Re: Databases as objects
Date: Tue, 26 Dec 2006 14:10:53 -0500
Message-ID: <F5SdncewL90l7gzYnZ2dnUVZ_vupnZ2d_at_wideopenwest.com>
David Cressey wrote:
> <snip>
> The object-oriented paradigm actually has room in it for both a data-centric
> world view and a process-centric world view.  But the majority of
> object-oriented SW designers focus on object analysis and design at the
> expense of message analysis and design.  They tend to treat the design of
> messages as a relatively trivial portion of overall system design.  Nothing
> could be further from the truth.
>   
I agree.  One cause may be the fact the OO books focus on OO and not 
systems that exists without OO, or existed before OO, or will exist 
after OO.  I don't believe OO is a fad, but too often its proponents 
behave as though OO is the only way problems may be solved--forgetting 
that problems were solved before OO and are still solved today without it.
> Yes, but what does all of this have to do with the quote I was responding
> to?
>
> Just this:  when you use procedures as the primary intervace to the database
> you have committed yourself to the process centric world view at the expense
> of the data centric world view.  When you call a procedure, you invoke a
> process, by name.  When you invoke a view,  you invoke some data by name.  I
> have more to say, but I need to stop and think.
>   
I'm curious to hear what it is.  Whether you use a view or tables, when 
execution crosses from the application into the database there's a 
boundary crossed.  The first question is whether or not a boundary 
should exist in the first place.  OOers, through various frameworks and 
tools are trying to eliminate it, and this requires (I think) a 
compromise between the best relational model possible and the best 
application object model.
If we instead respect the boundary I believe there's a place to cross 
into the DB and a place to return from it.  Whether that place is raw 
SQL or a SQL macro (procedure) makes a difference to how easily a DB's 
referential and semantic integrity is maintained, how reusable it is, 
and how much frustration programmers create for themselves.
>
>
>   
-- Visit <http://blogs.instreamfinancial.com/anything.php> to read my rants on technology and the finance industry.Received on Tue Dec 26 2006 - 20:10:53 CET
