On 11/19/2010 1:22 AM, Kjell Rilbe wrote:

> Maybe it would be possible to use a compatibility setting, along the
> lines of SQL dialect 1 and 3. That must have induced a lot of changes in
> a lot of tools, but "they made it". Same goes for identifier length, and
> would probably induce a lot less changes in the tools.
>
> I'm not sayin there should be a dialect 4 with longer identifiers, but
> perhaps a new DB-level setting?
>

No, no, no. Multiple SQL dialects was a disaster that was countenanced
only because there was an incompatible change to a datatype that the
developers couldn't find a way around. It was (and still is) awful.

The problem with extending the length of identifiers is just that it's
a lot of work. Existing tools can still work unchanged on databases
that still use short names - the API separates the field definition from
buffer definitions - so the transition does not have to happen
simultaneously.

No more modes!

Best regards,


Ann