[Axiom-developer] Axiom project goals, axiom testing

[root]

Vladimir,

The axiom project goals during the IBM years were basically to 
do research in computer algebra. There was no plan to commercialize
scratchpad (as it was called at IBM). It was only after IBM took a
major downturn and changed executives that efforts were made to 
make scratchpad into a product. Scratchpad spent about 21 years
being developed at IBM and was the basis of many people's research
work. 

Testing was done by the user community and the original developers
in a manner similar to the free development model. In fact several
people asked for and received the source code while I worked on the
project. It was only after scratchpad became commercial that the
source code was closed. 

It was normal to have several visitors to the scratchpad group at
any given time. Many of these people contributed large pieces of
code in their area of expertise. The actual group of full time
IBM developers was quite small compared to the number of contributors.
James Davenport, for example, wrote the original version of the compiler.
Barry Trager wrote the original integration routines as part of his PhD
thesis work. Scott Morrison wrote the original browser. Bill Schelter
wrote AKCL, the final lisp for scratchpad (prior to Axiom) which is
now GCL. William Sit wrote the differential equations package, etc.

As you saw in the book my job at IBM was essentially the job I'm doing
now. I am the person who "owned" the sourcecode. Everybody looked to
me for modifications, porting, bug tracking, testing, etc. The other
people in the group had various areas of expertise such as the interpreter,
the algebra, the graphics, the browser, etc. In the end, however, I was
the place where it all came together, making me the "low man on the 
totem pole". It's very similar to the way Linux is developed. Linus
"owns" the code but there are "maintainers" who have areas of expertise.
It appears that any large project must eventually reach this structure.

Tim