🚀 go-pugleaf

avg@rodan.UU.NET (Vadim Antonov) writes:
>The problem is not in searching -- the problem is in presenting
>the information and in regular expressions ([a-z] - does it include "o?)

Wrong question.  The correct question is: Which "o letters does [a-z]
include, if any?

Germans have "a after a and "o after o (or so I remember from when I
knew some German, about five years ago), Swedes and Finns have them
after z.  So, should [a-z] include the German letters or not?

The answer is that it depends.  If I'm using the German alphabet, they
should be included, but if I'm using the Swedish one, they should not
be included.

So, how does the stupid thing know which alphabet we're using?  Either
we tell it to it explicitly (which Vadim seems to abhor), or we use
completely separate codes for each and every letter of each and every
language, so that the Swedish a and z are different from the German a
and z.  The latter solution is ugly (IMHO), since it causes things
that are universally considered to be one to become many and that only
makes trouble.  The latter solution also is not going to happen, I
think.

>Let's solve simplier problems first. I merely want a character set
>which allows me to use my screen editor without fussing around
>every search pattern i use.

If I understand your proposed solution correctly, it is the "each
language has a completely different set of character codes" solution
which I have disagreed with above.

Now, I can easily understand that it works when the two languages are
like English and Russian, i.e. they use completely different alphabets
that only happen to have a couple of letters that _look_ the same (but
are still quite different characters, from what I understand; I have
no objection of the Unicode designer being flogged if they mix
Cyrillic and Latin letters :-).  It does not work very well when using
languages that have almost the same alphabets, with only a few letters
differing, such as Swedish and German.

Just the input problem is enough to kill this idea, I think: The user
would have to get the language of every character he types correct,
and as somebody who routinely (as in usually large amounts every day)
enters text in several languages (English, Swedish, and Finnish, with
a few quoted words every now and then from a number of other
languages, including German, Latin, and French) and often enough mixes
several languages in one sentence, let me tell you that it is _not_
going to succeed.  Having to switch language mode every few words, or
for every different document, is not going a workable solution.

The user is also going to have a hard time differentiating between
identically looking letters in different languages.  On fancy graphics
terminals one might use fonts or something to aid this, but that does
not exactly make for WYSIWYG, or even minimally good looking screens.
There is no such option for those of us using text-only terminals.

--
Lars.Wirzenius@helsinki.fi  (finger wirzeniu@klaava.helsinki.fi)
   MS-DOS, you can't live with it, you can live without it.