On Wed, Mar 9, 2011 at 15:53, Alan Cox <alan(a)lxorguk.ukuu.org.uk> wrote:
> Not true. Computers treat the two cases as separate characters
because
> there is no mechanism of having different versions of the same
> character. However, humans treat the two cases as two forms of the
> same character.
For a small subset of Western European languages and subject to all sorts
of caveats. Perhaps we should ask the same of say accented v unaccented
letters (where the policy of being the same is very language dependant)
or about languages of the rest of the world (the ones actually used by
far more people than English variants).
Do there exist Fedora packages with those characters in the package
names? How would users who don't have the proper keyboard layouts
installed install such packages if yum _didn't_ support converting รง
to c?
In other words, your argument is against the idea of
case-insensitivity in theory only (as no packages with such names
exist), and actually supports my position in practice (as there would
be no other way to install such packages).
> The question then reduces to: should the yum interface be
designed to
> be comfortable for a computer to interface with, or should the yum
> interface be comfortable for a human to interface with.
That question has no bearing on case sensitivity of yum if you bother to
think about it a bit more. A "did you mean" matching interface is quite
different to internal case sensitivity.
I agree that a "did you mean" option would be nice, as would a
case-insensitivity flag. It does not have to be the default behaviour.
--
Dotan Cohen
http://gibberish.co.il
http://what-is-what.com