Re: minimum package name length
On Thu, 2004-12-23 at 11:16 +0100, Frank Küster wrote:
> Tilman Koschnick <email@example.com> wrote:
> > Hi,
> > I'm intending to package the Cluster Command & Control tool suite ,
> > commonly refered to as C3. My package name of choice would be c3, but I
> > was wondering if such short names are acceptable. Is there any
> > documentation or general consent?
> I don't know whether there is general consent, but there are usually
> some people on -devel who speak up when a new package is planned with
> a two-letter name. I think the argument is that the risk of a future
> name clash is rather high, and existing packages with only two letters
> are just there because they have been there for a long time, and it
> doesn't make sense to change the names now.
> Indeed a search on Google for "C3 software" gives at least a Chemistry
> application (Windows), Cluster Command and Control, an Accounting and
> Managment system for Associations and Communities (Windows) and some
> educational software written in Java, with Websites designed for "a
> more recent version of Microsoft Internet Explorer" and a non-working
> free (as in beer) download.
> Therefore I'd suggest that you choose a longer name, but YMMV.
> Regards, Frank
I'm going to split this package into two binary packages anyway,
c3-master and c3-node, to be installed on the cluster's head node and
all the other nodes respectively. A good name for the source package
would be c3-tool-suite, I guess, thereby avoiding future name clashes.
One question though: Is there any problem if my
c3-tool-suite_4.0.1.orig.tar.gz unpacks into c3-4.0.1/? dpkg-source and
dpkg-buildpackage don't seem to have a problem with it, but are there
any other pitfalls with differing upstream-tarball names and