Re: An object-oriented network DBMS from relational DBMS point of view

On Mar 9, 3:44 am, "Dmitry Shuklin" <shuk..._at_bk.ru> wrote:
>
> I will be very appreciated for your feedback. Thank You for reading.

Hi,

So you've clearly put some time in to thinking about mappings between OOPL capabilities and RDBMS capabilities. But your essay said nothing about why you're thinking about this. What is the purpose?

There is a general tendency I have observed when people come across a new way of doing things that they try to map their old way of doing things on to the new way. That enables them to use techniques they already have under the new system. However this is not a native way of doing things, and it's always less productive than learning the new way, although it is typically faster.

But if one wishes to master a new way of doing things, one necessarily has to learn the techniques, best practices, and idioms of the new system, and not simply "write fortran in any language" as they say. It quickly becomes necessary to discard any mappings from familiar to unfamiliar, and instead, through study, make the unfamiliar familiar.

Also, if one is thinking of how systems work, it is better to understand and analyze the reasons for various features and see if there aren't other reasons to do the same things. As a simple example, OOPLs have encapsulation as a way of constraining the values that variables can take on. RDBMSs have a different way: declarative integrity constraints. If you come to understand how declarative integrity constraints achieve the same end, I expect that you will then view encapsulation as a poor second choice. At which point you would not see any value in trying to bring that facility into the RDBMS world.

Good luck.

Marshall Received on Sun Mar 11 2007 - 17:57:41 CET