Fables

Evolution

Not everything has an effect on evolution.  Some traits survive because they are linked to other traits that have positive results.  Why do we have vestiges of things we no longer need?  e.g. muscles to wag a tail, goose bumps to raise our fur for warmth an appendix?

Occam's Razor

Truth raises intelligence?

 

• A definition of intelligence

• Neural net formation? Neural nets and epiphany.?

• Jacob Bronowski’s Criticism of Science and “Simplicity”

• The Shaman

• Info Theory/Computer Thoughts

 

Modularity

 

• There is no map theory 

• There is no map of all maps 

• A map is a theorem (about what it shows)
• If there were a map theory, it would be a meta-theory that could construct all other theories. 

• Maps are constructs made up from the maker’s invention and conventions, which build up over time. If you think maps are obvious, you should take a look at city maps from other cultures e.g. Japan

 

• A program is an information map. Therefore there can’t be any programming theory. 

• Programs are constructs made up from the maker’s invention and conventions, which build up over time.

 

• Modularity ultimately fails when facing tight, elegant design.

 

• The real reason that Sony became the benchmark standard for TVs (esp. late 70's early 80's) is that the sets were tightly integrated and carefully built. (Hand eye coordination of Japanese women) Sony's were anti-modular at a time when the competition was touting component design. Sony sets were very reliable but, conversely, expensive and difficult to repair.

 

• Modularity is paid for by the interface. Modular products require expensive and finely constructed interfaces.

 

• Some examples; modular garden tools, and modular tools in general, cost more per module than each item bought separately. A modular handle costs about $10, a shovel module about $10, and a rake module about $10. Home Depot will sell you a rake and a shovel for about $8 each! and you don't have to switch handles

 

• In addition, there is the Shopsmith (a tool which can be reconfigured) conundrum where changing the tool to be something else significantly increases the time and labor involved. It was perfect for the hobbyist with limited space and lots of time.

 

• Each new wave of “solutions” to the programming problem is led by the very best people who move one to something else when the mob catches up. It’s similar to education fads, which produce great results as long as very bright, dedicated, energetic people are doing the teaching. As soon as mortals are involved, the fad fails. The same for programming.

 

• The list of “solutions” is long; Modular programming, higher level languages, structured programming, 4GLs, CASE, OO, etc. ad infinitum, ad nauseum!

 

• The current flavor of the month is object-oriented programming. It is not only being touted as the solution to the programming problem; it rode in on the back of the most over-rated language in computing, C++, which has been coerced, into a bizarre and unfathomable form! C++ is more peculiar to each local project than even the “standard” c libraries. Worst of all, OOP is the idea of modularity writ large. Once there are enough components, all programs will be simple snap-together projects!  Never hapens!

 

• Organizations, which zealously pursue OO soon, have standard components numbering 10,000 and up! It becomes a Shopsmith where people spend more time learning which component does what then they spend in actual programming.

 

• There is no standard OO interface and even tightly run organizations have trouble making an enforced interface work with all components. Objects become coerced to the interface at the cost of function!

 

• If you still believe in OO after all than please tell me:
• Why is software always late? 

• Why are major products buggy? Is Bill Gates stupid? Are Microsoft and the other companies too obstinate to use and realize the benefits of OO?

 

• Why did the company that made the biggest commitment to OOP die in a palace in Scotts Valley?

 

If We Can Put A Man on the Moon Why Can't We Solve the Traffic Problem? ...cure AIDs? ....(and other ignorant questions) 

 

The above question is a major cliché of our age. The premise is, if we can put a man on the moon then we should be able to solve everyday problems on earth. The question is usually a preface to a criticism of some policy, medical cure, common problem, etc. The sub-text is that the problem is solvable if only "they would do something" or "more money is allocated".

 

The Answer to the Traffic Question

 

The generic answer to asking such questions is generally a symptom of ignorance (more on ignorance later). The specific and long answer is that going to the moon an engineering problem, but traffic control is an emperical art and guess work. All the theory needed to go to the moon was in existence by 1900 and produced by people like Isaac Newton, James Clerk Maxwell, Alessandro Volta, etc. The traffic problem has no theoretical base. The science of physics in the 19th century produced thermodynamic theory. That body of theory lead to the development of various power devices. Power devices lead to the automobile that produced the traffic problem. There is no theory for optimizing flow through a network (that's why airlines fly spoke-and-hub). If you think you have a theoretical description of or a solution to, traffic problems, you will be richer than Croesus! The short answer is, we do not know how to do it.

 

"Aha," you may say, "the phone and power companies do know how to optimize flow through their networks!" Not exactly. The utilities have empirical knowledge from working with relatively fixed configurations and predictable traffic loads. The utilities also have better strategies for dealing with failure than do traffic engineers. In spite of the extensive knowledge base of utilities, power and phone grids do fail and often in unexpected ways. The failures occur precisely because there is no general theory for understanding the problem!

 

Ignorance and Profound Ignorance

 

The cliché question comes from ignorant individuals. Ignorant in the sense that they do not understand the link between theory and practice. People who don't understand the link, become angry when "they don’t do anything." Ignorance and the anger often lead to bad policy. The common result is that money pours into research not founded on theory. The Japanese "Fifth Generation" computer project and the US "war on cancer" are two examples.

 

Profound ignorance is the lack of theory itself. How do we invent new theory? No one knows and this essay does not intend to deal with epistemology (theory of knowledge). It seems to occur when genius has enough resources to do what it wants and is more or less left alone.

 

What Does All of This Have to Do with Information Systems?

 

Well might you ask the second question. About every five years, artificial intelligence crawls out of the grave and begins biting investors on the pocket book! In an industry driven by science and theory, one would expect recognition and dismissal of bad theory. Unfortunately, AI is back and causing problems again. The specific areas are handwriting recognition and voice recognition.

 

A little background may help. All AI theory had been expounded and put into algorithms by 1969. Time and again, AI produces demonstrations that never scale up to solve a real world problem. Examples include language translation, pattern recognition, and general problem solving. A famous program “Eliza” demonstrates that AI does not work but that a clever program can make people believe that it does. To this day, some (ignorant) writers cite Eliza as a “good” example of AI!

 

The recent bad example is handwriting recognition. The methods used are essentially the same ones that existed in 1969. The recognition rate has never improved since then but people still set out to create systems. What happens is that someone puts together a program to demonstrate “proof” of technology. Managers watch the demonstration and conclude sagely that “with a little tuning of the programs and better hardware” the technology should be ready for the marketin a few months, if not years. Time passes and with optimized programs and faster hardware, the AI programs have the same error rate but run faster!

 

An article in the Wall Street Journal criticized Apple's PDA, the Newton. Apple's response was the Newton’s recognition was as good as anything in the market. The claim resembles the aspirin companies claim that no competing product is stronger. They all have the same ingredients. The Apple claim is true; no other system produces errors any faster than the Newton! The Newton had many virtues butcouldn't overcome its (marketing) reliance on AI.

 

The chief scientist of a major computer company recently revealed that his company would soon have “workable” voice recognition. He seemed puzzled when I asked if any theoretical breakthroughs had occurred recently in pattern recognition. He was openly hostile to the issue of AI not being usable technology. It was a discouraging conversation since his company appeared to be following the same old AI pipe dream. He claimed unequivocally, that the technology would work “real soon."

 

Will AI ever work? If we have a theoretical breakthrough, then the answer is, of course, yes! If there is no breakthrough, then it requires brute force trial and error methods. Programs now play chess at the grand master level. The programs don’t use any general theory of chess, but rather keep many siituations on disk (as an example, over 150,000 opening games) and have different methods for different areas of the game. Handwriting and voice recognition may be tractable to such a brute force approach. The problem is that it took almost 25 years work by some of the best computer scientists worldwide, to produce the chess playing programs. Can the industry afford to fund hundreds, if not thousands, of PhD. candidates to work on AI problems?

 

A final note on AI. If you still believe that it is a workable technology, check into how many PhDs are being awarded in it each year.

 

Smart/Dumb

 

When are dumb people smarter?

 

• There is no greater or frenetic effort than a smart guy covering up a mistake

 

• An old saying is that the first step to wisdom is knowing that you don’t know

 

• Smart people think that they know, or can know, everything

 

• Dumb people know that they don’t know a lot of things

 

• There’s often a tortoise/hare effect with dumb people bossing smart people and/or doing better later in life

 

• Dumb people develop methodical ways for learning

 

• Some dumb people learn enough over their life that they are, in effect, smarter than a lot of smart people

 

Genius

 

• Genius, like power, is its own reason

 

• Preservation and self-promotion are the first and only priorities

 

• Enlightened self-interest has little or no meaning since the inertia is always

forward and goals always immediate. This leads to an internal logic that dismisses outside ideas and suggestions

 

• Listening occurs only at immanent danger (a gun to the head usually works)

 

Security’s Three Laws

 

With A Thanks to Roger McKee

 

The first law of security is that the major effect of security is the

punishment of honesty  i.e.security is a tax on honesty

 

If you’ve locked your keys in your car, tried to board an airline, or

had to fight your way into a pill container, then you have been

taxed by security.

 

You can’t stop the crazy or the highly skilled person from penetrating

your security.

 

Presidential assassination attempts, IRA crazies and their bombs, and Internet hackers are examples. You can prepare for the event but you can’t  guarantee that you can stop them.

 

The most effective security is to make the cost of penetration higher than the target is worth.

 

This leaves us with the crazies. Human nature is such that no matter what we do, no matter how good or fair we make the world, there will be people who hate the result.  9/11 is a notable example.

 

In Franz  Kafka’s short story, “The Burrow” AKA “The Labyrinth”, a mole is narrating how he is building false entrances and blind alleys in his home in order to outwit “the beast”. Finally, the mole reaches a point where one more blind alley will make his labyrinth perfect and the mole breaks through to the surface.

 

At that point, the mole realizes that “the beast” may have entered the burrow, and the mole has no way of knowing where it might be.  The tale ends with the mole quaking in terror on the surface, the riskiest place of all, and afraid to enter his secure haven.

 

No matter what you may value in life, at this very moment fate may be destroying it; your health, wealth, family, whatever. Fortune can be very cruel. The only security is knowing that there is no true security and that one must live life by taking risks since even the safest of burrows may be a hazard.

 

If you think that life and fate are unfair for putting what you value at risk, remember that life and the things of value in it are all gifts.  

 

Fables

 

Hunter-Gatherer

 

Light

 

Carrying a load

 

Organizing tools

 

The pantry vs. toolbox

 

Neatness

 

Hunter’s concentration and gatherer’s

persistence

 

Focus vs. gestalt – Tunnel vision vs. whole field

 

 

Cosmology

 

All knowing-all powerful conundrum

 

If a being is all knowing than it can’t be all-powerful. if an

all-powerful being could create a random event i.e. something that it

couldn’t predict or know. An all-powerful being therefore can’t

be all knowing.

 

Source of big bang?

 

Entropy release (as in magnetic refrigerator)