Re: Dreaming About Redesigning SQL

"Lauri Pietarinen" <lauri.pietarinen_at_atbusiness.com> wrote in message news:bmtgio$vvg$1_at_nyytiset.pp.htv.fi...
>
> Bob Badour wrote:
>
> >"Lauri Pietarinen" <lauri.pietarinen_at_atbusiness.com> wrote in message
> >news:bms3b2$lbr$1_at_nyytiset.pp.htv.fi...
> >
> >>Bob Badour wrote:
> >>
> >>>"Lauri Pietarinen" <lauri.pietarinen_at_atbusiness.com> wrote in message
> >>>news:bmrccu$uur$1_at_nyytiset.pp.htv.fi...
> >>>
> >>>>>>Now, the alternative COULD be to widen the domain of the DMBS.
> >>>>>>
> >>>>>Instead of widening the dbms, consider applying the dbms's
> >>>>>computational model to a wider scope of problems.
> >>>>>
> >>>>Yes, thanks for clarifying. That's what I meant.
> >>>>
> >>>I observe that the ability to apply the dbms's computational model to a
> >>>wider scope of problems suggests a single paradigm.
> >>>
> >>It's just two different appoaches to building applications. At the
> >>moment we are stuck in a situation
> >>where we have to live in two (slightly incompatible) worlds to get the
> >>job done.
> >>
> >>The billion dollar question is: could we get everything done in one
> >>world only, the relational one?
> >
> >Of course, we can.
> >
> OK, but playing the devils advocate, how can you be so sure that it would
> - scale
> - be manageable
> - adapt to all situations
> - not to talk about being able to educate programmers to the new
approach
>
> Say if you had a large insurance company with, say, 10000 rules, would
> it *really* work?

Yes, absolutely. It would scale at least as well as it does today. It would be more manageable than it is today because the important business logic would not be scattered among hundreds of applications. It would easily adapt to all situations. Why would it not?

To my mind, whether a particular piece of logic executes on a client or some kind of server is a purely physical issue. To me, physical independence means the dbms can divide the work among any number of processors to ensure scalability and responsiveness. Whereas we currently have database servers, application servers, web servers and clients, I see a dbms that encompasses all of these.

Competing products might divide up the processors differently in their internal architecture. A high-end product might have clusters of storage servers fronted by a logic server farm where traffic is controlled somewhat by a load balancing connection server, and finally accessed through client software capable of not only caching data but of executing logic to enforce many constraints without involving any servers or any latencies.

Educating programmers is another matter considering how few have any real education and how few desire any real education. The current structural changes in our market will force many people out of the market. The million dollar question is: Will it drive out the posers or the competent?

It is not clear to me which will happen. Received on Sun Oct 19 2003 - 15:26:37 CEST