Re: Object-relational impedence

Brian Selzer wrote:

> "S Perryman" <q_at_q.com> wrote in message news:frot80$5k7$1_at_aioe.org...

>>Their system used an RDBMS. And it performed poorly.
>>The same systems subsequently built on the same platforms (HW, OS, comms,
>>prog langs etc) , but using an OODBMS instead, performed orders of
>>magnitude better.

>>That's life.

> Funny, but this "orders of magnitude better" claim sounds like something a
> shifty politician like [ reference to specific politician snipped. ]

> I once altered a system that was designed to
> process only 14,000 transactions per hour so that it could process ten times
> that in the same time. That's orders of magnitude improvement

One order of magnitude.

> but I didn't
> change the DBMS, I rewrote some poorly written procs. I once altered a
> system that had a job that was taking over 25 hours to process so that it
> took less than 30 minutes by simply adjusting the way the hardware was used.
> Again the orders of magnitude improvement was not due to changing the DBMS,
> it was making the best use of the available hardware. It is also often the
> case that you can obtain orders of magnitude performance improvement by
> simply adding an index--at least for queries that can take advantage of the
> index.

We've all been to these places.

> Your claims of orders of magnitude better performance, therefore,
> are much like the claims of a used car salesman, or of a sleazy lawyer.
> Suspect.

As measured by the developers who did the profiling of the operational sequence on the system that used the RDBMS, and when the OODBMS was used instead.

This was a "mismatch impedence" of the purest kind.

Regards,
Steven Perryman Received on Wed Mar 19 2008 - 18:03:45 CET