🚀 go-pugleaf

hpa@eecs.nwu.edu (H. Peter Anvin N9ITP) writes:

>> There are at least a few real world applications (I don't talk about
>> trivialities like UNIX ls), e.g. catalogs of international libraries,
>> where an _easy_ to understand language independend sorting standard would
>> be extremely useful. Why not standardize one for Unicode in an easy
>> _implementable_ way? Then let's see, how many people will adopt it and
>> will quickly forget their traditional rules. Perhaps even many users
>> in Sweden. There are perhaps already some multilingual sorting rules
>> common practice, but I know no one that defines an easy implementable
>> deterministic and total ordering for ISO 10646 strings.

>For this application, I think English sorting order is rather well
>established.

What the hell is this "English sorting order"? Is it a standard that defines
a deterministic total order of all Unicode strings? Where can I get the
document that defines it? Perhaps we are speaking
about the same idea, an international DEFAULT sorting order, that will be
acceptable in many countries. There is no reason, why the USA and China
can't share a common sorting order, there won't be any collisions.
There won't even be collisions if you add France and Germany.
But I don't know that any has been defined yet. If people in Sweden need
a different order, than they should get their order, no problem.
But it would be a good idea to define a default order that might be 
easily refered to (e.g. as an ISO standard) and that satisfies the
needs of the greatest set of countries for which this is possible with
one sorting order. I am sure, that this is not possible with character
code based sorting, but that it is possible with a SLIGHTLY more
sophisticated algorithm. Why is everyone so uninterested in this idea?
I posted already many details of my suggested framework for this
standard. It consists of an algorithm and a big table that groups
characters that are sorted together (e.g. a, A, and all As with
diacritics are identified by this table to belong to one single group).
All of the many digit 7 versions in Unicode would be another group.
I don't have the experience to create the whole group table and the
(low priority ordering within the groups for totality of the relation),
but I hope, that the experts for CJK languages will also find suitable
sorting groups for these languages.

Markus

-- 
Markus Kuhn, Computer Science student -=-=- University of Erlangen, Germany
Internet: mskuhn@immd4.informatik.uni-erlangen.de  |  X.500 entry available
--- Wer, wie, was? Wieso, weshalb, warum? Wer nichts fragt bleibt dumm. ---