Michael (michka) Kaplan wrote:
> From: "Marco Cimarosti" <marco.cimarosti@essetre.it>
> > Yeah. Pity that the local code page is the default
> everywhere, and to use
> > Unicode in the GUI one has to dig deep in options,
> registry, manuals, etc.
>
> Well, I would not go *that* far.... in theory just defining
> _UNICODE is all you need. How far an app is from that
> theory varies, but its a measureable (and costable!) issue.
I assume that you are not talking about the point of view of an end user or
a system administrator, as "defining _UNICODE" would mean nothing to them.
But, even for a programmer, it is not as simple as "defining _UNICODE". A
medium size project normally involves several aspects:
- Writing code in C(++);
- Writing code in other languages (including SQL or scripting languages);
- Using ready made software (editors, programming environments, utilities,
etc.);
- Using ready made code (libraries, DLLs, 3rd party controls, etc.);
- Using database engines directly or through connectivity object;
- Moving text through networks (including Web and e-mail);
- Last but not least, using command line based scripts or commands.
The level of Unicode support in all these things, and the level of
dependency on code pages, varies a lot.
Some things support Unicode, some don't; some nicely convert to and from
code pages, some are more clumsy; for some things you need to do certain
things to set up Unicode support, for some others you do totally different
things; some things see Unicode as a "16-bit character set", some other
things are more up-to-date with UTF's.
We are talking about Windows NT, but I don't think that the situation is
much different on other systems.
> Well, like I said, the apps that stay Unicode throughout are
> ones that I would consider to be ideal.
That's clear, but we were discussing about round-trip convertibility between
Unicode and another standard, so Unicode-clean application clearly don't
apply here.
> > And truth is never wrong.
> But it also never stays the same.
But also it doesn't change overnight.
<RAMBLING thoughts of dubious value>
I think that local character sets will be with us for years to come.
Potentially, we will have to deal with them forever: if we need to read
Linear B tablets in year 2001, I don't see why our nephews in year 2999
should not need to read old e-texts in Shift-JIS or ISIRI 3342.
Mmm... Perhaps I should amend what I just wrote: almost no one in 2001 has
ever seen a Linear B tablet and, probably, would never have the possibility
of touching one. But everybody can go in a bookshop and buy a *paper* book
*printed* with reproductions of Linear B tablets. This is because the
technology has evolved and we now have paper in place of clay, and printing
in place of manual copying.
Similarly, the new technology must be able to reproduce and bring forward
the documents produced with older technologies.
If it was not for this need, I would be the first one to shout that a big
percentage of Unicode code points is just crap that could be wiped out.
Unluckily we also have to deliver the crap to posterity, just like clay
tablets talk us about slaves and human sacrifices...
</RAMBLING>
_ Marco
This archive was generated by hypermail 2.1.2 : Fri Jul 06 2001 - 00:17:15 EDT