Bug#302762: acknowledged by developer (cooledit removed)
On Sun, 01 May 2005 12:18:11 -0700
owner@bugs.debian.org (Debian Bug Tracking System) wrote:
> It has been closed by one of the developers, namely
> Matej Vela <vela@debian.org>.
> ...
> cooledit was removed from Debian unstable in January 2004. See
> <http://bugs.debian.org/229615> for further information.
Closing unfixed bugs for all removed packages is a bad idea, and arguably violates Debian's Social Contract which says:
We will communicate things such as bug fixes,
improvements and user requests to the "upstream"
authors of works included in our system.
I believe I understand why closing a bugs like this is done -- nobody maintains them, therefore there's no Debian maintainer to forward them upstream. Also Debian's BTS is used for statistics and planning, like a census; counting thousands, perhaps tens of thousands of orphaned bugs would create a misleading picture of the distro's current status.
The problem is real; but closing all open bugs is a poor solution...
Blowback 1: suppose package 'foo' is orphaned, and all its bugs closed. Later some new maintainer adopts 'foo'. Where are the orphaned bugs? In the archive. Some may still be valid. The maintainer might not have the time to search every archived bug just to find out which were closed due to the package's being orphaned, and if he did have the time, it still WASTES his time in having to review them when the bugs simply could have been left open.
Blowback 2: suppose a conscientious upstream package author occasionally reviews various downstream package BTS's for new bugs, in case nobody downstream remembered or cared to report those to upstream. Would such a conscientious author have time to review all the archived closed Debian bugs just to find the minority of still valid orphaned bugs? Even if he would, this would WASTE their time -- thus violating the spirit if not the letter of Debian's Social Contract.
Blowback 3: User Bob finds bug #X in 'foo'. Then 'foo' is orphaned, then months later adopted by a new maintainer who lacks the time to deal with Blowback #1. So bug #X remains closed, even though it's not fixed. Later user Betty finds the same bug in 'foo', and checks the BTS to see if its already been reported -- she finds nothing open, and reports it as bug #Y in 'foo'. It WASTED user time to report an already reported bug.
Suggested fix: create a new bug status between "Open" and "Closed" -- call it "Limbo" maybe. "Limbo" bugs would be any bug that was "Open" when orphaned, or any bug reported after a package was orphaned. Programs that tabulate bug statistics could consider "Limbo" bugs as "Closed" for most purposes. Upstream maintainers and users would still see them as virtually "Open" for most purposes. When any orphaned package was re-adopted, the "Limbo" bugs could be changed back to "Open".
NB: I've noticed many such closings of other bugs before, and the maintainers themselves seem to WASTE a lot of their time closing 'em. On 4/4/05 I wrote Colin Watson, the current BTS admin, to ask how he feels about a consequence or relation of this problem:
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=302814&msg=9&att=0
So far no reply, so I'll now reopen that bug just to keep the message in play. I think the whole problem of "a bug is EITHER closed OR open" constitutes a serious bug. It wastes the time of Debian Maintainers who adopt packages, and Debian maintainers who close such bugs and reports, not to mentions upstream authors, and last and perhaps least and perhaps also most numerous, users who don't know about these things and mistakenly send redundant info.
Reply to: