Re: A database theory resource - ideas

From: Bob Badour <bbadour_at_pei.sympatico.ca>
Date: Sat, 17 Mar 2007 05:09:54 GMT
Message-ID: <CKKKh.11344$PV3.117030@ursa-nb00s0.nbnet.nb.ca>

Bruce C. Baker wrote:

> "Bob Badour" <bbadour_at_pei.sympatico.ca> wrote in message 
> news:SIGKh.11289$PV3.116242_at_ursa-nb00s0.nbnet.nb.ca...
> 

>>Bruce C. Baker wrote:
>>
>>>"Bob Badour" <bbadour_at_pei.sympatico.ca> wrote in message
>>>news:yaGKh.11280$PV3.116035_at_ursa-nb00s0.nbnet.nb.ca...
>>>
>>>>Bruce C. Baker wrote:
>>>>
>>>>>"Bob Badour" <bbadour_at_pei.sympatico.ca> wrote in message
>>>>>news:TWFKh.11276$PV3.115808_at_ursa-nb00s0.nbnet.nb.ca...
>>>>>
>>>>>>Bruce C. Baker wrote:
>>>>>
>>>>><snip>
>>>>>
>>>>>>>Lots and lots of simple, tightly-focused examples. Think "RDBMS for
>>>>>>>Dummies"! :-)
>>>>>>
>>>>>>I don't see how that will help anyone learn to think in the abstract.
>>>>>>If one can only understand recipes, one can go work for a restaurant.
>>>>>
>>>>>
>>>>>A short, simple example can sometimes bring home the point of a long and
>>>>>detailed technical explanation. One reinforces the other. Don't take my
>>>>>"RDBMS for Dummies" suggestion /too/ seriously! :-)
>>>>>
>>>>>The "... for Dummies" books can indeed be too simplistic, but as a
>>>>>/starting point/ for a complete newbie, they serve very well.
>>>>
>>>>The single biggest problem in our industry is a lack of fundamental
>>>>knowledge and basic education. A recipe book and examples will do nothing
>>>>to address the problem and will serve only to perpetuate it.
>>>
>>>Agreed.
>>>
>>>Bob, have you ever read a "... for Dummies" book? If so, you'd know that
>>>they're /primers/, not cookbooks.
>>>
>>>I stand by my statement that a concrete example can reinforce and
>>>complement an abstract technical explanation.
>>
>>Yes, I have. And I have spent countless hours evaluating various books on
>>the subject of databases. One will find any number of cookbooks full of
>>concrete. One can go to any bookstore and clear out 99% of the computer
>>science section without losing anything worthwhile.
>
> I wholeheartedly agree. Computer science publishing is not what it once was.

If it ever was what it once was, it was before my time.

> Some examples of the ("classic" or current, database-related or otherwise) > one percent that doesn't suck are ...?

For relational dbmses, look for names like: Date, Pascal, McGoveran, White, Darwen

For the physical level, look for names like: Shasha, Garcia-Molina, Ullman, Widom, Gray

Other useful names to look for include Fagin, Kent

There are other useful names but you will never find anything written by them at your local bookstore.

If you are lucky, at any given time, you might find a single copy of two or so good books on the racks buried in a sea of wasted wood pulp.

I haven't spent as much time in bookstores the last three or four years so I have no idea if any new entrants have emerged.

On the programming side of things, one can look for names like Sipser, Knuth, Dijkstra, Plauger, Cargill, Stroustrop, Maguire, Ullman again, Aho, Ritchie, Kernighan, a bunch more from both Bell Labs and Xerox PARC. While things are much better on the programming side, the signal to noise ratio is still abysmal.

>>Even as vocational training books, they suck.
>>
>>Teach fundamentals and principles. Start small and build. Anything else is
>>a waste or worse.

> 
> "We don't need no stinkin' principles. We gotta get that app out 
> /yesterday!" *Sigh*

There is one way your suggestion might have some merit: programmed learning. I vaguely recall Fabian might have once been involved in some CBT material based on the principle. The approach works great for english grammar (see _English 3200_
http://www.eric.ed.gov/sitemap/html_0900000b800c0812.html http://www.eric.ed.gov/sitemap/html_0900000b800bdd24.html), and I assume one could make it work for data management.

However, even more effective than data management, I think a course in programmed learning might work to improve logic, empiricism, overcoming cognitive pitfalls etc. Imagine a programmed course of learning that teaches the lessons taught in _Uncommon Sense_, _How We Know What Isn't So_, _A Mathematician Reads the Newspaper_, _Darwin's Dangerous Idea_, and the Sokal Affair.

Programmed learning works more on the basis of questions and answers than on examples, though, and it definitely follows the principle of start small and build. Received on Sat Mar 17 2007 - 00:09:54 CDT