On May 30, 2007, at 3:40 PM, Neil Conway wrote:
> On Wed, 2007-30-05 at 21:23 +0200, Peter Eisentraut wrote:
>> I'm not sure what your rationale was for creating lower-case words
>> instead of upper case, except for it looks nicer. Is there a
>> technical
>> reason?
>
> There's no real technical reason: the standard says upper-case, but
> PG's
> general philosophy of case folding would suggest folding to lower-
> case.
> If we were compliant with the spec's case folding requirements then
> emitting uppercase would be the clear choice, but since we aren't, I
> don't have strong feelings either way.
Sorry for the late reply...
I'm worried that this would make us incompatible with cross-database
code. Granted, should probably be using a boolean representation, but
I'm not sure if that's universally true. And if we find out later
that lower case is a problem, it won't be possible to change it
without messing with the rest of our users. I think it'd be best to
go with the spec.
--
Jim Nasby jim@nasby.net
EnterpriseDB http://enterprisedb.com 512.569.9461 (cell)