Re: Nested structures

JOG wrote:
> dawn wrote:

>> Some, but not all, of you will find this new IDC paper entitled
>> "Because Not All Data is Flat: IBM's U2 Extended Relational DBMSs" to
>> be of interest.
>> [snip]

Like the earth?

>
> All information we communicate to each other may be stated as a
> proposition in first order logic. Everything.
>

Like: "When there is distress of nations, on the shores of Asia or in the Edgware Road, men's curiousity searches past and present, and clings to that dimension. But to apprehend the point of intersection is an occupation ..." (TS Eliot, apologies if I didn't quote it exactly.) Surely we are talking here about limited communication, ie., the kinds we find useful to automate with a db, not poetry for example.

> Hence, any collection of information may be organized into sets of
> propositions, as determined by commonalities in the structure of those
> propositions. The information world could be no flatter.
>
> J.
>

Actually, I've been a little distressed lately when thoughtful types like JOG and David C talk about flatness or star models. Maybe I'm just not catching some in-joke or other. Somebody else, sorry, I forget who, mentioned that some of these non-relational techniques can be more efficient/elegant from the programmer's point of view for certain limited applications whereas a pure RM implementation for such apps might be ponderous in terms of its code.

I don't argue against this but I also don't think this is an argument against strict RM, it just points out to me that we do not have agreed modelling methods that will express the effects of choosing those adhoc impl'ns in terms of classical logic and set theory. Surely predicate logic is the essential tool we have when we want to talk in unequivocal terms as well as when we want to predict results in common terms.

I haven't put that very clearly, but I'll let it ride along with saying that I think the discussion is comparing apples to oranges, ie., falling into the constant trap of confusing logical with physical. I think that unless you have some formal basis for describing a db, you never really know exactly what you've got nor all the manipulations you can achieve in a consistent way.

Just my rant for today.

p Received on Sat Sep 16 2006 - 16:43:16 CEST