Natural V/S Controlled (man made) processes

right in the first part of the sentence is the first error.

None of us consider nature as a process.

The second error is in the second part.

Nature is far from perfect

"Nature is far from perfect". True: far, FAR from perfect. But note that the original post was from someone who thinks that Jesus gave Martin Luther King a massage [sic], and contemplate where imperfection may have entered the discussion.

Depends on what you mean by perfect. Flaws and mutations drive evolution. No evolution, no diversity. So in a diverse and variable environment some random flaws result in enhanced survival and some enhanced extinction. Perfect conditions for evolution but also devolution. Now we intervene with human intelligence but perhaps with as much wisdom. One thing is for sure, time will tell.

If you want to look at a natural process how about the process of DNA replication. Every living organism on earth dose it a couple of billion times with an error rate so close to zero that it would be difficult to calculate. There is an ancient species of pine tree that has been around and remained unchanged for100,000,000 years and every single example today is genetically identical. That means the DNA has been replicating with 100% accuracy for all that time. You think ±0.5% is good over a couple of years with a few thousand copies, How about ±0.0000000000000000000000000% over a 100,000,000 years and I'm not even going to try to estimate the number of copies.

How would you know? with no reference sample from those ages long ago we must rely on the observed mutation rate.

In 1000 generations you can breed a mouse to the size of an elephant and by the time you get there the gestation time will be similar to that of an elephant.

Obviously the genes will not be the same, as they responded to the pressure. In the same way so did the tree or other organism

Aurizon you are of course absolutely correct in saying there is no way to know as we don't have a sample of the DNA from that far back. They do however have fossil evidence of the tree from around 100,000,000 years ago that show no discernable variation from the tree as it exists now. The other thing is that all the trees that exist today are have identical DNA so some time in the past there was only one tree that all the existing trees sprang from. The tree was discovered in a very secluded valley in the Wollemi National Park not far from Sydney Australia around 10 or 15 years ago. This link has more info about it.

http://www.wollemipine.com/

The exact location of the stand of pines is a well kept secret but samples of the pine have been cultivated and it is now available commercially. This is a true living fossil and just the fact that it has survived since the days of the dinosaurs is a credit to the accuracy of natures reproductive process.

I saw a TV program about that pine a few years ago. Like cheetahs, it went through an evolutionary choke point where there were very few specimens.

Nature is a compilation of a huge number of processes, many of which operate in conjunction with many other ones. Man has been able to "copy" a (relatively) few of these. I believe that nature AT THIS TIME is indeed imperfect. At present, entropy is increasing on a macro level. We as humans, with our God-given intelligence, are able to overcome this in some instances on a micro level. As for tolerances, is is we humans who developed them. Nature's processes are so intertwined that to even apply tolerances to them is almost pointless. In my opinion, nature is much too complex to compare with any man-made process (even very complex ones)!

I believe one of the logical errors here is he compared highly involved very complex natural occurences with very simple human processes. Just the number of variables that effect growth, compound the varaince. You would need to reduce the analysis by removing the variance introduced by all these variables down to something similar to that in the simple few degree of freedom human processes. The DNA replication process is a good example of much more limited degrees of freedom for comparison to human activities. A tree or human growth pattern to a simple machining process might be comparable to the development of a complete major metropolitan or national infrastructure system from scratch, with all new components, at an accelerated rate. Each time this infrastructure was completed in a new community would it be exactly the same as the previous with in 0.5% tolerance?

There are lots of examples of "natural processes" having very low tolerances, or the majority of results being virtually identical, just as there are many examples of man-made results having wide variations.

In another thread, it was pointed out that amoeba, one-celled animals, reproduce by splitting into two identical amoebas. If they are identical, with the same DNA, then all the offspring of these amoebas would have the same characteristics, so variability is nil. Has anyone ever heard of mutant amoebas occurring in nature? Yes, other animals have naturally occurring mutations, with fluctuation in DNA of offspring, but not in this case!

Then again, as a design engineer, I know that some man-made tolerances in manufacturing are selected almost arbitrarily, sometimes too low. This often results in a high scrap rate. Often, a re-evaluation of those same tolerances will reveal that they were artificially low and that the tolerances can, in fact be "loosened" and we will still have good products which meet all of the requirements that we need, but in fact there will be a wide variability in how the parts might measure. We often do this when we want to "safeband" something, i.e. to insure that all of our parts fall outside of some one-sided tolerance. We may in fact accept more variability than we might otherwise.

The flip side of this is that when tolerances are set too high and we have a high variability, we can start with practically no errors in production, but the product will not survive very long in use, due to mis-understanding of the actual requirements. One could say this "process" of setting requirements by man would then have a very high error rate indeed.

I think a good point here is not to confuse artificially set tolerances with "natural tolerances" or probability distributions.

Take the example of human height given by the author. This is so funny as to be ludicrous. This is a case of begging the question. Any process which sets limits on the basis of +/- 3 sigma (sigma being a variance equal to one standard deviation) guarantees that at least 97% of the values are within 6 standard deviations from the mean (using Chebyshev's inequality). Which means that this "natural" process has an allowable deviation from the mean about +/- 47.5 percent. If it turned out the distribution was a "normal" bell-shaped curve, the allowance would be +/- 49.85 percent! It would, in fact, be much higher than the measly +/- 30 percent which the author must have guessed at.

I saw segment on a TV show how burlwood smoking pipes are made. The craftsman was preparing blanks for carving. This process, of selecting the wood, choosing where to make the cuts and then cutting on those lines results in only one good blank out of three attempts. This is because of "defects" in the wood which would allow air flow through the side of the bowl. With two out of three blanks being unacceptable, our tolerance would seem to be +/- 17.5 percent. Yet, this is an acceptance versus rejection (error), not really a measure of variability of a dimension. So one must be careful about how one defines a "tolerance".

All in all, I do not think we are really comparing things fairly. It is almost absurd to think of natural process as having "tolerances", but rather to think of them as having "variability". If we compare "variability" of natural process to man-made process we might have a more valid comparison.

More to come!

You said

"Any process which sets limits on the basis of +/- 3 sigma (sigma being a variance equal to one standard deviation) guarantees that at least 97% of the values are within 6 standard deviations "

This is wrong

right value is 99.73%

Mr. Guest #11,

You had better check all of the words and figures before saying that someone else is wrong, or you will end up with egg on your face.

The statement you quoted is NOT wrong. Please note that I said "at least 97%". The figure you gave is for a bell-shaped "normal" curve. Please note that not all statistical distributions for processes follow the "normal" curve, but do comply with Chebyshev's Inequality, which gives a minimum of 97% for the 6 sigma distribution. I did go on to give the distribution for the "normal" curve, based on the rounded off figure of 99.7 % instead of the extra decimal place which you provided at 99.73%. Big Deal!

I think that you are wrong to say that I was wrong. How about that? Would you care to post a retraction?

Engineer! guest may be wrong but at the end he is guest?

eggs on the face of guest! ooooooooooooohhhhhhhhhhh how will it look?

rakesh_semwal, Now you are just being silly, so I will too! Firstly, he is not MY guest! He is only a guest because he did not sign in. Secondly, here is a possibility for "how will it look"!

Well engineered.

But I guess egg is semi boiled.