Re: Declarative constraints in practical terms
From: Marshall Spight <marshall.spight_at_gmail.com>
Date: 26 Feb 2006 13:05:57 -0800
Message-ID: <1140987957.052031.130800_at_u72g2000cwu.googlegroups.com>
> A little while ago, Kenneth Downs described an engine he was working on,
> called something like Andromeda. From his description, it seemed like the
> engine was sort of a supercharged data dictionary. And it could go back and
> forth between a declarative form and an imperative form for the same piece
> of data definition. Maybe I'm remembering this wrong.
Date: 26 Feb 2006 13:05:57 -0800
Message-ID: <1140987957.052031.130800_at_u72g2000cwu.googlegroups.com>
David Cressey wrote:
>
> A little while ago, Kenneth Downs described an engine he was working on,
> called something like Andromeda. From his description, it seemed like the
> engine was sort of a supercharged data dictionary. And it could go back and
> forth between a declarative form and an imperative form for the same piece
> of data definition. Maybe I'm remembering this wrong.
Yeah, I remember looking at that. It seemed pretty interesting.
> It seems to me that "CREATE TABLE ..." is imperative. When the imperative
> is carried out, two things happen. First the table is created. Second, the
> table's description is included in the metadata.
I'll buy that.
The big imperative constructs are insert, update, delete. (More traditionally, we would just say assignment is imperative.) One can consider CREATE TABLE to be an insert into a system table, although I know some people really don't like that idea. (Were you one of them? I vaguely remember.) Anyway, even if that idea isn't a practical one, it's useful as an illustration.
Marshall Received on Sun Feb 26 2006 - 22:05:57 CET