This isn't even intelligent design -- to have to make a choice between equally bad "poems" is simply random selection. I wanted a choice to kill both versions. Random selection operating on random phrases can only produce more random phrases.

There was no rhyme, meter, or any kind of consistent theme to make an intelligent choice of any of the versions I saw, despite 115,000 generations of selection. I predict the end result will simply be a small group of phrases, not even rising to the level of "bad" poetry.

Why not just pass out some typewriters at the zoo?

Word | Number | Number
length | of words | of keystrokes
4 | 2691 | 680
5 | 5553 | 10,700
6 | 9310 | 200,000
7 | 13100 | 4.3 million
8 | 14500 | 115 million
9 | 13698 | 3.6 billion

I computed this table from my copy of Webster's Dictionary.

A sentence that strings together five random English words of 4-6 letters each (assuming equal distribution) would require 8.4x10^24 keystrokes. That is equivalent to the entire earth's population of 6 billion people typing at 200 wpm for 2.6 million years.

A single _coherent_ 5 word sentence of 4-6 letters each (with proper placement of nouns, adjectives, and verbs) would require conservatively (I'm guessing here) at least 100 billion years.

(I should have explained that the above calculations are for Jim's hypothetical zoo animals hitting random keys.)

This is similar to a computer program that attempts to determine optimum network design. It is important to know that, except for trivial networks, the problem isn't solvable.

So the programs just assume some network that meets the requirements. Then they make random changes in the network design and check the results for performance. Those not good enough get "killed off." The rest get subjected to the process all over again.

The results always meet requirements, and always do so better than the original, human (i.e., intelligent) design.

The only difference between the above and the article is the selection means. One uses a computer to assess how good the design is, the other uses human taste to assess how good the poem is.

But in both cases, the design is solely the result of random mutation taking place within the computer.

The beauty of this experiment is that it highlights the difference between selection and design, and that it is easy to confuse the two.

"Here's a recent top vote-getter:

the thousand answer

room tonight

desire when scorn

though returning made and

somehow the

light

the nothingness corrugated"

Anybody feeling the power and beauty of evolution yet?

That sounds like what Paris Review publishes.

Rea does not know anything about darwinian evolution. His program does not, for example, include mutations.

Harry:

Could you give us a list of the people who DO know something about darwinism?

I did, couple weeks ago.

Harold Morowitz writes amusing but solid essays about it.

Peter:
One thing is for sure, people who can't tell the difference between design and selection are at least a few clues short of a full Darwin bag.

selection is design when done by an intelligent being

OJ:
Selection is selection.

When you chose this particular platform as the new host for BroJudd Industries, you selected it from all the others on offer.

You designed none of them.

Presume some computer program randomly mutated the features of web site hosting programs.

Picking one is an instance of selection, despite none of them having been intelligently designed.

Hence my computer network design example. The initial instance is intelligently designed, but none of the succeeding instances are. And the selection process is also not intelligently guided (unless you want to attribute intelligence to computers).

Every time the results--the program produces an equivalent of a phylogenetic tree--work better than the initial, intelligent design.

Jeff:

I present you with and apple and an orange, neither of whuich you designed, obviously. I tell you to choose one. The resulting design was indeed created by you, an intelligent being.

Among incoherent statements, that is right up there.

For starters, the apple and an orange were an apple and an orange before the choice, and would have remained so afterwards, no matter which, if any, I chose.

I give you a choice: Ford or Chrysler. You pick one. That doesn't by any stretch of the imagination mean you have even the tiniest notion how either was designed.

Nice dodge, though, on avoiding the highly successful implementation of Evolutionary Theory.

A Ford or a Chrysler is different than a Ford and a Chrysler. That's designing by choice. Every time an intelligent being makes a choice they design.

Your confusion arises because you believe design consists only of deciding on a car and ending up with what you'd imagined. Design can end in an entirely random result, so long as decisions are made every step of the way.