Information, again

[billwald]

From "Scientific American, "Parallel Universes:"

But an entire ensemble is often much simpler than one of its members. This principle can be stated more formally using the notion of algorithmic information content. . . . set of all intergers . Which is simpler, The whole set or just one number? . . . the entire set can can be generated by quite a trivial computer program whereas a single number can be hugely long. Therefore, whole set is actually simpler.

------------------------

This conforms, I believe, to Shannon's thesis - the exact opposite of Behe's interpretation.

 

[Athene]

I don't really understand this - what was Shannons original thesis and what was Behe's interpreation of this and why is it significant?

 

[Deamiter]

It's a bit of a fascinating idea, and quite valid for number theory! I'm not sure it applies to genetic populations however.

The premise (or my understanding of it -- I only read the OP) is that a set of numbers or a set of anything that can be described by an algorithm is simpler to describe than many of the individual numbers.

In contrast, I'm pretty sure the genetic makeup of any particular population would be quite impossible to describe by an algorithm (while there are patterns, deviation from the patterns are relatively random). Further, the number of base pairs in an individual is limited, so there is no infinite bound on the complexity of a particular organism's complexity.

I think it's just another definition of complexity (like the "information" that often gets debated here) that doesn't quite apply to genetics. It's a fascinating concept though, even if it doesn't really fit where you're trying to put it.

And please forgive any terminology errors -- I'm a physicist, not a biologist or mathematician, so I know that while my point is probably valid, I've probably massacred the use of a specific term here or there.