Re: Unicode Operating Systems

From: Adrian Havill (
Date: Tue Jul 15 1997 - 22:19:14 EDT

Kenneth Whistler wrote:
> Windows NT has been fully Unicode-aware, including Unicode text files
> almost since its inception, and now has Unicode editors available.
> This is sitting on my desktop now, in 1997. When Microsoft finally
> converges Windows 95 with the Unicode support in NT, this support
> should start to be universal for the Windows OS.

I'm sure you're including Windows CE, the handheld computer/device OS,
that is entirely Unicode.

This is significant, of course, to show that one doesn't need a massive
OS to handle Unicode (the Unicode by default JavaOS is also another good
example)-- the entire OS is 1.5 MB (including the hardware abstraction
layer) on a RISC (meaning fat binaries) CPU-- the kernel, including the
User and GDI, is 700 KB. I don't know how large the ROM footprint
increased with the addition of Japanese fonts, but I'm sure that it
mostly fonts, and not extra code, that increased the Japanese version's
ROM footprint.

And the porting speed due to Unicode is pretty obvious-- the Japanese
version of Windows CE 1.01 was put out amazingly fast, which reflects
four things IMO: 1) How important internationalization (well, at least
the Japanese portion of i18n) is to serious software developers 2) The
market no longer allows for you to release the English FCS first, and
then do the other languages (Japanese betas are appearing now before the
English FCS (e.g. Communicator), suggesting that finally something close
to simultaneous release is becoming a reality) 3) Unicode is the answer
that lets many do it quick and fast, and 4) Just because it's Unicode
doesn't mean it has to be massive (these are OSes for DEVICES, not
workstations) and complicated.

I realize everyone on this list already knows this, but I was so
impressed I thought it was worthy enough to once again praise i18n, m17n
trends in software and those using Unicode.

A recent Javaworld article noted that Mac users often complain that
JavaSoft managed to port the JDK 1.1.1 AND Hotjava to Japanese before
they could get a Mac release out. What the article doesn't realize is
that porting to other complex languages-- on the programming side at
least-- has suddenly become much easier (it's still hard though) due to

An interesting note about Windows CE: because CE's Unicode, parts of the
CE emulator in the Software Development Kit/Device Development Kit, etc.
are not available for the partially Unicode Windows 95. They're
available only to NT, its big, bad Unicode rappin' brother. (Source:

Even more interesting:

"Windows CE is a new operating system kernel and Microsoft expects heavy
international penetration. To this end, Microsoft made Windows CE
Unicode only. Due to the memory constraints on devices that run Windows
CE, Microsoft chose to leave out support for 8-bit characters entirely.
This means that you really must understand Unicode because you are
forced to use it when developing for the Windows CE platform."


Wow! 8-bit character support intentionally LEFT OUT. What's this world
coming to? (^_^)

Adrian Havill <URL:>
Engineering Division, System Planning & Production Section

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:35 EDT