On 01/19/2017 02:21 PM, Anthony Youngman wrote:
imnsho, "efficient" and "first normal form" are provably mathematically incompatible.
So? Codd made that clear back in the 1980s. That wasn't the argument, then. It was that the IT department took months to process even the simplest user requests. The possible relationships were all hard coded and THAT WAS THAT! The point, the whole point, was MAINTAINABILITY and FLEXIBILITY compared to the extant database forms where the structure and possible relationships were compiled in to the design of the database and only permitted the "questions" hat were part of that design. Since then, we've had many accommodations to achieve performance but many others are actually making the database 'fixed' by other means and adding new tables or restructuring breaks that and requires a 'recompilation. So its not really what Codd envisioned. An example of this is the (otherwise very *NIX friendly) Progress 5GL database system. In development mode it is a pure relational database, but once developed that database and the application are "compiled" and fixed. If you want efficiency you have to go further and that can't be done on trivial databases and simplistic relationships. once you get a few hundred entities and many thousands of basic relationships you can try for third or perhaps fifth normal form. BTDT, and it no easy matter but the factorization is no joke either and I've had over 1,00% speed up over 1st normal when dealing with some missives that way. It hurts the brain and you need some tools and it is an iterative and its not applicable for many cases. Well, OK, its only applicable in a few cases. I've never seen a web application what would warrant it -- A: Yes. > Q: Are you sure? >> A: Because it reverses the logical flow of conversation. >>> Q: Why is top posting frowned upon? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org