Re: OO and relation "impedance mismatch"
Date: Sun, 3 Oct 2004 22:38:38 -0500
Message-ID: <cjqgk3$q4u$1_at_news.netins.net>
"Marshall Spight" <mspight_at_dnai.com> wrote in message
news:ev%7d.197678$3l3.147655_at_attbi_s03...
> "Fredrik Bertilsson" <fredrik_bertilsson_at_passagen.se> wrote in message
news:31f7e57d.0410030945.5f3bbfbb_at_posting.google.com...
> >
> > I mean that the OO programming paradigm could be used with RDBs. I can
> > not see that non-OO languages should be more suited for accessing
> > relational databases. In an OO language it is easier to create
> > components for "seamless" access to a relational database.
> >
> > I think that the impedance mismatch is invented by some OO purists
> > that are frustrated that OO databases failed.
>
> I wouldn't say it was "invented" in the sense of something that
> doesn't exist. The difficulties matching, say C++ and Java with
> SQL are very real. But they aren't *fundamental* problems.
>
> I've never had a chance to work with SQL/J, but it looks to
> me like that approach might dramatically reduce the issues.
> Better still would be an OO language that had been designed
> from the get-go to work with the relational model. This is
> the idea that interests me the most.
>
Whereas I would like to retire the relational model as it is implemented today. I realize we can't just retire it cold turkey, but I'm hoping it's the beginning of the end for SQL-based DBMS's. They are one of the components that makes software development so much more complex and less flexible today than in the past, without enough benefits.
OO isn't any silver bullet, but doing both together will never be lean &
mean and I'll vote for keeping the OO and ditching the RDBMS's. We need to
get a better bang for the buck with software development and the database is
one area where I think we can make significant gains if we free ourselves
from the RDBMS approach. There are some new databases that would likely be
called file systems by many folks on this list that were developed by
putting non-RDBMS minds together and coming up with something more friendly
to OOP. I haven't spent much time looking at these, but I suspect many
businesses could do well to eliminate the RDBMS from their solution just as
many sites who never migrated to RDBMS's (oh, yes, there are many) are able
to move further faster and with fewer dollars (sorry to still be preaching
the same sermons, but the more I learn, the more I become convinced that we
can make some significant gains related to software development in the
database area -- as well as in the UI, business rules, and distributed
computing.)
--dawn
> Marshall
Received on Mon Oct 04 2004 - 05:38:38 CEST