Markus Kuhn wrote:
> Full ISO 10464 will not replace ASCII in the
> next 30 years in those 90% of applications that are not special
> i18n word processors.
Actually, most of Unicode (as opposed to "full ISO 10646") will probably
be adopted *first* by those "90% of applications that are not special
i18n word processors." The reason is that most applications that are not
focused on i18n specialty work will simply use the local OS's default
text encoding for purely local work, and cross-platform encoding
standards for networked data. Before long, both are likely to be
Unicode.
Most OSes are likely to be Unicode-based within the next few years
(Microsoft says Win98 will be their last non-Unicode OS and they're
encouraging people to consider upgrading to NT rather than Win98, WinCE
is Unicode, Apple's Rhapsody is Unicode, anything Java-based will be
Unicode, etc.), and since "network" already means the global Internet
most of the time, it's unlikely that a European-only encoding standard
will be widely implemented for Internet-based work, even in Europe.
As a result, long before your 30-years passes, most new applications
will be using Unicode by default. Using anything *else* will be the
thing that will require extra effort. I agree that most application
writers aren't going to want to implement bidi and composition
themselves, but a lot of this will be provided as system services by the
OS, and even where it's not, the app developers can ignore bidi and
composition in their apps and *still* use most of the scripts in
Unicode. Why would they artificially limit themselves to Western
customers who only communicated with other Westerners when they could
also have pre-composed, left-to-right Chinese, Japanese, and Korean, for
example, with just the addition of the appropriate fonts, by using the
default: Unicode?
I've observed that the fact that UTC members and other specialists spend
most of their time wrestling with arcane linguistic details on the
frontier tends to lead a lot of developers to think that using Unicode
in their apps will require them to do the same. We need to spread the
word that if you're one of those "90%", it's just not true. As a
first-order approximation (all you'll need for most apps), developers
can think of Unicode as just a bunch of additional pre-composed,
left-to-right characters. They just have to watch out not to assume that
a character == a byte (which they'd have to do for MES, too) and they'll
get most of the world's scripts for the same effort required by ASCII.
Defining a subset to make sure the user has necessary font/glyph
coverage may be helpful, but creating another encoding that needlessly
limits your apps to these scripts only, when you could have most of the
world's scripts for the same effort if you just used Unicode, is
counterproductive.
__Glen Perkins__
glen.perkins@nativeguide.com
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:35 EDT