On Mon, 19 Jul 2004 11:13:53 -0700, John Jacob wrote:
>
> In fact, *any* formal system at least as powerful as typographical
> number theory (which a relational language certainly is, even without
> relation-valued attributes) is subject to the Godelian attack, it says
> *nothing whatsoever* about whether or not the system is useful.
Not so fast. The relational algebra corresponds to first-order logic
without function symbols, which is a strict subset of FOL with function
symbols, and for that Godel is actually our friend because he has shown
that we have a complete axiomatization for that, known as Godels
completeness theorem. Which, by the way, doesn't mean things get suddenly
easy because FOL, even without function symbols, is still undecidable.
Received on Mon Jul 19 2004 - 17:56:36 CDT
