|
ABSTRACT
Most computer facilities have “program libraries.” All too often the library is inadequate because: 1. The latest and best algorithms are not used. 2. When several routines are available which purport to do the same job, there is little or no indication as to which one is best, and under what circumstances. 3. Certain basic programs are completely missing. 4. Certification is haphazard. Routines are not, and should not be, trusted. (By certification we mean a stamp of approval that someone other than the author has tested the program and that the tests were “successful.” Certification must be accompanied by documentation of the tests.) 5. Documentation is inadequate: the method used is not divulged, values of fundamental constants are not given, limitations are not properly described. 6. Often the routines are incompatible with respect to calling sequence or data format and cannot be used together.
REFERENCES
Note: OCR errors may be found in this Reference List extracted from the full text article. ACM has opted to expose the complete List rather than only correct and linked references.
| |
1
|
SHARE SECRETARY DISTRIBUTION SSD 174 U-79 Dec 1967
|
| |
2
|
N A CLARK W J CODY K E HILLSTROM E A THIELEKER Performance statistics of the FORTRAN IV (H) library for the IBM system/360 Argonne National Liabrary Report ANL-7321
|
| |
3
|
SICNUM NEWSLETTER (published by ACM Special Interest Committee on Numerical Mathematics) Vol 1 No 3 February 1967 p 4-5
|
CITED BY 6
|
|
|
|
|
James M. Boyle , William J. Cody , Wayne R. Cowell , Burton S. Garbow , Yasuhiko Ikebe , Cleve B. Moler , Brian T. Smith, NATS a collaborative effort to certify and disseminate mathematical software, Proceedings of the ACM annual conference, p.630-635, August 01-01, 1972, Boston, Massachusetts, United States
|
|
|
|
|
|
|
|
|
|
|
|
|
|