Re: Possible bridges between OO programming proponents and relational model

From: Cimode <cimode_at_hotmail.com>
Date: 15 Jun 2006 10:39:20 -0700
Message-ID: <1150393160.189930.86920_at_c74g2000cwc.googlegroups.com>


paul c wrote:
> Cimode wrote:
> > paul c wrote:
> >> Cimode wrote:
> >>> ...
> >>> I state that BB is wrong in saying RAM SQL tables representations are
> >>> multidimensional...
> >>> ...
> >> I doubt if he said any such thing. More accurate to say that RAM can
> >> represent multidimensional SQL tables, courtesy of human interpretation.
> >> A memory controller has no idea what an SQL table is. OO advocates seem
> >> to be forever boxing themselves into such corners because of neglecting
> >> to separate logical from physical. Alfredo put it exactly when he said
> >> the two are unrelated. Maybe one could define relations to describe
> >> this but it would be pointless since the HW mfr's are unlikely to burn a
> >> dbms into their controllers.
> > If you believe sticking an 'OO advocate' makes you feel more secure
> > about what you know then fine...I really don't care.
> >
> > It's not about boxing it's about discussing a limited scope issue which
> > is physical implementation. Nobody confuses logical and physical here.
> > It's you who refuse to discuss physical implementation issues.
> >
> > What do you think it would pointless to address such issue? The fact
> > that HW Manufacturers won't do it is not an argument to justify it's
> > pointless to discuss the issue of implementation.
> >
>
>
> This thread's subject says "OO programming proponents". Nobody called
> you one.
Of course when some close minded ignorant people refuse to adress specifically this issue, bringing back totally irrelevant comments about logical issues when we are specifically dealing physical implementation.

> I'm fine with discussing physical implementation issues. I don't see
> how one can realize a theory without that. There's some from me and
> others in this group's archives about it but I wish there were more.

Good for you..This could have been a much more productive thread if some people totally stuck on the traditional stupid and sterile *RM is good /OO is bad* (and vice versa) type of debate totally screwed it up.

They refuse discussion whether you bring sound arguments or not. Using buzz words they dismiss as *nonsense* anything that can prove them wrong...How practical...They are just close minded morons who can't think on their own feet....

> However, I don't see much potential in memory controllers since the kind
> mentioned have no programming interface that a dbms implementer could
> get his or her hands on.
How the hell do you know as a certainty that this is not an interresting track to follow? we are here to discuss theory aren't we.

As you said, maybe creating such programming interface may bring a solution...
Dual Core architectures implement a closer relation between CPU and IO physical adressing...As they are already programming interfaces to pilot CPU threading and context switching, what makes you think this is impossible?

> p
Received on Thu Jun 15 2006 - 19:39:20 CEST

Original text of this message