Re: pro- foreign key propaganda?
Date: Tue, 20 May 2008 22:25:18 GMT
Message-ID: <idIYj.152474$rd2.41414_at_pd7urf3no>
David Cressey wrote:
...
> Your guess is as good as mine, maybe better. But my guess is different.
>
> I believe the word "key" was as unhooked from "address" as any term of the
> day.
> I don't know a thing about IMS. For several indexed file systems, the keys
> were fraught with navigational consequences, but this was *not* because
> they were something other than data. The keys were just as much data as
> they were in the 1970 paper.
>
> In fact, when I was easing my way towards RDBMSes, I spent some time using
> indexed files, keys, foreign keys, and relational joins without the
> benefit of any DBMS. It's remarkable how much of "relational think" can be
> transposed to index files with little, if any, loss of conceptual clarity.
Yeah, I remember in the 1980's Unix had a file-based 'join' command. Not to criticize anybody, such as the physicists you mentioned - I have some sympathy for 'outsiders' who refuse to accept IT dogma and cause a course change in the computer field. But looking back on the data field's history, it is really just a series of stumbles and meanders, once in a while the latest direction looks sensible but is soon usurped with more chaos.
(Codd was a pragmatist for sure - twenty years after his first paper, he was making consulting bucks that would embarrass even the greediest senior law partners but was still willing to become architect of an obscure, non-relational product. Money wasn't at issue, just the choice of job titles between a couple of large egos put the kibosh on that deal.) Received on Wed May 21 2008 - 00:25:18 CEST