[comp.databases] Seamless integration of type systems

dennis@boulder.colorado.edu (08/22/89)

I would like to challenge the idea that seamless integration
of a programming language with the database
(aka persistent programming languages)
is a good idea.  I am involved in a project here at Colorado
which is already using the following programming languages
trying to access the same database:
	1. Ada
	2. C
	3. C++
	4. Prolog
We may eventually end up adding several other languages.
In this situation, there is no obvious one language in which to embed
the database type system.

Instead, we appear to need a database model that is relatively
independent of any one programming language, and that can be
accessed successfully, if not cleanly, from many different languages.
For our purpose, we have some small evidence that relational
models work better (in a best-fit sense) than OO models.
It is true that an OO model would fit C++ well, but it
doesn't fit well with Ada, and certainly not Prolog.
Whereas, there are well defined procedural interfaces
defined for relational models.

-Dennis Heimbigner
University of Colorado, Boulder
(dennis@boulder.colorado.edu)

moiram@tekcae.CAX.TEK.COM (Moira Mallison) (08/25/89)

In article <10931@boulder.Colorado.EDU> dennis@boulder.colorado.edu () writes:
>I would like to challenge the idea that seamless integration
>of a programming language with the database (aka persistent programming 
>languages) is a good idea.  

Dennis provides a counter-example to defend the above statement.  But
because it is not a good fit for your project doesn't mean it is 
not a good idea. 

>For our purpose, we have some small evidence that relational
>models work better (in a best-fit sense) than OO models.

At least Dennis qualifies his statement here.  It is not the
object-oriented defenders who are claiming to have a universal
answer here.  OODBMS's do not claim to be all things to all people.
There are clearly applications for which the relational model
works well.  There are also clearly applications which are
awkward and cumbersome when constrained to a RDBMS implementation.


Moira Mallison
CAX Data Management