Re: OO versus RDB

From: paul c <>
Date: Thu, 29 Jun 2006 21:21:18 GMT
Message-ID: <ivXog.106005$iF6.18306_at_pd7tw2no>

Bob Badour wrote:
> paul c wrote:

>> Bob Badour wrote:
>>> Christian Brunschen wrote:
>>>> ...
>>> Fair enough.
>>>  also because I am coming mainly from the direction of
>>>> being a pragmatic software developer rather than a deep theorist on any
>>>> particular subject.
>>> With all due respect, software development is applied mathematics. 
>>> One cannot be pragmatic about it while not knowing the underlying 
>>> theory.
>>> Similarly, pragmatic electrical engineers know Ohm's Law, Maxwell's 
>>> Equations and Stoke's Theorem. I cannot imagine an electrical 
>>> engineer claiming both pragmatism and ignorance of the theory of his 
>>> field.
>>> ...
>> I think this could be qualified a little.

> Anything can be qualified--the question is should it?

With all due respect, I think this statement should. A more basic question in any endeavour is "how do I know what I know?". I think it's basic because asking it can often save one the effort of applying what one doesn't know well enough to make any difference. In every place I've worked, the question wasn't asked enough, eg., in offices with only dozens of people it was clear that a different question, albeit a silent one, was being asked thousands of times a day, namely "what should I do next?"!

> Sometimes it is enough to

>> know that verging into certain territory will put one in over one's 
>> head.  I remember learning years ago that much of electrical 
>> engineering has to do with recognizing when the 'first approximation' 
>> is sufficient for circuit design, eg., certain circuits don't require 
>> the calculation of say, impedance or capacitance, ie., so-called 
>> second or third approximations.

> I will substitute inductance for impedance and assume you mean in
> circuits whose impedance is generally resistive.
> Applying a given rdbms product seems similar, for
>> example, it seems one could spend a few years trying to understand 
>> 3VL, or given the shortness of life, just say no and decide to avoid 
>> nulls.

> You seem to have veered off into a wild tangent.

It may seem so, and personally I suspect that exploration of 3VL as it may be applied to the logical roots of RT without starting from scratch is as irresponsible and arrogant as manipulating DNA to create a superior race, but I'll take ego license to embellish. Years ago a PL/I programmer said he envied me because I knew assembler, therefore I knew how the computer worked. I can't account for exactly why but this made me feel a bit of a fraud and so I signed up for night courses in electronic theory. (I already knew a little about how logical circuits were designed but nothing about how they were enabled.) Some of the introduction, like Thevenin, was compellingly self-contained but after the first few courses, I realized I could go for a PhD in EE and still not know what an electron was, so I stopped. Later, more elemental ingredients became part of common awareness, things that sounded like foreign currencies, muons and pesons and so forth. Meanwhile, electrical engineers of my generation continued to ply their trade without worrying about the really tiny stuff. (I worked with a number of them who went into programming because there was more money in it and there was just as high a percentage of them as of the other programmers who when asked why did they write such-and-such code could give no reason relevant to the task at hand.)

That story may strike as even more tangential (it does to me in the sense that I've long since discarded any pursuit of circuit design) but my point is that what seems tangential today may not seem so tomorrow.

I believe you understand that I'm in no way discounting the necessity of theory, in fact the pursuit of where a theory leads is what I call progress, like the meeting of many different theories that enable the machine I'm using for this message even if sometimes the tangents lead us to learn what NOT to implement. I mention this only to try to be clear to the more casual readers here, especially the border crossers who keep throwing OO stuff over the fence. There are very, very few professional db theorists but many amateurs like me who feel the urge to explore every tangent (there are many who think they are professionals, but if they thought more precisely would realize that they are not in fact being to do theory). In the end, the reason I like "tangents", even though you didn't ask, may be merely "because they're there" and deciding when to stop is as much a matter of taste as anything else. Implementation tangents are a waste of the originator's time, if not always for his observers, as for theory tangents, I think they're a human right, even for the OO crowd. So there!

Sorry for the lack of formal theory in this long post, couldn't resist, seems half the messages in this group are about motivation anyway.

p Received on Thu Jun 29 2006 - 23:21:18 CEST

Original text of this message