Re: Onto a potential relational manipulation language

On 12 déc, 00:22, vadim..._at_gmail.com wrote: <<
However, the models in Mace4 are generic, while in database world one would like to check assertions against objects that
are structured as tables (err, relations).
>>

I looked the Mace4 and link you provided and wished it could help me but it does not. The computing model I implement does not operate relations as tables (instead, one could see a table as state of a relation in a specific point in time).

A consequence of the above is that assertions do not tuple based assertions to implement relational operators.

To clarify the above explanantion, here is an example.

Suppose the 2 relations (apologies for the notation differences)

COMPUTER = [{HP1, HEWLETT PACKARD}, {HP2, HEWLETT PACKARD}] MAKER = [{HEWLETT PACKARD}] To implement the JOIN operator between COMPUTER and MAKER, the model I am exploring does not operate at any time the assertion between the tuples in MAKER and tuples in CONPUTER but implements an interval based INTERSECT between the domain of values of MAKER and the domain of values of COMPUTER reference to MAKER.

In that sense, the assertion in the traditional RA sense becomes a moot concept. I believe that you have been misled by the syntaxt I used in the language I am trying to build in the sense that you equate it with operation it does.

R3 = [R1 + R2] is not an assertion but the allocation of the result of relation R1 representation UNION the result of relation R2 into a new relation R3. The relations R1, R2, R3 become R-Tables only at runtime.

Hope this clarifies. Received on Fri Dec 12 2008 - 13:48:18 CET