Why I am trying what I am trying in my ode program

On 8/10/2012 10:09 AM, Dennis Darland wrote:
> My only comment is that I am not, as I have said at least twice before, trying to translate a FORTRAN program. (or translate any program) I looked at code generated by Chang's program (not the program itself mind you) once in the last year. I was having trouble with his equation for arccos (found in the appendix to his draft of his book), All I learned was that arccos was not implemented in his program. (I did locate a typo in my code a little later.) I know I could get this ACM article by signing up for a service with ACM (about $100/year I think). I seriously doubt I could find the other articles locally. Probably closest is Iowa City - about 60 miles, but I have not driven that far in about 15 years, and have no intent to now.
If you have a copy of the paper by Corliss etc, you can find details in 
their bibliography.
The book is Applied and Computational Applied Analysis  (3 vols!)

There are probably people who have subscriptions to the ACM digital 
library closer than Iowa City,
or people on the internet who can access items for you and email them.


It is sometimes said that a week of programming can save you 30 minutes 
in a library.

I think that understanding the fundamentals will suggest that
(a) some, perhaps most, of Chang's generated programs are necessary if
the programs are run on top of C, C++, FORTRAN, but are not necessary
if the programs are run in a computer algebra system.
(b) instead of simulating Chang's "generation of taylor series in FORTRAN"
you should just use the natural tools in a CAS to generate the taylor 
series.