Actually, I think there are remarkably few truly wrong answers, though it's
fairly common to see somebody misinterpret a question, or miss some
critical point, and therefore come up with a more or less correct answer
which unfortunately does not respond to the question.
As to following a thread, I think that's one of the weaknesses of the
archives in their present state. Threads are broken when Subject headers
change. That happens when people unwisely edit them by adding a superfluous
word or correcting a typo. It also seems that some e-mail software inserts
blanks as a side effect of wrapping long headers for display. Another
problem is the LISTSERV web interface, which does not recognize threads
which cross week boundaries.
On Tue, 23 Dec 2003 15:25:09 -0700, Jack Hamilton
>If someone follows a thread all the way to the end, then yes, the
>correct answer(s) will usually be found. But the Text Miner would have
>to be able to find "the answer you just gave was wrong, and here's why"
>messages, and not report the original wrong answers. That's what I
>think would be hard.
>And David's joking, of course: sometimes there are clearly and
>unarguably wrong answers on SAS-L. I've given them myself
>Manager, Technical Development
>Metrics Department, First Health
>West Sacramento, California USA
>>>> "Peter Flom" <flom@NDRI.ORG> 12/23/2003 2:17 PM >>>
>I think, actually, this is one of the great strengths of mailing
>SAS-L in particular: Mistakes don't survive.
>If you ask an "expert", he or she may get it wrong. Everyone makes
>mistakes (except Ian :-). People can also misunderstand questions.
>But, if you post a question here, and someone posts a wrong response,
>or if a question is confusing, things get straightened out, usually
>>>> "David L. Cassell" <cassell.david@EPAMAIL.EPA.GOV> 12/23/2003
>4:29:34 PM >>>
>Jack Hamilton <JackHamilton@FIRSTHEALTH.COM> replied:
>> An interesting problem with running text miner against SAS-L: how do
>> avoid returning incorrect answers?
>Come now. Surely you don't mean to imply that anyone ever gives a
>*wrong* answer on SAS-L? There may be sub-optimal answers, or
>not-well-worded answers on occasion, but none of the posters on SAS-L
>ever makes a *mistake* ...
>David Cassell, CSC
>Senior computing specialist