[Axiom-developer] Literate documentation

[daly]

Gaby,

I've had extensive discussion with 3 different funding groups at the NSF
and one at NIST. 

The problem isn't latex. The problems appear to be:

1) They won't fund work being done by commercial companies

     As long as Mathematica and Maple exist they will not fund Axiom.
     Of course, if Mathematica and Maple disappear it will be way too
     late to consider funding Axiom as the whole of computational
     mathematics will likely disappear down the wasted effort hole.

     When the Symbolics company folded it "captured" Macsyma, a major
     effort from MIT, and effectively killed it. The GPLed version,
     Maxima, only exists due to the effort of Bill Schelter and was
     several years out of date.

     When (not if, since there are very few companies that exist for
     50 years) Wolfram and Maplesoft die there is the likely possibility
     that the code will be lost to the world, as in the Macsyma case.
     The code represents a large pile of "intellectual property" and
     will not be freely released, and will likely not be available at all.


2) They won't fund work that is done by individuals

     Funded work needs an existing funding source that can handle the
     finances. A university has a provost to take 55% or more of the
     cash as "overhead". Open source does not have the financial
     machinery in place to accept grant money or manage the required
     reporting requirements.

     Funding an open source project with grant money has also
     raised the question about how to distribute and use the funds.
     Do we pay individuals? How do we judge the work? How is the
     work to be reported? Who manages the money?

     I've pointed out that open source could use the funding in at
     least four "group" related ways. One to fund a developer
     conference covering travel, etc. A second is to fund a
     "compile farm" setup at a data center which contains a wide
     range of machines and operating systems. Such a farm could
     be used by many open source projects. A third would be to fund
     the purchase of property rights in Macysyma (the last quoted
     figure I got was $250K for the source code).

     The fourth idea is by far the most interesting and ambitious.
     The idea is to follow Bruno Buchbergers lead and create a U.S.
     version (or branch) of the RISC (Research in Symbolic Computation)
     institute. RISC funded both research and industry work.

3) They won't fund developers outside the country

     The NSF and NIST are federal agencies and are not much interested
     in funding people who live and work overseas. They might fund a
     "visiting scientist" position for someone working at a university
     but that would not apply to an individual developer.

     I have not had any contact with INRIA or other European funding
     agencies but I suspect they might have the same objection. Since
     open source is worldwide there is no funding agency with the 
     span of control to handle funding. 


My conclusion is that the NSF and NIST are irrelevant to open source.

For a while Gilbert Baumslag funded Axiom development as part of the
CAISS project at City College of New York. Gilbert, however, is rather
more enlightened and sensitive to the long term value of computational
mathematics than the NSF or NIST management.

Thus, the funding issue is WAY above the latex question.



Tim