TCB in TN wrote:yet since that time, many of those "missing links" have been found and the fossil record is much more complete.
I think part of the problem is that most evolutionary biologists are still looking at morphology (a 19th century approach that looks at body shape, body plans, skeletal structure, etc.) to infer common descent, with the occasional look at phylogeny. Even with that, the fossil record pretty much falsifies Darwin's theory of common descent. However, we now know that the structure of a plant or animal is largely determined by the information encoded in its genome. When we look at the quantum leaps in genetic information between, say, a fish and an amphibian, or between a chimp and a human, the gaps in genetic information are staggeringly enormous, making gaps in the fossil record look like small potatoes by comparison.
To illustrate using MicroSoft Word and Apple's 'Pages', they are both word processing software that may have many similarities in appearance and end result. An evolutionary biologist who looks at morphology could conclude that the many similarities are a result of common descent (i.e., both evolved from a common older piece of software, without any help from intelligent humans, through a process of random changes to the code and when something worked, it's fitness preserved the changes). However, when we look at the actual coding differences between MS Word and Pages, that theory goes out the window.
Similarly, the chimp and human have a lot of structural similarities, and evolutionary biologists have inferred, therefore, that we have a common ancestor. However, with the sequencing of both genomes, we have found that there are approximately 3,000 genes difference between the two (not to mention massive non-protein coding differences). There isn't enough time in the universe to evolve one novel gene of average length, forget about dozens, hundreds, or thousands. Recent work has shown that the probability of finding just one average length protein-coding gene in sequence space is approximately one chance in 10 with 210 zeros after it. The upper limit for what the universe is capable of performing in a search over 10 billion years might, at best, locate a sequence that had a probability of 1 in 10 with only 119 zeros after it. In other words, the search capacity of the universe falls almost 100 orders of magnitude short of what is required to find any sequence at all that codes for just one average protein. The upper limit for the total number of trials on earth over a 4 billion year span is, at best, around 10 with 41 zeros after it.
Bottom Line: Writing any kind of computer software at all that is of any significant length, using a process of mutation/insertion/deletion is not going to work anywhere in the universe; the number of trial available is simply far too miniscule in comparison with what we see encoded in MS Word, which is far less impressive than what we are seeing encoded in the genomes of life.
Perhaps the best way to come face to face with the difficulties is to start writing genetic algorithms. They will not get off the ground without a fitness function and the fitness function must always encode sufficient information to properly identify the desired outcome. It turns out that to do this, the fitness function must always encode more information than the desired outcome. So before you can evolve the information required for life, you must first produce a fitness function in nature that actually contains more information than what you have to evolve. Unfortunately, most evolutionary biologists have no experience writing evolutionary algorithm software. Within that absence of knowledge, speculation can flourish unchallenged. That is slowly changing, however, with a few landmark papers on functional information and functional complexity having been published over the past year.
By gum! What am I doing! I come here to discuss shooting old leverguns!
