🚀 go-pugleaf

dik@cwi.nl (Dik T. Winter) writes:
>Wrong Vadim.  You cannot even do it in the European languages.  You cannot
>even do it in German.  How would you assign codes such that the German
>A-umlaut sorts as if it is the letter combination AE, and at the same
>time the umpteenth letter of Swedish (after Z).

(Second letter after Z, actually; the letter in between is
A-with-ring-on-top -- don't know its official name, if any.  The
Finnish alphabet has no A-w-r-o-t, so I guess they have A-umlaut after
Z, although they/we seem to usually include the A-w-r-o-t too, thereby
effectively using the same alphabet (we) Swedes use.  I'm a bit
uncertain of the official way, and I only have two dictionaries at the
moment and those are both for English (one by Oxford, one by
Webster).)

The immediately obvious solution to the problem with the German
A-umlaut not being the same as the Swedish A-umlaut (it looks like a
duck, but it doesn't walk like a duck, and it doesn't quack like a
duck, is it a duck?) is to assign them different codes so that you can
differentiate between languages.  Hm..., this could get messy though,
since then you get confused people: should a spelling checker for
Swedish accept the German a-umlaut in a Swedish word?  Should a
sorting program sort the Swedish a-umlaut differently from the German?
How many people are going to look at the 2^32 entries big table and
accidentally pick the wrong code, especially when they are not that
happy about foreign languages to begin with?

Personally, I think that giving two characters that look the same and
that are usually thought to be the same (as the German and Swedish
a-umlauts are, or the a-umlauts in other languages), even if they are
used differently in different languages.  A character encoding
standard should provide an efficient, simple way of referring to
different characters, not confuse things by trying to encode all other
possible information as well.  Single characters can be encoded
simply, what happens when they are used in groups to form words in
different languages is an other issue and should be dealt with
separately.  IMHO as a layman with only a little experience with I18N.
(I've written one program that was able to switch output languages --
Swedish, Finnish, and English -- thought that the issue was ugly and
unpleasant, if only because of the character set issues, and decided
to use English as the only language unless I have to use another until
I can understand the issues better and either hear of or come up with
a good, elegant and simple solution.  I'm still using English after,
what, five years or so.)

I don't know whether Japanese and Chinese have "characters" that are
the same in the way German a-umlaut and Swedish a-umlaut (or a through
z), since I know almost nothing of either language.  If the only
difference is that they are usually drawn slightly differently (for
all I know, the Germans might like to draw a-umlaut differently from
Swedes, by using a differntly formed umlaut higher up or whatever; its
still considered to be the same character), I see no reason why they
shouldn't use the same code.  But, since I'm linguistically
challenged, I'll have to take somebody elses word for it, though a
concise explanation would be nice.

Disclaimer: all my knowledge of the characters standards come from
netnews, a notorious source of unreliable information mixed with
absolute truth.  I don't even have a real stake in this game (like I
said, English is good enough for me almost all of the time, and
Latin-1 suffices even when it isn't), I just feel like rambling.  Take
it, leave it, or roast me via mail.

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