Re: Modernization and multi-member files

On Tue, 2008-07-15 at 13:12 -0700, Dave Odom wrote:

JK, et al,

Slow day? Rarely. Can I offer anything useful on how to go from muti-member files to third or more normal form, Relational architecture? ... Oh, yeah, but I'm not sure you'll like the answer because there is no silver bullet for good Relational design.

As to do I work for MickSoft?... Had the person's asking that question actually read, internalized and understood what I've been saying on here for years, they'd know better than to say that. Since I had the pleasure of working with the folks from the Santa Teresa Lab, I kinda know who invented the real DB2s and spent some time talking with Chris Date. Chris was very helpful in putting Ted Codds proofs into something that most of us could understand. This back when MickySoft was trying to figure out what a world class OS is; still are.

A beginning suggestion for moving from multi-member flat files to relational tables would be to read Chris Dates books and understand them.

Thank you, Dave, for the reminder. I'm back from my locker with "An
Introduction ..." in my hot little hand. I'm still working on the
understanding bit.

Also, some of IBMs books on the subject, especially the books talking about DB2 under MVS, VM and VSE. And, I suggest you become a member of DAMA, IDUG and other DB2-oriented local user groups. Those folks are a great source.

Both DAMA and IDUG have affiliates local to me, but I have yet to find a
compelling need to join either. Please feel free to contact me
off-list, or on-list if you think there is enough interest.

Of course, you can ask folks on here that have a background in relational database design, preferably those that did not try to learn in the IBM i world as their experiences is often limited.

Well, yes. But of what one platform would that *not* be true?

Hell, you could even ask me, but I doubt you'd want to. Will I know all the answers; NO. Can I help find the answer; YES. Will it be something you'll like or can work on the IBM i; one can only hope.

What part of Date's guidance can you not follow on the i? I admit, of
course, that the i encourages SQL, and SQL is "very far from being the
perfect relational language". And the i gives you a host of
non-relational tools, which you may choose to use--for reasons good or
otherwise. Still ... ?

Sincerely,

David Odom