FreeTDS Development Group

[Steve Langasek <vorlon AT netexpress.net>]

On Thu, 7 Jun 2001, James K. Lowden wrote:

> > UTF8 doesn't help CJK languages, of course.  Does TDS7.0 even support 
> > those
> > languages, though?

> Rather than tackling unicode -> ascii conversion (which will break down as
> soon as you run out of bits), I *think* Steve's suggestion would be the most
> direct path to unmangled Unicode transportation.  A clean 16-bit pipe, IOW.
> From what their documentation says, that would make PHP happy, too.

If it doesn't make PHP happy, then it should at least be fairly
straightforward to kick PHP into shape.  That's the Right Thing To Do, in any
case.

> Steve, what would it mean for FreeTDS to "support CJK languages"?  Isn't it
> enough to ferry the data back and forth?

The problem is that the 16-bit Unicode character space isn't large enough to
cover all of the oriental character sets.  If TDS 7.0 is native-Unicode
(where 'Unicode' refers here specifically to the Microsoft implementation,
i.e., the 16-bit space), then I don't imagine it's possible to represent some
of the Chinese character sets as text streams without some mutually
agreed-upon hackery.  If that's the case, it really has no bearing on a
decision to use Unicode vs. 8-bit charsets.

Steve Langasek
postmodern programmer