Re: what to do with `namespace-pollution'
On Wed, Feb 11, 1998 at 08:22:08PM +0100, Milan Zamazal wrote:
> >>>>> "MB" == Marcus Brinkmann <Marcus.Brinkmann@ruhr-uni-bochum.de> writes:
>
> MB: There is no need to keep a silly choosen upstream name. We
> MB: change a lot of things defined upstream (file location, etc),
> MB: and I don't think that changing a name from "B" to something
> MB: more readable is confusing (the opposite is the case).
>
> File location is no problem, most users don't care about it. Moreover,
> they expect e.g. documentation in certain directories on Debian
> systems. On the other hand they want to run the program by the same
> command on all systems. Moreover, makefiles, shell scripts, etc., may
> have problems, if the binary has different name on each computer.
Mmmh, I don't think we talk about standard packages here. I think most Linux
users (or even Unix users) would expect a "i", "l" or "B" binary *at all*.
I never saw so such names in make files, too. I am *not* proposing that cp
should renamed to copy ;)
> MB: Beside the reason you gave above, consider that a short name
> MB: gives *no* idea about the functionality.
>
> That's not true--`pl' (the main SWI Prolog binary name) is also the
> extension of Prolog source files (do not confuse this with the fact that
> Perl bigots chose already used extension for their source files) coming
> from ProLog. So it is clear for every Prolog programmer that binary
> `pl' has to do something with Prolog.
Sure, every rule has its exception. Most two letter names are well known, so
"cp", "rm" and "ls" speak a clear and well understood language. Again, I was
more thinking of "i", "l" and "B", I was not thinking of "X" or something
like this.
"pl" is fine for me, although it is not obvious for non-prolog programmers.
I would never think of suggesting renaming gcc to "this-is-the-gnu-c-compiler" ;)
But a program that is not yet standard of a well known programming language
or not a standard unix command (whatever well known and standard means)
should probably not be thrown in in the two letter arena.
> I repeat it again: I don't think it's a good idea to have few letters
> binaries, but I would rename already existing user programs only if some
> conflict arised.
I mostly agree with you, if you mean with "already existing user programs"
programs with at least *some* history. Note that there already exists "i",
"l" and "B" binaries (packages chris-cust and sam). Please check the
description of chris-cust, "i" and "l" don't seem to provide functionality
that you can't achieve better with an alias.
The one I could agree with is "B", but where shall be drawn the border?
Wondering,
Marcus
--
"Rhubarb is no Egyptian god." Debian GNU/Linux finger brinkmd@
Marcus Brinkmann http://www.debian.org master.debian.org
Marcus.Brinkmann@ruhr-uni-bochum.de for public PGP Key
http://homepage.ruhr-uni-bochum.de/Marcus.Brinkmann/ PGP Key ID 36E7CD09
Reply to: