[Axiom-developer] authorship

[daly]

I really hate to enter this debate because, apparently I'm at the
center of it. But there has been a policy of authorship and credit
in Axiom, although I probably never stated it clearly in one place.



I've held forth on the subject of authorship and/or credit in
Axiom before, for instance,

<http://lists.gnu.org/archive/html/axiom-developer/2003-08/msg00058.html>



Being a second-string academic all my life I've had a great deal
of experience with the topic, not all good.

I spent 19 years at IBM Research where "credit" is the coin of the
realm, so to speak. I remember spending years on a project, having
written a reasonable portion of the software, most of the tools, and a
large portion of a user's guide with 3 other people. The copy of the
guide I still have has 4 names on the cover. The DAY after I left the
project the guide was changed to exclude my name. I have a copy of the
"new" guide which is character-by-character the same except for my
name on the cover. 

I also worked on a patent, which my notes (prior to joining IBM) show
as my idea only to have my name excluded from the patent application.

I also generated an idea, did the background research and wrote the
code to demonstrate the idea but when the paper was written and
submitted my name was nowhere to be found. (Names and identifying
details withheld, despite my policy of "liberal credit sharing" :-) ). 

I spent years on another project which I rewrote in common lisp despite
protest (worse, a reprimand). The final project was shipped as common
lisp. The project generated a couple awards of $20,000 each. My name
was not mentioned in the announcement and I did not see a single
dollar of the cash.

Such events, no matter how well rationalized, always feel wrong.




I joined the scratchpad project and a few years later it was to
be sold. I spent time over several months trying to dig up the name
of everyone associated with the project. I put their name on the
documentation and I added a )credit function to the top level command
line of Axiom. I'm certain you don't know most of these people and
you do not know the roles they played in Axiom's history. 

When Axiom was released as free software and I rewrote it into the
current literate form I originally claimed that the author of every
work was Nicolas Bourbaki, thinking that was a clever reference to
a mythical mathematician. I was later informed that this is an official
society in France and changed every author to "The Axiom Team".

<html://lists.gnu.org/archive/html/axiom-developer/2003-08/msg00062.html>

Later the "Axiom Team" idea bothered me as there were files for which
I clearly knew the author(s). Their names should be on the work.
I spent time trying to discover or guess the probable author
of each of the nearly 400 files.

Volume 1, the Tutorial book, includes a page that has the name of
everyone who contributed to Axiom up to that time, as far as I know.

The role is not important, it's the effort that counts. There are
people who are listed as contributors and/or authors where the current
work no longer contains a single line of code or documentation. But
they were the original authors. 

I've just gotten permission to use portions of Joel Moses' work 
as part of the documentation effort I'm doing for integration.
It may be that after reformating/editing/rewriting/rethinking
the whole effort there may not be a single line of text that
comes from his original contribution. Yet clearly Joel will be
listed as an author. Authorship is ideas as well as text. And
it is important. So important, in fact, that credit was the ONLY
request made in exchange for his generous contribution.

When considering the issue of authorship and credit ask yourself what
happens when, in 30 years, Axiom contains not a single line of code or
text that you wrote. Are you still an author? Should you still be in
the credit list? Is it meaningful for co-authorship to exist between
people who are 30 years apart in time? Does it hurt to share
authorship or credit? Is ``diff'' the golden standard?

As a matter of (my) policy it has been the case that the most liberal
guidelines were used for authorship and credit. I believe this tradition
should continue.

Tim