On Fri, Dec 08, 2006 at 04:50:16PM +0100, Alban Hertroys wrote: > Why not use unicode symbols 0x2640 and 0x2642? A clever idea, however, it does not cover transsexual, female phenotype transsexual, male phenotype hermaphrodite, intersexual phenotype which we (GNUmed, that is) need to support in a medical database. In fact, most (all?) of the approaches I have seen in this thread lack that. Using bool would even preclude extension of the constraint in that direction if it were ever needed. So, one is better of with, say, char(2) or something similar. I would also suggest using a *coded* gender, not "male", "female" strings which will make gender-based calculations a lot easier down the road. Karsten -- GPG key ID E4071346 @ wwwkeys.pgp.net E167 67FD A291 2BEA 73BD 4537 78B9 A9F9 E407 1346
