Article View: comp.std.internat
Article #692islands
From: Erik M. van der Poel
Date: Sun, 10 Jan 1993 06:23
Date: Sun, 10 Jan 1993 06:23
51 lines
2659 bytes
2659 bytes
> plan 9 is inherently very distributed and so we see
> the difficulties of teh island approach earlier or more clearly.
The way I see it, Plan 9 is itself an island. The island may become
very large if enough people see that UTF is a good solution for
predominantly ASCII environments. The island may even subsume the
Latin-1 communities if they haven't invested too much already in their
code. I wonder how successful UTF will be in Japan.
> on the other hand, i'd be the last person to advocate massive system
> upheavals just for the sake of using 10646.
Yes, in some cases it's too close to an "all pain, no gain"
proposition, especially since the need for truly multilingual text
isn't as great as that for one's own language (plus English).
> if you are willing to be
> an island, and for many systems it may be the practical choice, then
> all power to you.
Bell Labs was willing to be an island, and UTF was a practical choice,
since most of their text files are ASCII. That is why Plan 9 could
take the UTF approach. (Congratulations, NOT flames.)
It will be interesting to see how organizations with truly distributed
systems (as in "selling hardware all over the world") try to cope with
this Unicode thing. For example, Sun may not be able to take the
all-the-world-is-UTF approach, since they have a fairly large number
of customers using non-ASCII text. Two major examples of this are
Japanese EUC and European EUC (aka Latin-1).
They currently use environment variables (e.g. "LANG") for "file
typing". This is how they can use the same "vi" binary in both Japan
and Germany. One conceivable approach is to extend this rudimentary
file typing to actual file attributes ("groan, not this guy again" I
hear all of you saying), extending NFS as well with a negotiated file
type option.
Then again, that may be too much pain too. Like I said, it will be
interesting to see what they do about Unicode.
I read that Microsoft is planning to support both Unicode and
Shift-JIS for Windows NT in Japan. I wonder if they have some sort of
file typing. Byte-order mark, perhaps? Anybody know?
Maybe Sun should go for such a "magic number" method?
-
--
Erik M. van der Poel erik@poel.juice.or.jp
Technical Details
Message-ID:
<C0MKEu.5or@poel.juice.or.jp>Path:
arc1-ca.pugleaf.net!sparky!uunet!zaphod.mps.ohio-state.edu!saimiri.primate.wisc.edu!ames!sun-barr!sh.wide!wnoc-tyo-news!hoffman!jmasgw!green!samrat!erikShow Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!sparky!uunet!zaphod.mps.ohio-state.edu!saimiri.primate.wisc.edu!ames!sun-barr!sh.wide!wnoc-tyo-news!hoffman!jmasgw!green!samrat!erik From: erik@poel.juice.or.jp (Erik M. van der Poel) Newsgroups: comp.std.internat Subject: islands Message-ID: <C0MKEu.5or@poel.juice.or.jp> Date: 10 Jan 93 06:23:18 GMT Organization: sometimes Lines: 51