Re: Possible bridges between OO programming proponents and relational model


Bob Badour wrote:

// Anything one wants that represents the SQL table. It might depend on
the
indexes used. It might depend on how one distributes the table. It
might
never get represented exactly as a table at all.//
You are wrong. Indexes represent disk organisation and what allow
consitution of execution showplan and calculation to isolate data in
RAM. IO representation is a result of loading data pages(index + data)
on RAM and executing specific showplan on theses indexes as well.
Results of execution traduced as SQL tables are stored according to
simple RowAddress/ColumnAdress bidimensional scheme. How in the hell at
that moment (just before display) can they be anything else than
bidimensional according to their according coordinates on
RowAddress/ColumnAddress.

// What in-memory representation do you want to discuss? A row store? A

column store? Some other representation? Others have offered you
alternative vocabulary for what they think you might mean, but you
totally ignored it every time.//
I did not this terminology for the simple reason that I don't know what
it refers to precisely. I will check it to see if it corresponds to
what I am refering to. I missed this comment.

//> > //Another completely incoherent paragraph.//
This is getting repetitive but I should clarify that...Do you deny that
the body of a relvar represented after a select x, y, z is ran is
anything else similar to

X-AXIS----------------
Y-AXIS 1,2,3
- 2,4,4
- 4,4,4
- 5,5,8

Do you deny that X and Y represent mathematical dimensions?

But the dimensions of what? A piece of paper? A computer monitor? That
is the point that seems to elude you. Neither X nor Y are properties of
the relvar or of the implementation of the relvar.
//
The media has nothing to do with physical addressing scheme. A
dimension according to the physical adressing scheme used in memory to
represent the result of the query. I perfectly know that this precise
X and Y have nothing to do with the dimension of a relvar which is at
logical level definition of RM. You keep switching back to logical
layer while I told a hundred times that there is no point discussing
logical level. Why do you keep putting in doubt logical definition of
relvars.

// Alfredo has shown no indication whatsoever of confusing anything. If

your only goal is to insult the people who are trying to help you//
This is BS. Prove it. I have not insulted Alfredo or anybody on this
board. If you feel insulted because I pointed out that some of your
comments were totally irrelevant or wrong because they were is not an
insult.

//, you
can go *** yourself.//
Bob you are going too far. If this is the way you *help* people on
boards? This kind of declaration are totally contradicting what you
present as noble intentions and supposed to be meaningful. I did
tolerate until now your confusions and I will not tolerate to be
insulted and simply ignore this kind of post. You should be ashamed of
yourself.

Cimode wrote:

Alfredo Novoa,
Cimode ha escrito:

<< This does not make any sense, the dimension of a relvar is the
number
of attributes.>>Absolutely. but SQL table implementations are not
relvar, just a possible representation of a relvar.

This does not make any sense.

It does not make any sense because you are confusing R-Table/SQL table
and their punctual in-memory representation at runtime.

//A SQL table with three attributes is always three dimensional
independently of how it is implemented.//

At logical level yes but not at physical level.

At any level, Cimode. If you don't mean to refer to the SQL table, then
refer to something else that you do mean. Any triplet, though,
represents three dimensions.

Are you referring to a two-dimensional array representing the table? Are
you referring to something else?


You advance a strong argument without anything to back it up Prove it.
What is the in-memory representation of a SQL table at run time?

Anything one wants that represents the SQL table. It might depend on the
indexes used. It might depend on how one distributes the table. It might
never get represented exactly as a table at all.


[snip]


Several people have already understood the issue
here but you are not one of them.

You advance a strong argument without anything to back it up. Prove it.


//<<I suppose that you know how to represent a cube with a table.>> It
is
an analogy used for communication's sake. You misread and
misunderstood my comment. I said that a face of a cube is
bidimensional and is a comparable to what a SQL Table is to a relvar.

And any representation of a face of a cube is bidimensional. However,
any representation of the cube is three dimensional, or it no longer
represents a cube. Don't confuse the projection of a cube with a cube.


//Your comparation is completely nonsensical.// This is getting
boring.

Are you suggesting that the memory is less physical than the

disk?>>Interesting question but the answer is obviously no.

Indeed but your previous post seems to suggest the contrary.

The
argument here is about getting better independence *progressively*
through thinking as to how in-memory representation of relvar could be
more truthful to what a relvar .

You have already stated that. However, you then immediately confused the
logical and the physical again.

Any in-memory representation of a relvar that represents a relvar is
truthful to what a relvar is. Otherwise, it doesn't represent a relvar.
Similarly, any in-memory representation of an SQL table that represents
an SQL table is truthful to what an SQL table is. Otherwise, it doesn't
represent an SQL table.

What in-memory representation do you want to discuss? A row store? A
column store? Some other representation? Others have offered you
alternative vocabulary for what they think you might mean, but you
totally ignored it every time.


//Another completely incoherent paragraph.//
This is getting repetitive but I should clarify that...Do you deny that
the body of a relvar represented after a select x, y, z is ran is
anything else similar to

X-AXIS----------------
Y-AXIS 1,2,3
- 2,4,4
- 4,4,4
- 5,5,8

Do you deny that X and Y represent mathematical dimensions?

But the dimensions of what? A piece of paper? A computer monitor? That
is the point that seems to elude you. Neither X nor Y are properties of
the relvar or of the implementation of the relvar.

The body of the relvar remains unaltered by the select as does the
implementation of the relvar.


Do you support that the above representation is the relvar itself.

No, it isn't the relvar itself. It is a representation of a relation on
my computer monitor (when I look at it.) X and Y are dimensions of my
computer monitor.


If
you answered yes to any of the above questions then you are wrong dead
wrong.

// The answer is evident to me: OO has nothing to offer. Both things
are
completely unrelated.//
Prove it. If your knowledge of OO mechanisms is as effective as the
confusions you are making between logical and physical layers,

Alfredo has shown no indication whatsoever of confusing anything. If
your only goal is to insult the people who are trying to help you, you
can go *** yourself.

.