Re: citations of nature
Date: Sun, 04 Jan 2004 06:46:50 GMT
Message-ID: <uFOJb.52423$I07.174447_at_attbi_s53>
"Dawn M. Wolthuis" <dwolt_at_tincat-group.com> wrote in message news:bt7n7f$iq5$1_at_news.netins.net...
>
> The definition I'm currently using is:
> Database: Retrievable data encoded on a persistent storage device combined
> with metadata - information about that data.
I think that too much thinking these days goes into the "persistence" part. Many if not most application programmers think of an DBMS as something that provides persistence primarily and little else. In fact, I consider persistence to be an *optional* feature of a DBMS. That is, the things a DBMS provides are all still useful even in the absence of persistence. Which is not to say that persistence isn't useful; it is. But simple persistence is actually not an achievement of much note. I think the applications programmer's emphasis on DBMS persistence is a kind of damning with faint praise.
I have to say I'm enamoured of Date's (or is it Pascal's?) "structure, integrity, manipulation." (Note that persistence is not mentioned.) I think those three things capture what a DBMS does pretty well.
Marshall Received on Sun Jan 04 2004 - 07:46:50 CET