Unicode comprimises

From: Dan Kolis (dank@hq.lindsayelec.com)
Date: Tue Feb 27 2001 - 11:13:32 EST

Hi to all,

My best second language is manual communication; (the many variants of sign
language), and all the time I try to explain to people the massive subtle
issues in how its deployed. The listeners eyes glaze over.... Human language
is so complex when you dig under the surface.

One of the most interesting things about the unicode stuff is how it is
forced to understand all sorts of conventions in written language just to
get on with the codepoint management.

You know what I mean... trying to grasp how diacritical marks you have never
seen work, arbetrary issues in the direction of written text, collation
sequences, etc.

But in its essence think of this thought experiment. You have a manual
typewritter and somebody takes 3% of the keys and disables them. Maybe they
hit a symbol you really need; a common letter, punctuation, etc.

But, can you still compose useful messages on it? Yes. Is it better than
having to build a typewritter from drawings as a one-of? Yes.

What I'm saying is to a purist, whether CH is kerned or two symbols, whether
an armenian infinity sign is included or not, is often a big, personal deal.
But even if a few percent of the unicode symbols where simply absent or
wrong, the entire initiative would still be worthwhile.

There are probably relatively few large scale projects which have the odd
property that even done somewhat badly there still very useful. For
instance, in the human genome initiative a guidline exists called the
Bermuda criterea and all the labs sign up to do it. It means the data is 99%
correct or its not used. They plan to potentially increase that requirement,
and that will be Bermuda II, or whatever it will be called. The 1% residual
error in that field is so high it makes the data close to unusable for many
proposed purposes.

On the other hand, there is some Maritime spec for oil tankers which is some
big fuzzy algorythm for swapping out single hulled oil tankers for safer
double hulled ones. Probably will effectively never totally eliminate single
hulled ships. But if 95% of those that crack up, don't leak its an amazing

What I'm suggesting as an interested outsider is the unicode overarching
goal is very difficult and I think its gone well and continues to go well.
The fact that people feel passionate enough about sub issues is a good
thing. If you loose your skirmish on a sub issue, there's no reason to give
up. You can win your points another day, but only if you continue to try!


Even done with residual boo boo and omissions, the initiative is clearly
enabling, worthwhile, etc even if a little Klingon and off center CJK sneaks
in once in a while.


This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:19 EDT