On Thu, 15 Sep 2005 22:34:21 +1000, in php.internals [EMAIL PROTECTED]
(Leigh Makewell) wrote:
>The people who submit bugs are helping you make the product better. You
>should never insult a person for trying to help you. Tell them you are
>unable to reproduce the error and then point them to the relevant
>documentation on how to give you better information. In other words,
>help them help you!
This is a very delicate situation.
Of course bug submissions are created in the first place for improving
the project. The main problem in many large projects (e.g. php, mysql,
apache as they all have bug reporting systems that resemble each
other) is that at one point the threshold for actually bogus bug
submissions reach a critical mass.
At that point the work on the bug database changes from primary being
a benefit to simple end user support. At the exact same point the bug
administration work seems irrelevant and nobody wants to do it.
There is some rationale in this though. Those people's time could be
more well-spent just developing on PHP instead of checking bug
reports.
The problem is of course that the number of relevant bug submissions
is probably still the same, so there surely are important bug
submissions. But the number of bogus bugs are greater than ever
before.
I don't think there is any easy solution. If those looking at bug
reports weren't doing it, there might probably be no-one to do that
work, and then the relevant bugs wouldn't be looked at.
Comparing php bug reporting to mozilla and mysql (the latter being a
bit unfair comparison with professional workers) I think that mozilla
bogus bugs have better references (to dupes, etc.) and mysql bogus
bugs are even more polite in their template answers.
Some of the stuff I see on bugs.php.net that I'm not too happy about
are "."-answers (maybe just to satisfy a requirement of some kind of
answer) and "This is not a bug"-templates several times in a
bug-thread.
When dupes are reported there might be a point by making a
reference/link to that bug. It would help the bug submitter and more
importantly help future people who find the dupe bug (which might have
a better summary to search for than the original bug). But of course
it would take more time. I guess there is also a bit of "internalitis"
about it ("We have discussed it... sometime... somewhere... with some
conclusion... search the archives!" :-)
When bogus bugs are reported, this is a case between bug work and
simple end user support. I wouldn't say that the user always "should"
be helped, but a link to the specific section/function in the manual
might help. Anything might be better than "."
To quote Poul-Henning Kamp in a tongue-in-cheek post:
http://www.bsd-dk.dk/archives/2002/Oct/0105.html (in Danish,
translation follows):
==
RTFM is in fact a protocol but some don't know the implementation.
Just sending "RTFM" is an error in the protocol that recipients should
just ignore. The correct usage is to indicate what FM that contains
the information:
RTFM: disklabel(8), games(6) and/or uniq(1)
The same goes to STWF
==
I have experienced the second issue where the "This is not a bug"
template is misused. First the template and a reference to "you should
use this instead". Then a reply from me that the problem was also
present under these conditions. That was followed up by even another
"This is not a bug.."-template and another wild comment. It is hard
not to see that as arrogant as the first comment obviously was wrong.
I suppose theses situations make some people think that "Bogus" is the
default option. It's the equivalence of calling an ISP, listening to
their first level support (that would like to ask for the size of your
mouse when their mail server is spewing out errors), wishing to say
the magic word to penetrate through and getting in touch with a
technician instead.
Furthermore there are the attacks like "Try this on a real OS, like
Linux" (bug #34483). I don't get the "Running a PHP for 24hours (under
windows) is REALLY not supported or suggested." as well . I hate when
some users in cross-platform projects (like PHP, Apache, MySQL,
Mozilla, etc.) assume that stuff should only work on their platform.
Of course, all of the above is really more easy to notice than to
implement and more just an observation than solutions. Some of the
suggestions might even be patronizing ("you should reply in a good
manner, not in a bad manner").
I suppose there are reasons for many things. I suppose a lot of the
.-answers are when there are many open bugs and a nice developer
decides that he would spend the evening on reviewing/closing these
bugs instead of watching a good movie.
I guess my only real suggestion is that some of the template-answers
could be better worded.
--
- Peter Brodersen
--
PHP Internals - PHP Runtime Development Mailing List
To unsubscribe, visit: http://www.php.net/unsub.php
