Re: I think that relational DBs are dead. See link to my article inside

Hi Cimode

> > > How are they attached?
> > Via configuration DB. It is like a RDB, it contains tables, rows, ...
> > but it is not RDB. It is RDB emulation on OODB. So tables, rows and
> > columns has different behavior. For example row can be contained in
> > many tables at one time. So if you change row field via one table it
> > automatically changed in all tables which contains this row.
> You lost me. On one side, you told me there's no concept of table but
> now you use table concepts...Please clarify...

Yea, there are no tables as concept and at the same time there are "tables".

On the most low logic level of DB the network of nodes or or-graph is exists. There are no tables, rows, objects, ... only nodes and links. All links are single directional. Each link has a "color" or identifier. Each node can have as many links as needed. But each node can have only one link with one color. There can't be two or more links with the same identifier directed from one node. There can be many links with the same color directed to the one node. There can be many links with different identifier directed from one node to another. From some node you can found other nodes only when you know identifier for link to these nodes. Each node knows about and owns links dercted from this node. Node doesn't know about links directed from another nodes.

Each node can have a .NET object instance attached. In the database the root node is exists. the root is the beginning of the database. Each persistent object can discover what node is attached to and root node.

links are just pointers. and all works very fast.

Thats all about Cerebrum.Runtime.dll

Also Cerebrum.Integrator.dll has some meta information implemented. Each node(object) at the same time can be considered as collection of related objects. If we take two nodes and call one 'columns collection' then call another 'rows collection' as result we get a 'table' of rows. Table is a node related to two another nodes - columns and rows. Each row is a node too. And as node each row has links to some related nodes. Columns collection contans links to some another nodes. They are AttributeDescriptors. Attribute Descriptor knows about identifier (color) of corresponding attribute. So when we have a rows collection and columns collection we can navigate to row attribute instance. It is a logical model of 'table'. Also there table of tables is exists.

> I was refering to principle as what set of concepts permit to guarantee
> that the system is for instance less hardware dependent..

I think it is absolutelly hardware independent - on the logical abstraction level. As implementation it depends from Win32 and .NET . the same DB storage file works ok with .NET 1.1 or .NET 2.0 compiled. So DB storage file format is independent from .NET version or from user DLLs version where methods of user objects is implemented. It is because of the low logic level as single direction graph.

> > > Don't this approach increase resource consumption at linkage editing
> > > time. (compile/link/run?)
> >
> > may be, may be not. it is relative to point of view. i afraid about run
> > time more then about devtime.
> Maybe/maybe not? Does not this question require more particular
> attention if you think that a system would be more performant.... In
> most computing cycles the primary ressource consumer is not run time
> (execution time) but more compile and link edit time.

Hm, i think that a run time and development time (not compile and link) are matter.

> If your primary worry is devtime, it is an additional reason to spend
> it carefully by using sound logical principles on which to build on.

yes.

> Do not base abstract reasonning on current implementations
> technologies(DOT NET).

It is not logical restriction. It is by design. Even more, the Cerebrum Kernel is written on C(not C++).
.NET and kernel glue is written on MC++ If i need i can port kernel to some another plaform.

> By their semantics and current capabilities,
> they should not guide your reasonning. The opposite should happen.

yes. i understand.

> I have explained to you that metadata in RM would be treated at compile
> time only.
> run time is decomposed into compile/link/execute. The method you
> suggest requires definitions to be executed no matter what. In RM,
> definitions, value and operator constraints are implemented at compile
> time only.

I think that RM is theoretical concept and can't have a compile time, is it?
I am about real compile time in real PC.

>There are unfortunately no RDBMS existing today

Too bad for RM ))

> > My DB also allows use many abstract models. For example, i implemented
> > simple RDB like database to store configuration metadata.
> I believe you implemented a SQL Table like DB.

Sorry for bad english. I want to say simple RDB-like database. A database which similar but not equal to RDB. I talk about this earlier when describe how table is implemented as sub-graph

> > No it is not needed load all network into RAM. You just need to load
> > one instance which holds count attribute. that is all.
> So you store the count value? .

Yes. Collection stores value inside own instance. It is one of optimization. The most optimization i done - O(1) when searching object instance by it ID.

> Again there's no such thing as an RDBMS already implemented.

Ok then when i say 'RDBMS' - read 'table-based DBMS'. For me while it don't exists yet it is no matter if MS SQL truely RDBMS or not. It just called RDBMS.

WBR,
Dmitry Received on Wed Jul 05 2006 - 18:30:33 CEST