Date:         Tue, 9 Apr 2002 13:11:01 -0700
Reply-To:     Cassell.David@EPAMAIL.EPA.GOV
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "David L. Cassell" <Cassell.David@EPAMAIL.EPA.GOV>
Subject:      Re: Are You a SAS Expert Test?
Content-type: text/plain; charset=us-ascii

SAS Master Ian wrote [in small part]:
> I am always interested in the why, so why can so many things go wrong
in a
> SAS merge?  Perhaps it because the code is so simple, or rather so
simple in
> comparison to what it does.

Or perhaps it is because so many things can go wrong
under any circumstances, when we use a programming
language to manipulate data which may [or may not]
be what we expected.  Consider the following code:

z = x + y/w;

Now list 25 things that can go wrong here.  You
can apply a lot of Roger DeAngelis' "43 Things That
Can Go Wrong In A Merge" list here too.  If x, y, and
z come from different data sets, then all the complexities
of merging/modifying/updating apply here too.  And don't
forget all the issues of roundoff error, division
by zero, numerical instability, loss of precision,
_et_cetera_ ...

Perhaps the moral is that what we do is not anywhere
near as simple as we like to think.  After all, if this
sort of programming were really so simple, why would
anyone ever buy a programming book with the word "Dummies"
in the title?  :-)

David
--
David Cassell, CSC
Cassell.David@epa.gov
Senior computing specialist
mathematical statistician