Re: Possible bridges between OO programming proponents and relational model

From: vc <boston103_at_hotmail.com>
Date: 3 Jun 2006 15:17:21 -0700
Message-ID: <1149373041.672770.194980_at_i40g2000cwc.googlegroups.com>


Cimode wrote:
> << Hold on right there. No one is talking about implementation, much
> less about 'direct image implementation', whatever it is. Please
> state clearly what the difference is between two expressions
> describing
> the variable that can contain a three dimensional object (except minor
> syntactical pecularities).>>
>
> I have stated as several times in the initial question (feel free to
> reread) that the main purpose of the thread would be exchange on that
> matter. If you believe this subject is uninteresting then feel free
> not to participate.

No, that is not an answer. It's called a non sequitur. The question was not about "the main purpose of the thread". You've stated earlier that there is some difference between the relvar expression describing a three-dimentional object and the SQL expression doing the same. I asked a specific question: what is the difference between those two expressions except minor syntactical differences? Are you unable to answer the question clearly ? Do you refuse to do that ?

Relation 'flatness':
Are you saying that the mathematical relation is somehow 'flat' and due to the purported 'flatness' is unable to express anything about, say, geometrical objects like a sphere or a cube ? Since you've stated earlier that the database relation (which is an implementation of the respective mathematical relation with some unimportant pecularities) is somehow 'flat', then you must admit that the modern math is unable to cope with three dimensional geometrical objects. Do you still insist on the database relation 'flatness' ? yes or no ? why ?

>If you do not know what a direct image
> implementation is I suggest you do some reading on transrelational
> model by Steve TARIN. As my time is as limited as yours, I suggest you
> do some research on that before we can exchange in a fruitful manner.
> Check the following link.
>
> http://en.wikipedia.org/wiki/TransRelational_model

Your suggestion does not carry much force because a) you appear to be unable/unwilling to answer truly trivial questions. Your refusal/inability to do so does not inspire much confidence in your ability to discuss more complicated matters than the school algebra notions. b) since you apparently advocate a novel data management approach, you should be able to state briefly what advantages the new approach has in comparison to the relational model and whether those advantages are sufficient to make people convert to the new way of doing thing. Given the level of the RM understanding you've exhibited so far, the likelihood of your being able to present such comparison in a meaningful way is very low.

So, for now let's postpone discussing the fancy stuff like the 'trans-relational model' and try to settle the issue of whether the database relation is flat or not. OK ?

>
> <<What's that supposed to mean ?>>I am suggesting that a physical
> adressing scheme based only on a vertical and horizontal dimensions
> (not relation dimensions but mathematical dimension) can only allow
> only a bidimensional representation of a relvar at runtime.
>
> <<That does not make any obvious sense>>I perfectly know this is not
> *obvious* which is the reason I am opening a thread to exchange with
> the nice fellows here. But not obvious does not make it nonsensical.
>
> Thank your for your reply.
Received on Sun Jun 04 2006 - 00:17:21 CEST

Original text of this message