🚀 go-pugleaf

In article <berger.725045292@gram>, berger@imag.fr (Gilles BERGER SABBATEL) writes:
> dhd@exnet.co.uk (Damon) writes:
> >IE, don't be parochial and assume that because it solves *your* problem
> >it's more then moving from a 1% solution to a 1.1% solution...
> 
> Sure!  But the problem is that we could relatively easily move to
> ISO-latin-1, (many softwares appear to work with 8 bits chars) so that
> it would be a first step to experiment and solve the problems.  Whenever

Experiments along your lines are going on, in the sense that a lot
of old software *is* being used with 8 bits today, with or without
modification.  It's hardly even a 'first step', as I18N in some
form has been going on ever since vendors started shipping computers
across national borders.  So, isn't everyone happy then?

> First, ISO-latin-1 appear to require a minimal conversion of softwares,

Well, 'minimal' is not very well defined here.  It may be a minimal
change in terms of source code lines, but distributing the output is
a whole different matter.  Without any factual evidence for my claims,
I assume it would be easier to update every piece of software at a
single large site (say, MIT) than to get the updated version of a
single piece of software (say, sendmail) out to every user who would
need to install it (end-user or not).  Haven't we been through this
already?

> able to write japanese at all!  So, the use of ISO-latin-1 in France
> would not disturb anybody outside of France, and it contributes to the
> solution of the problem - at least it poses it!

Controlled experiments in a limited environment is fine; I'll leave
it to the French computer users to decide whether they want to take
part in that experiment.  However, I imagine some people outside
France want to make it quite clear that they are not going to support
whatever local conventions are established locally, especially when
they can provide superior solutions.

As for your statement that use of ISO-latin-1 in France won't
disturb anyone outside France, well, I prefer to remain silent
until I understand what it means, or I notice disturbances...

However, I don't see the real need for your experiments here.
I consider it an obvious fact that a serial line can handle 8
(or even 9) bits just as well as 7, so what is it that still
needs to be tested before you can go into production use? :-)
--
Anders Andersson, Dept. of Computer Systems, Uppsala University
Paper Mail: Box 325, S-751 05 UPPSALA, Sweden
Phone: +46 18 183170   EMail: andersa@DoCS.UU.SE