Re: Can an intrinsic function be replaced in FTN95

In article <%AEEh.6221$iF.6109@trndny03>, michaelmetcalf@xxxxxxxxxxxxxx says...


Well, only to the extent that your processor can distingush -0.0 from +0.0
and -0 from +0. This is a very marginal change in most cases.


That's the problem. It is an huge numeric model where the original authors
decided to use sign for that wery purpose. There are about 150 places where
they do this. I wanted to write a replacement that behaved like the FTN90
version.

Anyhow, I'll try your EXTERNAL SIGN suggestion.

Thanks, Tom

.