Quadpack and maxima



--- Richard Fateman <fateman@cs.berkeley.edu> wrote:
> 
> 
> C Y wrote:
> 
> >
> >Dr. Fateman, it would be very interesting and possibly useful
> >if you were someday able to write up all the little tidbits
> >like this about Maxima - 
> >
> 
> Isn't that what mailing list archives are for :) ?

Point :).

> Do you think this is really that interesting?  As pointed out, there
> are now better quadrature programs (quanc8? ) in maxima.

I didn't mean this specific case - you know a great deal about a large
number of the internal systems of Maxima we haven't delt with yet.  The
quadrature example was just a good example - knowing there are better
ways out there than the one we impliment lets us add a possible future
todo or enhancement possiblity.  Eventually I suppose most of this
would be come clear as we grind our way though, but if we start out
with an overall map of weak and strong areas of the system we might be
able to pinpoint more easily the better approaches to take (in broad)
to fix deep rooted errors or limitations.  Just at thought - not really
essential.  Part of my thinking is that while some early pictures exist
showing the "flow" of maxima's overall structure, I'd be very surprised
if they are current to even the 1982 version - we don't have a really
good overall summary of how the various parts of Maxima work and are
handled.  Maybe one isn't desirable or possible, but it struck me as a
good way to introduce new programmers to the system.

> More generally, I guess that some of the "information"
> about maxima is in the source code as comments, or should be there.
>  I put quotes around information
> because it is kind of like the newspapers in the novel 1984. People 
> rewrite them.
> If you look at some of the comments, you see people, sometimes 
> correctly, and sometimes ignorantly, shouting at the previous coders.
> I suppose they could do worse. They could just delete previous code
> and insert new (maybe buggy...) code.

True.  But new coders might have difficulty sorting all this out, and
an overall "map" with comments on which areas of maxima are strongest
and weakest and why (nothing as detailed as code level - merely stuff
like "This ability is not aware of property X, which it should be for
full evaluation capabilities") would be a good way to point people in
specific useful directions when investigating the code.

Just a thought, and again probably more work than it's worth.  

CY

__________________________________
Do you Yahoo!?
Free Pop-Up Blocker - Get it now
http://companion.yahoo.com/