Hello all.
I share the opinion of most (I assume) participiants of this forum
that an OO DB is the future. Although I think that OO DB alone can not
become the one.
I think that it must be integrated with an OO OS.
Only then it may become fast, powerfull and useable.
I'm looking for collaborators to write such OS.
Everybody is wellcome.
One can find more information at
http://www.yaooos.org
Michael.

"Michael Groys" <michaelg (AT) alzt (DOT) tau.ac.il> wrote

Depends on what you mean by OO DB. If you mean a network model dbms with
some user-defined type support, only the ignorant or misinformed will agree
with you. Sadly, they might form the majority.

Hello Bob.

Bob Badour wrote:
Actually I worked only with regular databases.
In addition I wrote ODBC driver for some database project.
By OO DB I mean the database that contains objects of different types
and relations between them. As you know regular relational databases
contain tables with some predefined set of fields per each entry.
Then query is done on some expression which is based on values of these
fields.
In the case of OO DB I think about queries based on expressions over
object methods (which are virtual in the sense of OOP). The most
important feature as I see - is the possibility of handling objects from
different types together.
All this can be done effectively only in the scope of OO OS.

Best regards and don't be sad.
Michael

"Michael Groys" <michaelg (AT) alzt (DOT) tau.ac.il> wrote

You seem to misuse the term relation. A relation is a set of n-ary tuples.
Is this what you mean?

I suspect you really meant to say 'objects of different types with pointers
among them'.


No, I don't know that. A relational database is a set of true fact
statements.

Bob Badour wrote:
May be I indeed misuse it, but what I actually meant is that the OO DB
have to manage some connections/associations/relations between objects
and/or basic types.
How it is organized is not important for the concept.
It can be pointers or n-ary tuples or something else.

When you perform sql query - you access some columns of some tables.
You can perform very complicated queries that will access number of
tables with resolving some n-to-m associations or whatever you want.
You can describe your databases in macro level in any way you want,
but at the lower level it will remain the set of tables with different
columns that can be handled in row wise or column wise fashion.

"Michael Groys" <michaelg (AT) alzt (DOT) tau.ac.il> wrote

As I mentioned previously, a relation is a set of n-ary tuples. Why do you
persist in misusing the term? Connections and associations sounds an awful
lot like pointers to me.


I disagree. The logical data model has the utmost importance.


If it is n-ary tuples, it is a relational dbms. If pointers or
'connections', I suspect you mean a network model or hierarchic dbms.


Are you suggesting that SQL defines the relational model? Ridiculous!

Using a relational dbms, one derives new true fact statements from the
database using the rules of symbolic logic.

Bob Badour wrote:
a set of n-ary tuples! Designers of DB just invented tables of ntuples
and called them - "relational" database.
If somebody has a son then there is parental relation between them
without any "tuples".

with projection operators.
And what if it contains both pointers and ntuples.
People use databases to store information in an efficient way.
They access it by providing some boolean expression to select
tuples with additional expressions to specify what data they
want to retrieve and how it must be organized.
In this way it is done in any query language excluding latex or doc.

"Bob Badour" <bbadour (AT) golden (DOT) net> wrote


Making a literal interpretation, I don't think he is misusing the term
here. What he said is the same as: the database that contains
relations composed by values of different types.

A relation value is a relation (or relationship) among values, and
objects are values.

BTW the common mathematical definition of relation is: a subset of the
product of 2 sets. But of course "2" should be replaced by "n".

Very probable because what he said could be an RDB.


Regards
Alfredo

"Michael Groys" <michaelg (AT) alzt (DOT) tau.ac.il> wrote

With all due respect, you are posting to a newsgroup within the
comp.databases domain (internet usage of 'domain'). In the scope of
databases, 'relation' has a very precise and well-defined definition. If you
wish to communicate effectively in a database newsgroup, I suggest you stick
to that definition.

When you check Webster's, you will discover that aunts and uncles are
relations, and I don't think that definition applies here.

Unless you mean sets of n-ary tuples, I suggest you use another term that
implies exactly what you do mean.


Your ignorance is showing. Dr. Codd, a mathematician and computer scientist
working at IBM, applied the mathematical definition of relation to the
problem of managing databases more than 30 years ago. His work has
established the state of the art and science in data management. Thus far,
no serious contender has emerged, which is not surprising given that Dr.
Codd's idea allows one to apply logic and mathematics directly to the
problem of data management.


Again, your ignorance shows.


Network and hierarchic both use the concept of 'record', which corresponds
to n-ary tuple.


Your ignorance is not only showing but is also profound. Please stop trying
to spread your ignorance; the goal is not worthy.

Let me one more time show my ignorance.
First of all I want to ask you few questions.
1. What is the purpose of database?
2. What part of database users, programers and administrators use
mathematical models in their work (and know about Dr. Codd).
3. For what purpose YOU use database?
what kind of database YOU are using?
How do YOU access it?

Second, I thought that this group deals with object oriented
databases. Tell me if I'm wrong.

Third, I have some comments on my ignorance that you showed.

Bob Badour wrote:
Prof. Turing invented Turing machine and other very nice concepts,
But this doesn't mean that we have to use Turing machine in programming.
In any case I don't want to solve problem of data management,
I just want to provide people that didn't read Dr. Codd's books
with an easy and efficient way of handling data.

I think that the real database must solve the problem of managing of
concrete data.

You didn't tell me anything new except definition of word relation.
In any case I know a lot of people that use and manage databases and
do a good job without need to apply to any mathematical models.

Best Regards, Michael