heather@MATH.UCLA.EDU (05/28/88)
Please forgive me if this question has been asked before. One of the staff members here at UCLA is porting a statical analysis application from an IBM mainframe to his departmental VAX. The erf() function, which is supplied with the FORTRAN compiler on the mainframe (VS FORTRAN), is not present under VMS FORTRAN. We believe that this function is part of the FORTRAN 77 standard and are therefore unsure why it is not supplied. Be that as it may, DEC Tech Support has claimed no knowledge of this function. Have any of you come across this problem? How did you solve it? We heard that the solution may be found in the DECUS library distribution from 77/78 but in perusing the catalog that describes the contents of that distribution have found nothing that obviously solves the problem (although there are some possibilities, like a STAT subroutine package, that we can check in to). Anyway, if any of you could tell us how we can get source code to this function, written in FORTRAN of course, or, worse, how to buy it, we'd be most grateful. Please reply directly to me as I am not a regular reader of the newsgroup. I will be happy to summarize to the group. Heather Burris Advanced Workstation Consultant
ok@quintus.UUCP (Richard A. O'Keefe) (05/29/88)
In article <12789@shemp.CS.UCLA.EDU>, heather@MATH.UCLA.EDU writes:
[Asking about the ERF() function.]
I have replied by E-mail, but I thought it might be worth following-up
anyway.
An obvious place to start looking is CALGO. (The collected algorithms
from the ACM.) The Computer Journal (the journal of the BCS) has an
algorithms series. JCAM [J. Comp & Appl Math] has some good stuff.
Applied Statistics has a series of Fortran-coded algorithms.
In this particular place, the easiest place to look is
Numerical Recipes,
Press, Flannery, Teukolsky, & Vettering,
Cambridge University Press, 1986
ISBN 0-521-30811-9
The answer is found in section 6.2, pp 163-164.
It would be interesting to know what the relative quality of the VS FORTRAN
version and the Numerical Recipes version is.
Perhaps we could have "readily available Fortran source code" as a topic?
I've seen at least two other journals which regularly published Fortran code.