On Fri, 27 Jul 2001, Steve Greenland wrote: > On 26-Jul-01, 19:50 (CDT), David Schleef <ds@schleef.org> wrote: > > I would strongly suggest a different package name than 'new', since > > it is too generic and misleading. The command name is, as well, > > but I think that is less of an issue. Well, I disagree. I think BOTH the package name AND the command name are generic, and bad choices. > I dunno. It's actually fairly natural to type 'new file.c'. The package Exactly my point. How many people do you expect to think like that and name their macro processors, source beautifiers, template processors and who knows what else "new" ? If the answer is not "none", you already know why we're warning you. > name doesn't conflict with anything else, and we seem to operate on a > first-come, first-serve basis. It matches the upstream name, which is > important. You are not supposed to propagate upstream <expletive ommited, I'm sure you can imagine a strong enough one by yourself> down the chain, you know. Debian maintainers are supposed to be far more than packagers; they're supposed to be good downstream maintainers, to alert upstream about the difficulties their code could and will cause in a distribution, and to work around such breakage where possible. You're right that we do operate on a first-come, first-serve basis most of the time -- and AFAIK, "new" could very well be yours to take by that rule. Still, that does not change the fact that using a command name such as 'new' is asking for namespace colisions (someone else is sure to have the same idea). BTW, I should add that IMHO, "new" is a bad package name because it is extremely non-descriptive AND a very common english word -- which is a pain for searches. I dislike the bad examples set by imagemagick among others. I think the cyrus-imapd maintainer showed a lot of wisdom in fixing upstream bad naming by prefixing the cyrus utilities with generic names with "cyr". Too bad the postgresql maintainer did not do the same... and the list goes on and on. > If you want to talk about misleading and generic, let's talk about > imagemagick's 'identify' and 'convert' commands :-). I would not expect one to shoot oneself on the foot just because someone else did it. If you do not want to change the names of your package and suggest upstream to do the same for the template processor, that's your prerrogative. But don't complain if someone files a bug requesting that BOTH be renamed later due to namespace colision. -- "One disk to rule them all, One disk to find them. One disk to bring them all and in the darkness grind them. In the Land of Redmond where the shadows lie." -- The Silicon Valley Tarot Henrique Holschuh
Attachment:
pgp7MEicK63qN.pgp
Description: PGP signature