[Frumious Bandersnatch]

I think there have been wholly unjustified literal comparisons between DNA and blueprints and DNA and computer programs.

It was your analogy, not mine.

In bacteria, survival is improved by replication errors. The analogy to blueprints is mostly invalid.

Not necessarily.  Firstly, we don't know if these are replication errors, or are a response to environmental stimulii that are within the bacteria's genetic coding.  Secondly, if these are indeed random mutations, they are more likely to cause problems than solve them, so this argues heavily against random chance.  Thirdly, the changes themselves have not been shown to cause a disparate species.

[js1138]

Not necessarily. Firstly, we don't know if these are replication errors, or are a response to environmental stimulii that are within the bacteria's genetic coding.

I don't believe this. For example, if mutation is a "designed" response to antibiotics, then the mutation rate would be very high in the presense of an antibiotic. After all, a given colony of bacteria is composed of nearily identical clones.

But the rate doesn't change. all that happens is that survivors increase in numbers because they are the only ones left standing.