Lucas Wiman wrote:
>Even if they (George and Scott) did this, then there would
>still be MacLucasUNIX, or everything else in the mers package, as well as
>Ernst's program, and good ol' lucas.c. Any of these could be used. We've
>really got to put our feet back on the ground here. If we did put a license
>change on all of George's program derivitives, we would still have to get
>Will and Ernst to change their copyrights, and Richard Crandall.
In fact, I've been thinking about these issues for some months now.
I already have GNU-style copyleft language in the upcoming release of Mlucas,
but the monetary issue complicates things. My personal view is, if I spent
a lot of time and wrote the program, I should be entitled to a share of the
money (and being between jobs at the moment, it would come in handy :)
But, how does one combine that with a copyleft? If I give you the right to
to freely modify and redistribute my code (with original header intact) that
would seem to imply any associated prize money split percentages remain in
force, but at what point (i.e. many modifications later, say 5 years after
I put it out there and others pick it up and I'm killed by a freak asteroid
impact on my way home from the Coffee Society) does it cease to be my code
in any legally enforceable sense. And, to my knowledge, the whole issue of
a copyleft has never been subjected to a serious legal challenge, so its
legal status is unclear.
>From an algorithmic/prime finding perspective, I don't want to be too greedy,
since fewer people will want to run my code if they get only 10% of any money
(let's face it, that's human nature), nor do I want to put too strict a clamp
on the source code, since I want to encourage algorithmic improvements,
assembly
implementation, interfacing with a PrimeNet GUI, etc., all those things which
need others' input and effort, even if I'm still closest to the basic
algorithm
at the heart of the thing.
As in so many other arenas, money is nice to have, but it can really
complicate
matters.
>In fact, is the DWT patented?
Crandall and Fagin published the details in 1994 in Math. Comp., which
(I believe) precludes any patenting. Had they patented it first, they
would be licensing others to use it, meaning that the EFF money would
likely not be an issue, since GIMPS would not have gotten started without
free software for all to use.
Aaron Blosser wrote:
>And it was mentioned before...if George and/or Scott setup such a legal
>contract regarding software usage, there probably would be people writing
>their own software, hoping to get the big cash all to themselves.
I think Aaron understimates the amount of work it takes to write a really
good program of this kind. Even $100000 would not reimburse George for
his work on Prime95, at any reasonable hourly wage for a skilled programmer.
And even if you've written a great code, you need to get *lots* of people
to run it - insisting on keeping any prize money for yourself makes you
seem greedy, and it's that perception more than anything else that would
hurt your recruiting.
-Ernst
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
