Re: underscores
ian@chiark.chu.cam.ac.uk (Ian Jackson) wrote on 28.05.96 in <[🔎] m0uOE4q-0002aDC@chiark.chu.cam.ac.uk>:
> Kai Henningsen writes ("Re: underscores"):
> ...
> > Why do we need two versions of the file name?
>
> The architecture is quite long, and it is totally redundant when the
> filenames appear in the per-architecture archives.
In comparision to the rest of the name, the architecture portion is
actually quite short.
> It seems to me to be a good idea to try to save people who have to
> type filenames (not everyone has completion all the time !) or scan
However, they usually either have wildcards, or need to use 8+3 filenames.
The cases remaining after that are, IMHO, _very_ few.
> directory listings (will use up more space and be more cluttered) the
Not appreciably, at least to me.
> If we design the filename structure right then it can be totally
> unambiguous whether an architecture is present.
That's not the problem I'm trying to solve. (And actually, that was never
a problem; all you need for that is the short list of all allowed values
for that field.)
> > Why not simply leave the name as-is? It seems to me that would reduce
> > possible confusion and even make those scripts simpler.
>
> As I've said before, we should be designing for people, not for
> computers - computers are here to serve us, not the other way around.
Actually, if you look at my argument again, the benefit for the scripts
was a lucky side effect, not the reason I prefer it that way. The main
argument was designing this for people - IMHO, there is simply no good
reason to have the same file with two subtly different file names, and
there is a very good reason not to.
Keep things simple. One file name is enough. And it certainly eliminates
the possibility of typing in the whole name in the wrong variant, only to
get an error message.
MfG Kai
Reply to: