Re: Nulls, integrity, the closed world assumption and events

dawn wrote:
> [snip]
> I have mentioned before, but perhaps not recently, that other than
> performance issues, I have little interest in what happens under the
> covers, outside of the interfaces between humans, clients and the DBMS.
> So, if it is best to put data in 1NF behind the scenes, so be it. I
> am not convinced that is best, but nor is it a question that concerns
> me.

If it eliminates query bias and allows more powerful manipulation of propositions then it is a benefit. If it doesn't affect your particular practices then no problem - but better to have that capability there for those who do use it (and hell do I want to use it).

Look, a 'proposition' can have multiple values for the same predicate role (but when they are these are conjunctions - and I find this much more appealing than a fudging some set based value artifice to represent it). Its just that base relations have more _manipulative power_ with single values. But importantly this does not preclude having an interface that suits you in terms of a virtual relation (and that a coder doesn't have to generate every single time via an OR mapper). So as I said bath-water and babies - I am adamant that you do not need to throw out anything to get the interface you want.

> Developer data modeling is related to the interface between the
> DBMS and the developer (for all functions a developer might do,
> including debugging). That is where I would like to see better
> optimization.
>
> > I see no
> > reason why there cannot be a solution to this that accommodates both
> > standpoints.
>
> I agree. What do you think of products such as Intersystems Cache'?
> What would you suspect (or know) might be missing for such a product to
> meet your needs? --dawn
Received on Mon Jan 22 2007 - 17:06:33 CET