[Martin Taylor 2017.11.07.09.44]
Sure, Martin, but if one bypasses mathematical correctness, ignores the rules of evidence, and fails at logical reasoning, no matter how one BELIEVES (!!) "PCT to be essentially not only correct, but necessary", what is said and done is not science anymore �not eveen philosophy� but some sort of pseudo-scientific religious email
jousting which, honestly, is what CSGnet looks like most of the time.
I agree with all that, bur the reason I wrote my "Why PCT" message at that particular moment was that you seemed to be conflating CSGnet with PCT, and throwing out the baby with the bathwater. Since I know of no other forum for discussions of PCT with others who take it seriously, my feeling is that the best approach is to try to oppose the nonsense and offer some science. There are others who do the same, and your sane voice would be a welcome addition, not just in opposition but also in support of science.
PCT needs serious examination, in a way that in other domains called "stress testing". Your power-law problem is one such test. Kent McClelland's sociological enquiries are another. Trying to explain the appeal of Donald Trump could be a third. Maybe these tests are too complex, in the way that to derive a molecular structure from quantum mechanical first principles was too complex until supercomputers became super enough to allow a few simple molecules to be computed, and compared with observation that demonstrated the accuracy of the theory. If they are too complex for today's PCT, is the problem inadequate computing resources or inadequate theory? We won't know until we compute and get the wrong answer (or the right answer).
Maybe what is needed is not more basic compute power, but data from experiments on sub-problems. This latter was the approach Powers preferred and for which he often argued. My preference is a blend of both approaches, understanding that there will always be some problems too complex for first-principles analysis, as there are in chemistry and biochemistry.
How to fix CSGnet is one such stress-test for PCT. If PCT is accurate, it should provide a base for developing techniques to fix the problems you correctly describe. To change the metaphor, if CSGnet is sick, PCT is not the disease (or so I believe).
Martin
···
On 2017/11/7 4:16 AM, Alex Gomez-Marin wrote: