Re: Real World (Re: Mixing OO and DB)
Date: Mon, 17 Mar 2008 17:49:56 -0700 (PDT)
On 18 mar, 00:14, topmind <topm..._at_technologist.com> wrote: [Snipped]
> > The most basic declarative solution such as setting up a constraint
> > involves more abstract thinking (intersect beween a relation and an un-
> > ary domain attribute) than the most sophisticated procedural
> > subroutine could ever provide.
> > Only ignorance and stupidity can hide that fact from your puny mind
> > who tries to lock us up into low level procedural approaches.
> That reply was not to me, but I'll chime in anyhow. Declarative
> "programming" using sets etc. indeed can be tricky at times.
Who mentionned trickiness ? Who tried to sell that a set based thinking would be natural or intuitive ?
The only sale here is the one attempted by the Kazarov moron who's got the nerve and arrogance to try selling us fuzziness under the wax of abstraction when he can't even make the difference between abstraction and low level thinking even if ihe's burried in it.
> Procedural is more natural to most people because it mirrors the step-
> by-step process descriptions we grow up with. Plus, its often easier
> to "intercept" the intermediate steps for debugging.
I must admit that procedural is more natural than declarative. Natural, intuitive can not be abstract. 8 out of 10 times it simply leads to fuzziness.
> This gets to the issue of whether we should try to bend our mind to
> fit our software or vise versa.
RM is about total control over programming. A program is simply an implementation of a relvar under a specific presentation protocol and nothing else. The mind bending is simply about delimiting and formalizing the segment of reality we are interested in.
What is mind bending is the fact of introducing science fiction into the equation , put on a crappy OO procedure in place (that works one out of 10 times), and let others figure out our mess. OO programmers are usually the poorest programmers that can be. In 30 years I have not met a single one that could get a program to work right.
> There may indeed be a declarative
> genius or two who can write a general ledger system with just a 10-
> line relational formula. However, it is unrealistic to expect this of
> everybody. The goal is not job security for the select few Einsteins
> who can pull that off.
One does not have to be a genius to favor set oriented approaches only *aware* of its superiority. Gymnastics and discipline do the rest. Received on Mon Mar 17 2008 - 19:49:56 CDT