On Thu, 30 Nov 2000, Toon Moene wrote:
> This is my every-day experience in maintaining a medium-large weather
> forecasting code: If it goes berserk, it could be I'm:
[snipped list of possible reasons]
> No matter how well I know the code (and I've really really studied it
> the last 8 years), I cannot *without thorough investigation* determine
> in which of the 6 classes the problem falls.
Yes, agreed. But the thorough investigation can be done using the
entirety of the code. But the last step, sending in a bug report, can
only be done after a lot more work - reducing the code sample to a
manageable size and pulling out all the external references. My programs
often use a large number of external libraries, often written in languages
other than Fortran. In the case of commercial libraries I don't have the
source code and the licences would not *allow* me to send in the object
code. And very often you find that in trying to reduce the problem to a
small self-contained program, the problem goes away. Sometimes this is
because the problem is in one of the external libraries, but mostly it
isn't. It just takes time, as I say sometimes days, to make the *right*
subset of the code that still shows the problem.
> > Yes, of course. But in turn vendors should help potential bug reporters
> > by giving them information on known bugs.
That's all I am asking. Some vendors used to do that so it's clearly not
impossible and though some descriptions were too vague to be useful, many
were not. But the number willing to list known bugs seems to be smaller
now, perhaps because they don't like washing their dirty linen in public.
But I have much more confidence in vendors who are willing to admit to
their (fixed) past mistakes, than in those who prefer not to admit to
having any.
Regards
--
Clive Page,
Dept of Physics & Astronomy,
University of Leicester.
|