[Axiom-developer] life and death

[daly]

For a few years I maintained a CD collection called Rosetta. 
I collected every computer algebra system I could find and the
list numbered over 100. Most of them (over 80%) were started by
people who unified the idea of language typing and the idea of
operator overloading to get a CAS. Most of them were in C++.
Almost all of them stopped after implementing some form of
polymorphic polynomials. Almost all of them cannot handle
issues of non-commutative arithmetic. All of them die when
the author moves on to something else.

The Rosetta project shows that there are hundreds of man-years
wasted making new systems (each project being about 2 semesters
long). If the same effort was put into a single Axiom domain
it would live longer, be generally more useful, and if properly
documented it could even be maintained.

I find it frustrating to watch yet-another-CAS being born.
There are 30 years and hundreds of man-years invested in a
system the size of Axiom. No single person, and likely no single
group, is going to get the long-term, expensive support to do
this again from scratch. There are no funding visionaries left.

So the issue isn't really python-vs-spad. The issue is whether
you care to spend your time writing code that will die or code
that will live.

Tim