--- Richard Fateman <fateman@cs.berkeley.edu> wrote:
> I think that, more than many other projects, the maxima program
> lends itself to be divided into sections, with different people
> taking primary responsibility for these sections. This is not
> to deny the advantage of having a single leader, but just
> to say that quite a few things can be done by distributed
> programmers. Coming up with a new release periodically requires
> coordination and perhaps someone acting as a leader.
That seems to make sense.
> Why do I say it can be divided? ...
> Here's why.
> The lisp code is divided up into files, each one of which (or
> perhaps a collection of a few files) defines some capabilities.
> For example, numerical evaluation of special functions. Some
> require the correct behavior of many other parts, but through
> interfaces, often declared in the files themselves, though
> not always. (Integration is in a few files, but uses other
> capabilities including simplification, rational functions,
> polynomial arithmetic, etc.) It is possible for someone to
> be "in charge of integration" to ask someone who is
> "in charge of simplification" to add a feature.
That also sounds good.
> There are a number of modules that could, in principle,
> be entirely replaced by new code: re-doing polynomial
> arithmetic; re-doing the display; re-doing the front-end
> language parser. Even though some of these are at the
> core of the system, they act as dispatchers to other
> programs. The model, for those who have learned lisp or
> scheme up to a modest level of achievement, is one of
> a read-eval-print loop. And hanging more stuff off the
> eval part is always possible.
>
> A more sophisticated view of what maxima does would have
> to include the side effects of eval-ing a command to
> some kind of cummulative data base of facts, definitions,
> etc.
>
> There are other portions of such a project that I
> think of as entirely separate, namely the porting to
> different lisps and machines. This sometimes requires
> a longitudinal look at ALL the files to fix dependencies
> of some sort. It should not require such an effort,
> but in reality, recompiling 200 files or so on a new
> machine/host system ordinarily finds "something".
>
> RJF
If we go that way, we need to divide up the tasks into
pieces which will be maintained. Richard, you're almost
certainly the best qualified of any of us to handle that
and be overall coordinator - if everyone agrees, would
you be willing to take on that role?
__________________________________________________
Do You Yahoo!?
Make a great connection at Yahoo! Personals.
http://personals.yahoo.com