[Axiom-developer] Re: Axiom and commercial success

[root]

> | I encourage you to raise your eyes to the horizon and ignore the
> | rocky road under your feet.
> 
> that is well put; however, the rocky road needs to be dealt with, not
> just ignored.  How do you get people on board for long-term projects
> when you don't make room for them to make sure they can embark on the
> long journey?

Do research. I define research as "doing something nobody in the world
knows how to do". 

People can do "big research". They can pick a rock that is distant and
hard to reach like 
  * "define B natural"
  * "define a new Aldor"
  * "define 
            n
    +-    -+
    | a  b |
    | c  d |
    +-    -+
  * "merge GAP and Axiom"
  * "document the algebra in ALLPROSE""
  * "define a complete, correct type coercion graph"
  * "define an [algebra domain] <=> [relational database] isomorphism
  * "symbolically solve for points on an elliptic curve"
  * "automatically compute the complexity of axiom algorithms"
  * "build a computer algebra test suite mechanism"
  * "define an algebra <=> graph isomorphism and reduction algorithms"
  * "add higher function support (eg polylogs) globally"
  * "define a "delta-epsilon algebra"
  * "architect a classroomt textbook mechanism"
  * "automatically derive a numerically stable approximate poly algorithm"
  * "define a "numerically stable simplification" algorithm
  * "create a cluster-aware axiom"
  * "create a parallel axiom"
  * "define a group theory type hierarchy"
  * "define a chemistry type hierarchy"
  * "organize an axiom video tutorial series"
  * "automatically generate flash-based time series graphs"
  * "define infinite reals"
  * "deeply embed literate algebra"
  * "entwin ACL2 and Axiom to do proofs"
  * "do automatic literate technical paper classification"
  * "define a NIST symbolic classification"
  * "automate a general bibliography mechanism"
  * "handle wave-equation mathematics"
  * "organize an existing math topic, collect papers, make them axiom-literate"
  * "add general genetic algorithm machinery"
  * "define propositional logic in axiom"

People can do "little research"
  * "expand SExpression to accept Maple syntax"
  * "expand SPAD to handle dependent types"
  * "implement symbolic summation"
  * "implement a clifford algebra"
  * "document an algebra domain in ALLPROSE"
  * "graph the existing algebra lattice""
  * "document the axiom domain to algebra database implementation"
  * "define tools for elliptic curve algebra"
  * "automatically compute the complexity of factoring polynomials"
  * "build a computer algebra test for a domain"
  * "define a way to graph polynomial operations and do them graphically"
  * "implement polylogs"
  * "handle stress/strain using a simple "delta-epsilon algebra"
  * "implement Strang's MIT linear algebra course"
  * "hand derive a numerically stable approximate poly algorithm"
  * "derive a "numerically stable simplification" polynomial simplifier"
  * "write a cluster-aware matrix domain"
  * "write a parallel processing matrix domain"
  * "extend axiom's group theory type hierarchy"
  * "implement a chemistry domain"
  * "make an axiom video tutorial"
  * "make a flash-based time series graph"
  * "implement an infinite real domain"
  * "implement literate )compile"
  * "decorate Axiom functions with ACL2-style axioms"
  * "investigate automatic technical paper classification"
  * "organize and name existing algebra for symbolic classification"
  * "create a build-time bibliography mechanism"
  * "accept and compute with feynman diagrams"
  * "pick an existing paper, make it axiom-literate"
  * "write a simple genetic algorithm solver"
  * "make the boolean domain work right"

Axiom is an ideal research platform and we're in the first 36 years
of the collision of mathematics and computation. Somebody just handed
us 36 years and 300+ man-years of research. We have complete and
total freedom to do anything we can dream.

> when you don't make room for them to make sure they can embark on the
> long journey?

Room? I see intellectual room in every direction. I could have a team
of computational mathematicians, add a new person every month, and
never live long enough to cover the space.

If you suddenly got a job working on Axiom and were given complete
freedom to do anything you want, what would you do?

Room really isn't the issue. Time and money are.


\begin{flame}

All existing computational math platforms have suffered from the 
total lack of vision by the traditional funding sources. The 
NSF/Universities/Research Labs (traditional funding sources) used
to consider "computer math" as leading edge research. They now 
consider it a "solved problem with commercial implementations" (NSF).
Or not a good revenue stream since it won't make money this quarter (IBM).
Or a cross-discipline crevice (Universities).

Traditional funding sources, (at least in the US, the rest of the world
seems rather more rational) seem unable to understand that the science 
of computational mathematics is hardly born yet. It's like the funders
have said "Plato discovered geometry... math is complete."

The NSF has a policy that "if there is a commercial implementation we
don't fund it" and feels that Axiom has something to do with MMA and
Maple.  We are being compared with the "commercial systems" and
considered losers.

"Do not compare yourselves to others or you will become vain or bitter"

\end{flame}


Axiom is a research platform in computational mathematics.
There is no competition, commercial or otherwise.
Real science is not a commercial enterprise.
Fund it with your time and energy.

> How do you get people on board for long-term projects

I don't know what motivates people to skip sunshine and wrestle
with hard problems. I'm irish. I don't do sunshine :-).

My primary job seems to be to carry this software thru the funding
dark ages. I'd make progress quicker if I could but it takes a LOT
of time to make sure it "just works". Most of the things I'm
actually working on will take years to complete. But I'm in no hurry.

Tim