On Sat, Dec 16, 2006 at 10:59:04PM -0500, Stavros Macrakis wrote:
> Some basic questions:
>
> 1) Do we know who the current owner of the commercial Macsyma is?
> Rumor has it that Andrew Topping died intestate. If that is true, it
> could take years to resolve.
This is a good question. If it turns out that Andrew Topping did own
it, then the legal issues are a pain, but perhaps it also becomes an
opportunity because the time is right to liquidate.
> 2) Do we have any evidence that the current owner is interested in selling?
Macsyma has not been commonly available, has not been developed much
over the last 5 to 10 years, Windows XP and Vista make Macsyma
potentially increasingly out of date and incompatible with current
operating systems. Mathematica and Maple and Maxima are all
competitors. Maxima is a direct strong FREE competitor.
Clearly the economic value of holding Macsyma is declining. In fact
the primary reason to hold it is to try to sell it to someone
else. Because the market for a commercial Macsyma is dwindling vs. the
Mathematica and Maple competition, it seems less and less likely.
> 5) Why do we believe that a Google or whoever would be interested in
> making Macsyma available under a free license?
Google is probably not interested in creating a commercial CAS. But
Google is very interested in using free software. Google supports free
software development with money through the SoC. They probably pay out
several million a year through that program.
Google has a lot of smart and motivated engineers, and has plenty of
problems that a CAS would help solve (reliability modelling, energy
consumption, thermal loads, financial risk, bayesian estimation in
machine learning problems, whatever ...)
IBM is in a similar technical position. Jimbo Wales is directly in the
position of being interested in education from a nonprofit
standpoint. All of those were good ideas for people to approach.
Here is a simple model for the value of Macsyma. Assume that you could
currently sell 1000 copies at $250/ea. (I think this is a stretch,
what about advertising expenses, distribution, who would buy Macsyma
over Maple? Support costs... etc). Assume with no further development,
the sales decline by 50% per year. Assume that the current interest
rates for money invested elsewhere is around 8 percent. After 10 years
you've got essentially no sales. Maxima calculates the value of
Macsyma as
(%i9) 250*1000*sum(2^-k*exp(-.08*k),k,0,10),numer;
(%o9) 464208.675620015
So a rational choice would be to sell Macsyma for something around
$500k. Let's say that my answer is good to within a factor of 2 either
way.
Now, alternatively, imagine that with Macsyma's internal code merged
with Maxima you'd get a pretty darn useful tool (especially since
Maxima is a pretty darn useful tool already). Imagine that Google has
1000 engineers and by using Maxima they will on average save 1 day per
year of work over the next 10 years, with the same discount
rate. Imaging the average google engineer makes 75000 per year, there
are 365 days per year but only 5/7 of them are work days.
(%i17) 1000 * 75000/(365 * 5/7) * sum(exp(-.08*k),k,1,10);
(%o17) 1902002.558429897
So the case could be made that Google could pay $500k for Macsyma,
release it to us for maintenance, wait a year or so (k=1), roll it out
to their engineers, and save money over the next 10 years. Not to
mention the value to the world. Not to mention the fact that by giving
away Macsyma, Google could probably write off the cost as an expense
against taxes.
Macsyma has become far more valuable as a free tool than it is as a
way to make revenue from selling the product under this model.
--
Daniel Lakeland
dlakelan at street-artists.org
http://www.street-artists.org/~dlakelan