Re: Example of expression bias?

Bob Badour wrote:
> > Whilst it's difficult to be absolutely certain, I think 'no' is a
> > reasonably safe answer. I'd be very happy to be proved wrong in this.
>
> Am I totally way off base or would prolog approximate one?
>

Not really; Prolog is a combination of a theorem prover, a database (of sorts) and a unification algorithm. The lambda calculus doesn't really enter the picture. (First order logic, particularly Horn clauses, do though).

Prolog also has a couple of very non-relational features; the order facts & rules are asserted to the database affects the order of evaluation and the order in which results are returned directly, and attributes are accessed by position rather than by name. Prolog is also effectively untyped. Some of the operators are strict on their types, but that's about it. Asserting rules in the "wrong" order can lead to a non-terminating program.

That said, a couple of years ago I put together a relation calculator that implemented a fairly butt-ugly version of the relational algebra and allowed you to add "types" and operators to it. Dirt slow, and I keep meaning to redo it from the ground up to do it properly, but most of the work was done in a couple of weeks. It's a great language for building quickly in, as you concentrate on *what* you want to do, and leave the computer to the grubby *how* of making it happen.

And, any language where you can give an answer and let it work out the question is fine by me :) Received on Wed Jun 21 2006 - 02:18:38 CEST