Re: computational model of transactions

"paul c" <toledobythesea_at_oohay.ac> wrote in message news:goQBg.331006$Mn5.38111_at_pd7tw3no...
> Brian Selzer wrote:
>> "paul c" <toledobythesea_at_oohay.ac> wrote in message ...
>> You're right about "sum" being disjunctive. I don't know what I was
>> thinking. But I don't get what you're trying to say about a minimal
>> logic. Should set theory be stripped away? I guess you could do that if
>> you think of a database as one big statement. Even if you did that, I
>> still don't understand how you can avoid order unless the database is
>> static. If the database can change, then the operation that changes it
>> creates a new database. I suppose you could ignore the old database but
>> that doesn't mean that it doesn't exist.
>
> As others here state, set theory is already pretty minimal but complete
> enough at the same time, so what would be left? Sets have these operators
> that are almost direct parallels to inferences in predicate logic which
> makes a pretty powerful and flexible combination - I'm not smart enough to
> think of something better, but I suspect the nuances have not yet been
> fully explored, eg. logical equivalence and rva's, not to mention type
> theory which I'm mostly ignorant about.
>
> The vocabulary you use is full of words that suggest time, eg., order,
> static, change, new. The other problem I see is that you consider the
> database part of your universe. I think to make any progress you must
> separate the system from the environment - in my opinion, most, maybe all
> systems don't go far enough in this regard, which is why they encumber the
> user with their own 'environments' and users continue to think in ways
> that are too dependent on implementations. However, following anything I
> say is probably not good for your career.
>

I've thought about this, and I still don't get it. Given a proposition S, and a second proposition A, then "S or A" is a distinct proposition: let's call it T. If S is a database state, then T is also a database state. T is the state that includes A, and S is the state without. Which do you query? How do you know?

I don't see anything wrong with including the database as part of the universe. A statement about a thing is a thing, and as such can be part of the universe. I think it's important to be able to discuss what is known, and especially in a temporal database, what was known at a specified point or interval in time in the past. It should be possible to draw conclusions about what is known and then for those conclusions to become part of the database.

I agree that people can get stuck on the mechanics of one implementation and amidst the daily hubub of deadlines and cross pointy-haired people can lose sight of the foundations, but I also think that a solid understanding of at least one implementation is essential, and more is better.

> p
Received on Tue Aug 08 2006 - 05:51:42 CEST