the state of the Share code. It is widely varying as far
as to quality and documentation. Only a few packages have been through
even a basic functionality check, let alone correctness testing. At
this point, a not fully tested directory would be completely redundant.
It is likely to be some time before we need to make any distinctions.
> Many thanks for all of this information. The best would be that
> even the commercial Maxima code could be included in Maxima
> some day. (You are aware about Axiom!) This requires some effort.
> Perhaps Professor Fateman could help more. We will see.
Axiom is a different case, and I am indeed happy to see it released.
But the company who owned it was willing from the beginning to release
Axiom, and the reason it took so long was ironing out the legal
details. Every indication to date I have seen points to the current
owner of the Macsyma Inc. codebase being completely uninterested in
releasing it.
> Yes, yes, you are right. I still prefer a second share directory such
> as that I proposed previously without bug fixes for the moment. Of
> course, perhaps Wolfgang could take care of the packages under
> question (after a permission of course). A CLEAR distinction between
> the tested share code (the old one in Maxima) and that not tested or
> tested a little (the new one in Maxima). That's my opinion.
Someday what I'd like to do is add a unstable or notworking directory
to the share system, and have all packages start there and move to
their proper place in the share structure once they have been checked
out. Then we could add a toggle USEUNSTABLE to the load mechanism for
Maxima - if it is set to TRUE, search the unstable directory when a
load call is made, but if it is FALSE (default) ignore that directory.
That discussion is for the future - currently essentially all of share
is "unstable." No doubt the more skilled people among us will have good
ideas - but since we wandered onto it here's mine.
> > If I understand you correctly, you wish to locate older copies of
> > DOE-Maxima people have and incorporate code from those
distributions
> > into this distribuiton?
>
> Yes! This is exactly what I mean. The DOE-Macsyma code, with the
> exception of bugs, has been stable since 1982. No changes, no
> additions, just bug fixes. This is my impression based on what I have
> read. I do not know any details, but you know what happened.
Well, Bill made some changes, but it should be basically the same, yes.
> This is a point of disagreement, but I do not belong in the
> development
> team of Maxima, I do not decide, I am a simple user. My impression
> is that when DOE (rather a successor of DOE) gave the permission of
> the code to William F. Schelter, they essentially meant the whole
> code distributed by them (DOE-Macsyma) not with the exception of some
> files.
> But, to be on the safe side, we can come into contact with them and
> clarify that they did so, i.e. in my case, the permission included
> not only the ODEAUX.MAC file (already in Maxima), but the dependent
> files too. I wish to believe that the lack of inclusion of the
> dependent files in Maxima is just an omission and not an indication
> that these files have been exempted by DOE. Which is your opinion?
Erm. I'm not really qualified to have an opinion on that one. Our
attitude to this point with the DOE has basically been "let sleeping
dogs lie." It is also possible that the original DOE version didn't
have these files - they might have been given exclusively to the
commercial codebases. I'm much happier to deal with the original
authors - they have more of an interest in seeing their efforts live
on.
> On the other hand, for old code, mainly of 1970s, I doubt whether we
> could find the authors. This code may have been made in M.I.T. by
> persons not known to the computer algebra community you belong.
This is a bit of a problem. I've never successfully contacted anyone
at MIT about any Macsyma stuff they might have. Has anyone else ever
had any luck, or know who to talk to? I've always wondered if they had
some of this missing code or scribe versions of the original
documentation they could release.
CY
__________________________________________________
Do you Yahoo!?
Yahoo! Mail Plus - Powerful. Affordable. Sign up now.
http://mailplus.yahoo.com