All Claims Made as the Result of a Computer Simulation Should be Considered BS, Until Proven Otherwise
| July 20, 2012 | Posted by GilDodgen under Intelligent Design |
No one should give any credence to computer simulations who is not familiar with what they are and what they can do.
In my work in aerospace R&D I was recently alerted to a severe problem in a computer simulation. The original simulation said that a mechanical part was half as strong as empirical testing said it was. I triple-checked all the material properties of the part in question, and ran two simulations based on the original model. Both simulations were clearly in error.
I then noticed something. The solid finite elements were modeled with four-node tetrahedrons, not ten-node, second-order tetras that would provide the requisite degrees of freedom to make the model work correctly.
I realize that what I have written will make eyes glaze over, but the point is:
If a single oversight such as I have discussed — in a computer simulation using the most advanced technology available, with the most experienced people in the field doing the work, and with all the material properties thoroughly known and characterized for a simple mechanical part — can result in a conclusion that cannot be trusted, how much trust can you put in claims made by those who produce computer simulations about what the earth’s climate will be like in a hundred years, or about how computer simulations verify that Darwinian mechanisms transformed a microbe into Mozart in 10^17 seconds?
My advice to all real skeptics — as opposed to Darwinists, who are pathologically immune to any skepticism concerning the creative powers of Darwinian mechanisms — is: Assume that a computer simulation is completely unreliable (and even worse, possibly modified to produce a result consistent with a conclusion that was reached in advance), until proven otherwise.
13 Responses to All Claims Made as the Result of a Computer Simulation Should be Considered BS, Until Proven Otherwise
Leave a Reply
You must be logged in to post a comment.
Gil, you might be interested in this article that came out today:
Reminds me of this comment from Paul Davies:
Although they are still probably a very long way from having a accurate model of the simplest organism known to man (You can obviously comment more completely on that Gil), it still is, at least, a far more fair simulation of reality than evolutionary algorithms.
Speaking of evolutionary algorithms, I had this exchange with a neo-Darwinist yesterday on Evolutionary Algorithms:
me
he replied
to which I replied:
and he replied:
to which I let it end on his vaporous dreams of scientific integrity
BA77,
I automatically assume BS, until proven otherwise. We see claims like this all the time, each one more absurd and ridiculous than the last, and promoted with evermore certitude about how the problem of the origin of life is about to be solved in purely materialistic terms. Computer simulations prove it!
This keeps the public funding flowing for the rainbow chase.
I’ve written software of all kinds for almost 40 years, I’ve taught a range of undergraduate CS and CIS courses, and consulted in many areas including software quality assurance. No non-trivial program is bug-free; no, not one. Two things cause people to earnestly believe that their simulations are reliable – hubris and agreeable results.
Gil, I’m going to keep beating this drum. A qualitative theory is to a computer model what Rolling Stone articles are to formal logic. It’s all just a manner of reasoning, for some value of reasoning, but the computer model is at best a concrete and falsifiable presentation of a theory that was previously written on the back of an envelope.
The results of a computer model are the hypotheses to use in science. They are in no manner the actual empirical observation, out there in reality, of what occurs with highly symmetric children.
All simulations are based on simplifying assumptions.
The consequence of this is that one cannot take a simulation as giving truth. But a simulation can be a useful guide. The value of the guidance given needs to be subject to empirical testing
I’d say that the view that you (Gil) expressed is a bit too negative. You would not be using simulations at all if you really believed that. However, one does need to understand the limits of simulation when using it.
As an example, take the simulations of Darwinism. My personal reaction is to not pay them much attention. That is to say, I take them with a grain of salt. And if I (as an evolutionist) don’t find them persuasive, I keep wondering why anyone would expect a critic of evolution to be persuaded.
The main problem with simulations is that you cannot simulate what you do not understand. And that is where evolution comes in- no one understands it so no one can simulate it.
to bornagain77-
Liz’s GA does not use natural selection and does not create CSI as she starts with the very thing that needs expalining in the first place- self-replicators. Ya see Liz has her binary sequence of coin tosses replicate with variation and that is the very thing that needs to be explained, THAT is the CSI she has smuggled in.
Unfortunately she is too dim to grasp that even though I told her (and provided a reference). That other evos are jumping on her wagon just demonstrates their ignorance and desperation.
Rule for models, computer and otherwise: validate, validate, VALIDATE. Through real-world observations. Never forget, GIGO. And, where somebody is trying to sell something, whether for cash or an idea or agenda, never forget that liars and fools can figger, too. KF
I forget, no computer simulation is equivalent to a real world experiment or observation.
Gil: “Assume that a computer simulation is completely unreliable (and even worse, possibly modified to produce a result consistent with a conclusion that was reached in advance), until proven otherwise.”
Like this one:
http://www.youtube.com/watch?v=WnXeUIaYj3k
Gil: “Assume that a computer simulation is completely unreliable (and even worse, possibly modified to produce a result consistent with a conclusion that was reached in advance), until proven otherwise.”
Especially when they won’t release their computer data:
http://cryptome.org/nist070709.pdf
A simulation of a life process is subject to the representational vaules set within the simulation, and is either useful or not based upon those values. But the simulation of the origin of a life system must simulate the rise of representational values themselves. Pattee once stated that life is matter controlled by symbols.
Good luck simulating the rise of a representation.
Joe,
The main problem with simulations is that you cannot simulate what you do not understand. And that is where evolution comes in- no one understands it so no one can simulate it.
Amen raised to the Xth power, where X is large. Joe’s comment encapsulates the essence of my thesis, expressed with far fewer words and much more eloquently.
I’ve always been a detail guy — whether learning, memorizing, and performing every single note in an 80-page piano score of a Rachmaninoff piano concerto, or writing complex computer programs — but immersion in detail can deceive one into not being aware of the overarching theme.
Darwinists are hopelessly lost in the speculative details of their theory, which they must propose and defend at any cost — forget reason or evidence — because to do otherwise would be to admit that their materialistic worldview is indefensible and that design is evident.
The reason I make my posts at UD is that I was once in the Darwinist camp. I was thoroughly indoctrinated with Darwinian argumentation from childhood, through college, and beyond.
But then one day I was exposed to rational argumentation and evidence that convinced me that I had been deceived, and that Darwinism was never about science, but about a long-awaited creation myth for a materialistic, nihilistic religion.