bug reports

["Andreas C. Lemke" <andreas@BOULDER.COLORADO.EDU>]

I guess, it's ok to forward this message from Johanna:

Date: Wed, 1 Nov 89 14:16 EST
From: Johanna Rothman <JR@YUKON.SCRC.Symbolics.COM>
Subject: bug reports
To: andreas@boulder.Colorado.EDU, JR@YUKON.SCRC.Symbolics.COM
In-Reply-To: <8911010011.AA06934@sigi.colorado.edu>

    Date: Tue, 31 Oct 89 17:11:13 MST
    From: Andreas C. Lemke <andreas@boulder.Colorado.EDU>

    [...]

    Actually I forgot to ask one question:  How many doubles do you get? I
    mean, how many incidents refer to bugs that were already reported?

I don't have any data to back up my gut reaction, but I think it may be
as high as 10%.  Remember, this is an almost useless number, because I have
not analyzed the reports.

    An answer to this question would say something about the extent of the
    problem of programmer's wasting time to repreatedly track down a
    known problem.

One of the issues in determining if a bug is already known is to be
precise about the bug report.  It is sometimes quite difficult to reduce
the problem to the smallest possible case.  Frequently, that is what is
necessary to really see if this problem is the same as a known problem.