[clean-list] IEEE floating point and Clean

Siegfried Gonzi siegfried.gonzi@kfunigraz.ac.at
Wed, 12 Sep 2001 13:04:06 +0200


M.Kesseler@aia.nl wrote:

 
> I don't think - although he's quite verbose - that he is exaggerating the
> problem, but it IS questionable whether the problem affects you.

It doesn't affect me (at least not yet). But I think such papers as
Kahan's one could scare away (not me; I would be a fool reacting this
way) people from one language. I mean not the masses (as it is true for
Java) but scientists. 

On the other side, I would be interested in, whether people strongly
disagree with Kahan. Kahan has got a name in the field, but that doesn't
mean he is an institution.


As I understand Kahan now. He wants that a computation doesn't interrupt
but the user should get the possibility to set a flag do work off the
place where the infinity/over/underflow... occured.

An obtuse question: Why isn't this possible with Clean or are there any
well planed reasons to implement Clean as it is now? It is likely that
the Clean guys were aware of exception-handling.

Clean team, please do not get depressed, I still like Clean. But I think
some themes are not discussed or unspoken and this could make confusion.
You may not forget, there are a hundred of "silent" readers who maybe
get a wrong impression.

Meine Verehrung,
S. Gonzi