Interesting article, especially for those of us around 30(!) years ago
for OCO battles.
http://www.scientificamerican.com/article.cfm?id=secret-computer-code-threatens-science
Some comments are scary -- like first one, 35 years geological research
and not understanding this simple article. And others suggesting that
software results be verified by rewriting software and comparing
results. But not saying what to do when results compare unequal. Rewrite
again and accept two-out-of-three? Maybe five-out-of-nine would suffice.
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
