Re: i

[Marcus Brinkmann <Marcus.Brinkmann@ruhr-uni-bochum.de>]

Hello!

On Wed, Feb 11, 1998 at 12:09:28PM -0500, Raul Miller wrote:
> Marcus.Brinkmann@ruhr-uni-bochum.de <Marcus.Brinkmann@ruhr-uni-bochum.de> wrote:
> > I don´t want to imply that this makes sense, I actually find it quite
> > stupid, so we mustn´t follow upstream authors who choose "B" as a
> > name.
> 
> Are you saying that you want to preserve the name B for use by an author
> who has a good reason for using it?  Or, are you saying that B is such
> a bad name that you don't think *any* author should use it?  I think
> you're saying the latter -- but I don't think it's a part of the debian
> charter to pass esthetic judgements on the naming conventions choosen
> by other groups (except where this causes conflicts).

I meant the second interpretation - mainly because short names are likely to
be used by shell aliases or other customization. Or do you think that "B" is
a very clever name for an executable ?

However, there seem to be two things to consider:

- namespace pollution. This is to take care of too general names like
address, fax, text, edit, compiler, play, show, view, etc...

Those are likely to be provided by several packages or are ambigouus ("what
is to play" (a game, a sound file, what format, a video file, what format)).

- very short names (one or two characters). Those are likely to be used by
aliases or other local customization and are not very descriptive.

My favourite example is "ll". This is a common alias for "ls -l", and no
package should provide a binary called "ll".
 
> It *does* make sense to pass esthetic judgements on the naming conventions
> choosen by people writing debian-specific software.  [Including absurd
> packages which don't do anything but provide tens of thousands of 
> executables.]

Yes, this definitely. However, as you may guess, I would extend this to
upstream packages, too, because there is hardly any difference. My
argumentation was, that Debian can't care about "silly" assumptions made by
upstream authors. We change a lot of details, like file location. I don't
think it is confusing to change one letter filenames to something more
descriptive. Certainly this should be well documented etc... you get my
point.

There are only 52 possible one-letter file names. Some more if you also
allow numbers and other characters.

There are only 52*52=2704 two letter words. Debian has far over 1000
binaries.
 
> Here's an excerpt I wrote from bug report #18062.  Christian had raised
> an objection to the name B, and I was responding to him:
> 
> 
> > B is the command which adds files to an existing sam session. [Or, in
> > the debian version, creates such a session if there is not one already.]
> >
> > [sam is basically an immensly powerful version of ed.]
> >
> > B works the same both inside and outside of sam.  If you want to change
> > it, you should probably address re-building sam's command language
> > (which also happens to be the language used to communicate between
> > the X interface and the editor).
> >
> > Of course, if a higher-priority package requires the B command, then
> > we'll have to do something about the issue. However, note that sam is
> > the standard editor of the plan 9 operating system, so it's unlikely
> > that gratuitous debian changes will be accepted by the upstream author.

I don't know what the plan 9 operating system is. I don't understand if the
package will break in some respect when you change the name of the binary
(as you say one should rebuild the command language).

I have stated my personal opinion, and have tried to give some reasons for
it. I'm sorry if I can't explain what I think about it in best english, feel
free to ask if I was unclear.

I have also tried to summarize a little, so we can start to discuss things.
There are certainly more aspects to care of, so we should extend this
discussion and let the result become policy.

One note: I hope I didn't sound harsh in this mail or my last. I didn't
intend so. Please excuse any misleading or ambigouus expressions on my side.

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


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-request@lists.debian.org . 
Trouble?  e-mail to templin@bucknell.edu .