> I think you underestimate both the number of EBCDIC systems extant,
> and the number of cars driving on the left. Neither is by any
> stretch of the imagination a niche, and neither is going to go away
> in the forseeable future.
> Tony H.
Fortunately, mainframe gateway protocols confine the former to the
mainframes, and Chunnel gateway protocols confine the latter
In some ways, this is wandering off the subject, but in
others, it's right on the mark. First, Ken, driving on the
left is hardly confined to Great Britain. It also is the
method used in Japan, Hong Kong, Singapore, and Ireland.
I'm sure others on the list will supply additional examples.
I almost changed the Subject: line when I started this reply,
because it seemed we were going too far away from how universal
Unicode will be in the future. Glenn Adams had predicted that
all systems will migrate to Unicode as their default encoding;
I said otherwise, and offered examples of where problems have
NOT converged on a single solution.
To expand on Tony's point, I think many are underestimating
how entrenched various solutions are. Almost eight years ago,
I worked for one large computer company, and it was trying to
figure out how it would move customers from its proprietary
Western European encoding over to ISO 8859-1. Most agreed the
standard code set was preferrable, and plans were put in place
to make the migration.
Eight years later, the migration is not complete. Customers have
software and data that uses the proprietary encoding, and they
may see no advantage to making a change.
Technology changes with blinding speed, but plain old everyday
users may change at glacial speed. Just as left-hand driving and
EBCDIC are not as confined as some may think, other encodings are
not only cowering in niches waiting to be swept away by a convergence
to a single encoding. There's no doubt companies and countries are
reducing the use of proprietary encodings, but reducing does not
mean the same thing as eliminating.
Sandra Martin O'Donnell
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:36 EDT