"How do we distinguish a sufficently complex adaptation from a non-complex one for the purposes of your criteria?"
Most adaptations are complex. There are a few that aren't, mostly that only require changes within one gene of only a few amino acids.
"what qualifies as a random mutation?"
Anything that is not organism-directed (either the organism itself or other organisms).
"Does retrovirus infection count?"
Nope.
"UV damage?"
Yes.
"Changes in ploidy?"
This has not been sufficiently determined. I would lean towards "no".
"Or only errors in replication?"
Interestingly, I'm pretty sure much of what is considered "errors in replication" are processes we simply don't understand.
"Normal sexual recombination is random to a certain extent."
Are you sure about that? I would ssay that the only thing "random" about it is that we don't understand the process.
So you're defining a complex mutation by the number of base pairs that change?
How many DNA base pairs would need to change to constitute a complex adaptation as compared to a non-complex one? Does it matter where on the genome the DNA changes, considering that tweaking developmental genes leads to exponentially different characteristics whereas changing base pairs in other parts of the genome may do nothing at all?
On top of that, if you're going to define a complex mutation by the number of base pairs involved (since you effectively disregard recombination as nonrandom) and you want to see this scale of viable mutation in the last hundred years in order to demonstrate evolution in action - then you're actually asking for evidence against evolutionary theory, which indicates that random mutation causes large changes over millions of years, not hundreds.
Interestingly, I'm pretty sure much of what is considered "errors in replication" are processes we simply don't understand.
No. Errors in replication are a fundamental part of the chemistry involved in DNA polymerase activity. We understand it very well.
"Normal sexual recombination is random to a certain extent." ; Are you sure about that? I would ssay that the only thing "random" about it is that we don't understand the process.
Then you would be wrong. We understand the process very well. Segregation of chromosomes at meiosis during gamete production is a random process. It's a coin flip whether the sperm gets the maternal chromosome or the paternal one. Even genes on the same chromosome are randomly recombined due to crossing over.