Re: DEC multilingual code page, ISO 8859-1, etc.

From: Erik van der Poel (
Date: Tue Mar 28 2000 - 14:17:30 EST

Frank da Cruz wrote:
> Yes, this is the conundrum we all face. In a more controlled environment,
> there actually could be a cold cutover, as there was, for example, when
> the ARPAnet switched from NCP to TCP/IP.

The Net was *much* smaller in those days, and even then, it was
difficult to orchestrate the big switch-over. These days, you wouldn't
be able to pull off such a stunt. There is a new protocol at the IP
level. It is called IPv6 or something like that, and it is at the same
level (roughly) as the current IPv4. I suspect that "they" won't have a
flag day to do a cold cutover to IPv6. No, they will have negotiation
mechanisms. Requestors will first ask responders whether they support
IPv6, and if not, they will fall back to IPv4. Please correct me if I'm

> The question is whether new applications should continue to do this. In
> my opinion, if an application needs to use characters that are not in any
> standard 8-bit character set, then that is the time to introduce Unicode.
> It is NOT the time to pump still more private code-page data into the
> public network.

That depends on the particular private code page. In the case of
windows-1252, there are far more users with mail software that works
with windows-1252 than UTF-8 (or any other encoding of Unicode).

There is an old saying "Be conservative in what you send, liberal in
what you accept". As far as windows-1252 vs UTF-8 is concerned, sending
UTF-8 is *not* conservative. We need to wait until more people have
UTF-8 capable software installed.

> A standard character set is one that is in the ISO Register -- the IANA
> registry doesn't count.

No, both the ISO and IANA registries do not count. Only the installed
base counts.


This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:00 EDT