| > I think that ISO-8859-15 would be more appropriate than -1, it fixes
| > the bugs in -1 and adds the euro.
|
| For better or worse, I believe 8859-1 is more common and should be
| required, but I don't have a string opinion on the subject.
It is the case today but it is unlikely to be the case when sieve
become a standard. The most important reason is the replacement of the
generic money symbol (0xa4 ¤) with the euro symbol but it also includes
some of the characters that they missed in Latin1.
Expect the impact of 8859-15 (so-called Latin 0) to be negligible. The only
vendors who are supporting it are those who have weak Unicode implementations.
Had anyone asked me to predict the impact of iso-8859-15 in advance this is
certainly what I would have said. However, this is not what I'm observing in
practice -- we've had a number of customers ask for 8bit charsets that contain
the Euro and iso-8859-15 support in particular. (And this despite the fact
we've had reasonably complete Unicode support in place for years that nobody
seems terribly interested in.)
The dominent desktop is Microsoft Windows, and it supports the Euro just fine
through Unicode; I have yet to see any application on MS Windows that supports
8859-15.
This may be true but I think it misses the point. Having good application
support for iso-8859-15 isn't really necessary; all you have to have something
that supports an 8bit single byte charset and then all you have to do is use
the right character tables. By contrast, getting full support for Unicode in
place is vastly more difficult.
However, I'm far from convinced this will result in wholesale replacement of
iso-8859-1 with iso-8859-15. At most I'd be willing to add iso-8859-15 to the
list of must-decode charsets, but that's as far as I'd be willing to go.
Ned