Re: OS 99

From: Paul Keinanen (keinanen@sci.fi)
Date: Thu Nov 12 1998 - 06:42:07 EST


At 17:07 11.11.1998 -0800, texin@progress.com (Tex Texin) wrote:
>It was an attempt to distinguish whether you could enter and display
>Unicode characters from the monitor or terminal, as opposed to only
>having system services that support Unicode internally (e.g. just for
>conversions) but could not be used for a true Unicode-all-the-way
>application.

While the question of what Unicode "support" really means is a really muddy
thing, your attempt to define it did not make it any better, unfortunately.

With a terminal generating Unicode characters and capable of displaying
Unicode characters would make most operating systems Unicode capable in the
sense that you can write Unicode text processing applications for it. If the
operating system support setting the terminal line(s) to 8 bit pass-all
mode, either UTF-16 BE or UTF-16 LE would be supported. Even the RSX-11 on
PDP-11 from the 70's would qualify with this definition. With UTF-7, almost
any operating system would "support" Unicode.

Regarding the system service (API) support, if UTF-7 or UTF-8 strings are
passed, does the system service "support" Unicode. The only thing that would
make a system service to _not_ support Unicode in this case is that the
maximum length of a string parameter is not long enough to accommodate the
longest UTF-7 encoding of a single Unicode character or if the UTF-7
encoding contains characters that are illegal in that parameter (e.g. + or -
or no lower case ASCII support).

>I hope that helps, it was a fuzzy question, because the vendors were
>being fuzzy in their statements about whether they "fully" supported
>Unicode. (In my opinion.)

I fully agree with this and with a bit of creative thinking (as my example
above) and with a little marketing hype, you could claim that nearly every
OS "supports" Unicode (even those written prior to Unicode), while in
reality the degree varies greatly from OS to OS.

While for some operating systems the interesting thing what parts support
Unicode, the interesting thing about more complete implementations is try
to find features that are _not_ supported in Unicode.

Paul Keinänen

 



This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:42 EDT