From: "Jianping Yang" <[EMAIL PROTECTED]>
> "Carl W. Brown" wrote:
> > If there are no surrogates in the database, is there any reason that I
can
> > not change the database from UTF8 to AL32UTF8?
>
> You can change the database from UTF8 to AL32UTF8 in this case. Also you
can
> use Oracle database scanner to scan your UTF8 database and if you find
there
> is any pair surrogates inside, you can convert UTF8 into AL32UTF8 through
> imp/exp.
This really does weaken the whole argument of the naming here. There are two
possibilities:
1) Oracle has scores of customers who are using supplementary characters in
their prior version products using the irregular UTF-8 byte sequences.
2) Oracle does not have such customers but they figure they will, now that
there are assigned supplementary characters.
Now, you have previously said that (1) was not the case. You have also
explained that (2) is quite true and will be a huge problem.
Therefore, there is no problem in your NEW product with having compliant
UTF-8, named UTF-8, and calling it a day. The strange new name can go to
your non-compliant UTF-8 and stay as internal as you like it!
I think that Oracle is really going to need to explain WHY this is a problem
given that they (you) are claiming that #1 is not true. Oracle has a chance
to prove that they are willing to do some work here to avoid what many
consider a major compatibility issue that will in fact weaken the standard.
MichKa
Michael Kaplan
Trigeminal Software, Inc.
http://www.trigeminal.com/
