Re: TMQL Issue: Functions and Predicates as first-class topics

On Fri, Mar 09, 2007 at 06:18:03AM -0800, Frederic Andres wrote:
> regarding TMQL Issue related to Functions and Predicates,
> wouldn't it be easier  and cleaner to define a many-sorted algebra. 

It would *definitely* be very clean. And also accessible to optimizers.

I only wonder how complex a definition for 'date' would be. date and
datetime are *incredibly* tricky and manyears have been invested
already.

With all these formal things, the first yards are extremely elegant and
effective. But the final mile often adds a prohibitive cost.

\rho