[Axiom-developer] defintrf.spad, polycat.spad

[daly]

Bill,

I understand how you feel. I had extensive conversations about that
email with several people. It was a carefully crafted statement.

Axiom is no longer an IBM project. It is no longer a NAG project.
It is an open source effort that names a particular savannah and
sourceforge project. When a project forks, the new project needs 
to choose a new name. Otherwise the Axiom project will have to deal
with bug reports, complaints, or other problems over which we have no
control. The "Axiom" name defines our effort and its goals.

Carnegie Mellon University Common Lisp (CMUCL) was a long running
project at CMU under Scott Fahlman. I communicated with that team
while at IBM because we used CMUCL. A fork was created to "improve"
CMUCL. That fork is known as Steel Bank Common Lisp (SBCL). Both works
still exist. Same initial code base, different goals, different name.




> I think your email to Waldek is grossly unfair and inaccurate. I think
> Waldek (building on top of the work done by Gaby) has contributed more
> to the Axiom project in the last 6 months than any other Axiom 
> developer has in the last 5 years Your claim that he would "rather not
> contribute to Axiom" is ludicrous. 

Define contribute. As I defined contribute, which I believe is the
common open source understanding, it involves creating patches
against the main line of work and publishing them. Show me patches.
I asked for a changeset for hyperdoc so the trunk would be fixed.
I see no changeset posted. That is a statement of fact which is not
"unfair and inaccurate".




> ... I think it is clear from the axiom
> developer email archive that Waldek initially did not "prefer to
> create his own project", rather it seems obvious to me that the
> creation of a separate project was entirely in response to actions and
> statements made by you.

As adults we accept responsibilities for our actions. I made the
statements that if you join a project but do not agree with the project
goals, if you create a separate branch and promote it above the main
trunk, and if you fail to publish your work back to the main trunk
then you are creating a fork, not contributing to the project. I stand
by those statements. I believe that several people on this list also
agreed that the Waldek branch looked like a fork. There were two
possible responses. Either contribute by posting patches for individual 
features or create a new project. 

Waldek chose to create a new project. I didn't "make him do anything".
I'm quite curious to see what the new project goals are. He is likely
to fill a gap "in the market" that we don't cover.  I harbor no hard
feelings about his choice. I wish him success.




> About the name "Axiom": I believe that the most accurate statement
> would be that this name still "belongs" to NAG through the copyright
> statement that NAG insisted on as part of the open source release.

Actually, you're wrong about that. First, the license clause never
mentions the word "Axiom" (see license/license.axiom or any file).
The license is a copyright issue, not a trademark issue.

Second, "Axiom" is a trademark issue, not a copyright issue. I had a
discussion with NAG about transferring the trademark. They did not
renew the trademark, at least in the U.S. NAG withdrew the product
from the market and thus their trademark has lapsed.

I paid for an attorney to discuss the trademark issue. I paid for a
trademark search.  I could apply for a trademark on the name Axiom in
the computer algebra area (there are many "Axiom" trademarks) but I'm
under no obligation. I can contest the attempt by anyone else to
trademark Axiom in computer algebra because I have been using the term
continuously and have the association. So in essence I have a common
law trademark on the term Axiom in the computer algebra area despite
it not being federally registered to me. It is quite expensive to
register a trademark and I don't feel like spending that much money
without cause. The actual comment is:

 "Federal registration is not required to establish rights in a
  trademark. Common law rights arise from actual use of a mark. 
  Generally, the first to either use a mark in commerce or file
  an intent to use application with the USPTO has the ultimate
  right to use and registration."

According to the attorney I believe I have the common law right
to assert that Axiom is an unregistered trademark associated with
me and this project. I have asked that it not be used by FriCAS.




> To associate "Axiom" only with this particular open source project
> as you have defined it, would be misleading and probably wrong.

Really? How curious. We've done it for years. Please explain.





> ... But of course
> it would be inconvenient to have two projects with exactly the same
> name so identifying itself as FriCAS (AXIOM fork) Computer Algebra
> System" seems entirely appropriate to me.

But it isn't appropriate. FriCAS is a separate project with different
goals and different code.  By using the Axiom name, the FriCAS code
will trade on the goodwill and extensive effort over the last 7 years
associate with the Axiom name. FriCAS needs to develop its own
goodwill, user base, mind share, developer team, bug reporting, user
set, web site, mailing lists, goals, policy statements, and all of the
other "non-code" portions of this project we've spent years
creating. A project is not "just code".





> I am *very* disappointed that you apparently intend to force the
> FriCAS project to remain separate from what you are calling the "Axiom
> project".

It *IS* a separate project. Check the sourceforge link you sent. 
I believe it has been officially started as a new project.
I didn't create it. Please stop trying to make it "my fault".
I have no association whatever with the FriCAS project.





> You seem to be unaware of the damage you are causing.

Damage? Bill, you seem to think *I* set up the site.  I simply asked
for the professional courtesy not to have the Axiom name associated
with this new project. We have all worked very hard to develop as "Axiom"
a project dedicated to developing correct, well-documented CAS with a 30
year horizon focus that can be the basis of computational mathematics
in the long term. This should NOT be associated with a project that
does not share those goals.

It's a simple request. Change all instances of Axiom to FriCAS.




> And I fear that this might force people to "vote with their feet"
> rather than collaborating in the manner that one expects in an open
> source project.

I stated what I believe is "the manner that one expects in an open
source project" as a means of collaboration. I participate in other
projects, lead other projects, and am on other project mailing lists.
I've been doing open source since the Pinger project in 1997.
I REGULARLY see diff-Naur patches. Several a week. People create a 
branch, develop a feature or fix, and post patches. They contribute.

I have no opinion about the choices other people make. It is perfectly
reasonable to participate in more than one project, even more than one
computer algebra project. I'm also a developer on Magnus, another
computer algebra system. And I started Doyen, which Jose has been
managing quite well for years. I have commit access to the
Common Lisp Cookbook project. People should feel free to participate
in FriCAS if they agree with the project goals. I haven't seen the
project goals so I cannot comment. But you seem to think this is an
"either-or" decision. It isn't.


But I do request one thing. Stop using Axiom in the FriCAS project,
if only as a matter of professional courtesy.


Tim