=========================================================================
Date: Thu, 18 Jul 2002 19:44:41 -0400
Reply-To: Quentin McMullen <QuentinMcMullen@WESTAT.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Quentin McMullen <QuentinMcMullen@WESTAT.COM>
Subject: Re: Division by Zero/Missing Value Errors
Content-Type: text/plain
Jim Wrote in part:
> Having a simple global switch that users could turn on to ignore
> division by
> zero errors only is a great idea that SAS should consider including in
> their
> product (i.e., work like nonotes, etc.). Sure, one can get around this
> error with lots of code, but that's true with just about anything.
> Might as
> well make it easier (more useful, more productive).
Hi all,
My tendency is to lean towards Don's reasoning on this, certainly it's not
an option I would want to turn on. But I guess that is not reason enough to
be against it.
In fact in general, I'd be supportive of options that give the user more
control over what sort of message level (i.e. none/NOTE/WARNING/ERROR)
should be reported for particular events. I would love to be able to make
uninitialized variables, automatic conversions, etc. into a WARNING.
I wonder Jim if you would write a bit more about how you would see this
option working. Do you need an option specific to division by 0? Or would
you consider a more general option that would drop notes from all such
invalid arguments, e.g. sqrt(-1), log(0), etc. ?
Seems to me if you want to offer such control to the user, it probably makes
sense to think in terms of logical groups of messages that could have their
message level determined with a single option. Of course, if SAS had error
numbers identifying each message, I guess it wouldn't be too hard to give
the user explicit control over each individual message.
Kind Regards,
|