| Re: a plea for more reliable mathematical routines [message #17102 is a reply to message #17029] |
Sat, 11 September 1999 00:00   |
meron
Messages: 51
Registered: July 1995
|
Member |
|
|
In article <7re25a$dvm$1@news.doit.wisc.edu>, "Liam Gumley" <Liam.Gumley@ssec.wisc.edu> writes:
> I believe there is a market for either an add-on Mathematical Toolbox, or
> preferably built-in access to a selection of routines from a well-respected
> mathematical library like BLAS, LAPACK, CMLIB, NAG etc. For example, NAG
> developed an add-on library for Matlab:
>
> http://www.nag.co.uk/nagware/NN.html
>
> I think many people would be more than willing to either upgrade their IDL
> version, or buy an add-on toolbox, if it gave them access to a set of
> high-quality numerical routines. A user survey would no doubt tell RSI very
> quickly which routines people would like to see (Bessel functions and random
> numbers have been mentioned).
>
Well, I'm using my own math routines for anything that really matters,
at least I know what's in them and how far I can trust them. Some
significant problems with the IDL math routines I can think about off
hand (other than outright errors, at times, are:
1) Special functions only working for real input, not complex.
Downright tidiculous implementation since usually same algorithm that
works in the real domain will work in the complex one as well.
2) The implementation of double precision borders on the fraudelent.
What I mean is, most special functions accept a keyword /DOUBLE and,
when set, will return a double precision result. When one checks the
output, though, one finds that this is still a result of a single
precision calculation, only recast to type DOUBLE. This is worse than
being said on the onset that double precision is not available. And,
again, it is ridiculous since same algorithms that work for single
will work for double as well.
Mati Meron | "When you argue with a fool,
meron@cars.uchicago.edu | chances are he is doing just the same"
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
