On Fri, 17 Jan 1997 John Plaice wrote:
> The private positions are fine if they are used internally by me and my
> close correspondents, and we all concur on what those private positions
> actually mean. But if some other group, using the same approach, but
> with different values for characters, uses the same spaces, and then there
> is communication between us, we will need to distinguish the different
> versions of `Unicode+private extensions'. Now if there is no way to
> identify this in Unicode, then ad hoc measures WILL need to be used
> to be able to distinguish the two.
Neither ad-hoc nor official measures for identification solve the
basic problem. For an examlpe, let's have a look at ISO 2022, which
has ample possibilities for code switching. To identify code tables
with escape sequences, you have to either register them (for this
you need an approved national standard or so, so you will have to
do a lot of paperwork) or you can use a private sequence, in which
case you have no guarantee that there are no conflicts.
So you see, the problem is not solved, it's just moved one level
higher. Better to solve it directly than just to move it around.
> Standards development is a very important process. But we should never
> forget that everything changes (Heraclitus), and that versioning is a
> perfectly normal situation. We have even seen versioning take place
> in the Korean space WITHIN Unicode, so I do not see why we should suppose
> that PRIVATE spaces will not themselves be versioned and that standardized
> means should not be available to distinguish/separate them.
The Korean mess (as I call it) is a very important example. It
showed to everybody the big problems of reallocating codes, and
helped strengthening the firm consensus that no code reallocation
should ever happen anymore. From now on, versioning in Unicode/ISO
10646 has to mean extension only, and this can be dealt with without
version identifiers.
Regards, Martin.
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:33 EDT