On 12/12/06, Johnny Billquist <bqt@softjar.se> wrote:
> Unless you're of the opinion that Unicode itself is a mess that we would
> be better without.
> But by now that's hardly a choice anymore...
>
>         Johnny
>
> Gilbert Fernandes skrev:
> > On Tue, Dec 12, 2006 at 12:49:55PM +0100, Christian Biere wrote:
> >
> >> UTF-8 uses 8-bit bytes. MS claims or claimed that UTF-16 is the best
> >> performance/memory trade-off.
> >
> > What matters most is not performance nor performance/memory tradeof,
> > but more compatiblity, especially because UTF is encoding of strings..
> >
> > It's for a very very good reason (and a clever one) that Ken Thompson
> > and Rob Pike did an UTF-8 out of UTF : the encoding of byte codes
> > for UTF-8 is the same as for ASCII for low-bit values. This makes
> > UTF-8 very interesting when UTF-8 documents are displayed on ASCII
> > using systems : there is little or no change to do to have a proper
> > display of UTF-8 there.
> >
> > UTF-8 seems a better tradeof.
> >


Here's a netbsd amd64/openldap bug:
http://www.openldap.org/its/index.cgi/Incoming?id=4675#themesg

I couldn't duplicate this with netbsd-sparc64.
I've also heard there are some driver problems/incompatibility on
amd64 in 64-bit mode, but that's all hearsay.