Re: Objects and Relations
From: Bob Badour <bbadour_at_pei.sympatico.ca>
Date: Mon, 26 Feb 2007 14:59:59 GMT
Message-ID: <PBCEh.1520$PV3.21889_at_ursa-nb00s0.nbnet.nb.ca>
>
> In RM domain is used as a synonym of type and it is completely
> different to relation.
>
> When you say "relation" we assume that you mean "relation value". If
> you mean "relation variable" please write: "relation variable" or
> "relvar".
>
> On the other hand, a domain is a set of any kind of values: scalar,
> relational, tuple, array, etc.
>
>
>
> I don't see any sense here.
Date: Mon, 26 Feb 2007 14:59:59 GMT
Message-ID: <PBCEh.1520$PV3.21889_at_ursa-nb00s0.nbnet.nb.ca>
Alfredo Novoa wrote:
> On 24 feb, 13:43, "Cimode" <cim..._at_hotmail.com> wrote:
>
>>If you define class by domain then what is the RM equivalent for
>>relation?
>
> You probably mean the OO equivalent for relation.
>
> OO lacks an equivalent for relation. That's why it is so difficult to
> write database applications using OO languages.
>
>
>>In RM, a domain is *not* the same* as a type or a relation.
>
> In RM domain is used as a synonym of type and it is completely
> different to relation.
The above is arguable and depends on one's definitions for domain and type. Specialization by constraint makes them essentially synonymous.
>>( >>domain = set of *possible* values for which a relation draws >>*possibly* values
>
> When you say "relation" we assume that you mean "relation value". If
> you mean "relation variable" please write: "relation variable" or
> "relvar".
>
> On the other hand, a domain is a set of any kind of values: scalar,
> relational, tuple, array, etc.
>
>
>>typing = making a domain from a relation body and name (therefore >>only *existing* values for a specific relation).
>
> I don't see any sense here.
As usual, Cimod uses a definition derived solely from his own imagination. Received on Mon Feb 26 2007 - 15:59:59 CET