Re: State of IT and DBMS expertise

"Dan" <guntermann_at_verizon.com> wrote in message news:1YCjd.3500$DB.482_at_trnddc04...

> In general, I couldn't agree more. I think one of the problems is that
> implementing an SQL table is soooo easy that designers just implement
> without considering or incorporating an integrated design. I can't say
for
> sure if this is the case or not. Another issue might be that IT
personnel
> like to construct things, even if the wheel is already built and
> sufficiently functional. It's a crimp on creativity.

Yes. This touches on my own personal history. In about 1978, I was exposed to a CODASYL type DBMS by a colleague. If I had wanted to, I could have gone on to being an instructor on a particular product for Digital. I looked at DBMS and said (to myself) "Not simple enough. There has to be an easier way"

A few years later, the same colleague introduced me to some magazine articles about "how to use relational principles, even if your stuck with indexed files." This stuff was beautiful, and relevant to my work at the time. And I didn't even have an RDBMS (or SQL) available to me at the time. When one came along, I used it.

There are some ways in which a DBMS is better than a file system. There are some ways in which it adds, as Dawn keeps pointing out, bulk, cost, and delay.

So, as a consequence of my own personal history, my tendency is to do the best I can with the available tools. Someone in this newsgroup once called that "the Anglo-Saxon infatuation with pragmatism." Well, gee whiz, pragmatism may be ugly, but it works. I am sure you, Dan, see the subtle humor in the self referential nature of the preceding, but some people won't see the humor, unless there's a smiley face. So here it is: :-)

> I do see a greater number of people who haven't necessarily gone through a
> formal computer science or information technology program being involved
in
> information technology, however. It is easy to see how formal fundamental
> concepts just might not have as much relevance as the numbers grow larger.

And these are precisely the people that Dawn cites as the reason for the durability of Pick based applications. The fact that IT amateurs built them. There are really two ideas there:

The first is that certain subject matter experts have a knack for understanding cybernetic systems (and by "cybernetic" I mean manual systems as well as automated ones), and for specifying how they should be automated. The best set of specs is a working system. Maybe it's easier to learn Pick and write a system in it, than it is to learn how to write specs that will effectively communicate with IT professionals. I actually think that this is where Dawn's contribution has been underestimated in this forum.

The second idea is that IT people are basically stultified, uncreative, and lazy. I can't express my disagreement emphatically enough. IT people were the heroes of the 20th century. Yeah, the physicists provided the theory for the atomic bomb. But IT people cracked the Enigma machine and the Japanese code. And that's just for starters.

Without IT, the forward progress of industrialization would have stalled at the 1950's level.

>
> I also see a disturbing increasing reliance on tools. To exacerbate this,
> (this is speculation, mind you) businesses don't want highly trained and
> expensive personnel. They want to have systems simple enough where the
> normal person can walk in and pick up everything that is needed for their
> function within a reasonably short period of time.
>
That's another long topic. In short, the special forces and the infantry don't recruit the same kind of people.

> IMHO a good data modeler could possibly know/discover even more than them
> (as individuals but not necessarily collectively).

Good modelers may be IT people, or they may not be. The best DBAs I've ever seen are world class modelers.

> Not a problem. We can all benefit by letting off a little stream I think.
> >
> > But I'd have to say that it's unusual to run into DBA's who don't know
> > simple design.
> >
> You have yet to be where I've been. :-0. In all fairness, many of them
do,
> but mention serialization, atomicity, the ANSI-SPARC 3-schema
architecture,
> or other rather esoteric things and their eyes roll back. Actually,
though,
> it is very unfair of me to generalize in such a fashion. There are good
and
> bad everywhere.
>

Ooops. I know a lot about ACID and transactions, but I don't know ANSI-SPARC 3-schema architecture. So maybe you've caught me in your net. Does it have anything to do with conceptual data modeling, logical design, and physical design?

If so, then I have a point to make: I'm more ignorant about computers than I was 40 years ago. It's not that I'm getting dumber. It's that the amount of stuff worth knowing is expanding faster than I can learn. Maybe a lot of people have the same problem.

> I agree, though I also agree with those that hold the possible hope that
> things can be made even better.

Strongly agreed. We are not at the end of the road. Received on Mon Nov 08 2004 - 13:19:47 CET