In recent years there have been a number of discoveries in genetics and embryology. In the light of these we can view these both as digital/computational processes (albeit highly parallel ones). In the past evolution was often described in analog terms "gradual change, imperceptible differences...". I think this is rather misleading.
Firstly I should present two key differences between analog and digital processes.
1) Analog copying inevitably leads to increasing degrees of error until the original signal is completely lost. Try copying a cassette tape from a copy of a copy of a cassette tape and you'll see what I mean. Digital copying supports error checking and copies of copies of copies etc. remain almost entirely identical to the original.
2) When a digital signal is used as a program to direct some process (such as decoding a compressed film) errors in the signal can lead to much more complex distortions than errors in analog signals. Analog errors we are familiar with include the fly specks on film reels (which are pretty boring as mistakes go). In contrast an error in a mpg file can often result in trailing series of discoloured boxes or a distortion of the entire image. If fractaline compression were used for images one might expect even more interesting errors to crop up.
Our genetic code is entirely digital. Because of point (1) above it was recognised that this must be the case even before the role (and form) of DNA was discovered. For an interesting book predicting this (in advance of the Watson and Crick) see "What is life?" by Schrodinger.
Now for the new discoveries:
A) Two species both belonging to the same genus (which would normally be incapable of interbreeding) were crossed successfully once a crucial gene had been removed from one of the two species (the offspring were fertile).
This demonstrated that not only can speciation events occur by chance polyploidy events but also by the creation or loss of a single gene. This adds to a growing repertoire of sudden genetic events that can result in speciation.
B) Active DNA sequences have been found which occur not only in one gene but as part of many different genes. These sequences act very much like subroutines providing the same functionality in many places of your genome.
This show that your genetic blueprint is structurally similar to a software program. Such a structure makes clear evolutionary sense. If you have evolved one active site on an enzyme why not use the same active site for other enzymes which need to do related things.
C) Not all genes are expressed all the time. There are gene switches which can turn genes on/off (mainly during embryology). Often genes can be switched on or off by other genes. Because of this the gene switches form networks of interactive units. The main differences between species may well not lie in their genes (which are remarkably similar) but in these networks of gene switches.
These networks are like large networks of logic gates regulating gene expression.
D) Various evolutionary "throwbacks" have been observed including feat on whales, tails on humans, scar-less healing and antiviral cellular processes. Such throwbacks are difficult to understand if you regard evolution as an analog process.
The one thing that all of these examples have in common is that they haven't been present in the adult form of the animal for millions of years (in a few cases much longer). So the question has to be "How have the genes to create these processes or structures been maintained for this period of time?".
It appears that humans have the capacity to heal without scars but this capacity is rarely utilised because in most parts of the skin a scaring process gets there first. Foetuses possess scar-less healing as does the mouth (in good conditions). There is scientific speculation that if wounds are kept free of infection and the scaring process is switched off that wounds would then heal in a scar-less manner.
My hypothesis is that these throwbacks are all produced by gene programs that have been built on top of (and hence hidden from view) by a new tier of genetic programs. When mutations occur which disrupt this higher tier the older programs get a chance to function in their original manner once again. This is evolutionary feasible because changing the original genetic program may often be too dangerous as it would disrupt anything that had been built on top of it. So tails don't grow on people because the program is terminated early or interrupted by another gene program. However, as the original set of genes is still there it is quite possible for a mutation in the interrupting subroutine to allow the original tail building program to run. The set of genes that produce the tail might be needed elsewhere (perhaps in the growth of the spine) so evolution has not removed them.
This concludes my collection of evidence towards a digital understanding of evolution.
I think the analog metaphor for evolution was chosen originally because serial digital processes are inherently brittle and do not respond well to error. But analog processes are prone to build up of error and are incapable of computational behaviour in practise. Thus analog processes cannot account for much of the most interesting qualities of evolution and embryology. It seems that parallel digital processes are crucial to understanding biological processes and probably useful in genetic algorithms research too.
I will leave you with some possibilities highlighting the significance of computational evolution:
Z) Barriers between distantly related species may be breached by careful study of the interaction between the genomes of those species. Where species have different numbers of chromosomes artificial intermediates may first need to be formed which share the same chromosomes (but different genes).
Y) It may be possible to trigger genetic processes which have not occurred for geological time spans at both the morphological level and the biochemical level. This has great medical potential including (but not limited to) stimulating the regrowth of limbs.
X) Much information about the past form of an organism's biological ancestors may still be present in their genomes today.
W) Advances in embryology may follow from switching on and off various gene programs allowing us to observe echoes of how embryological development used to proceed.