QuoteYes, he has contradicted himself on a number of occasions, and he is not very measured in what he says. What conclusions do you take from this?
That he is a wannabee inventor who could easily be self-deluded, and that he is (in the normal meaning of the word, given the context that he wants people to give him money and accepts this when offered) dishonest. Perhaps he does not think he is dishonest.
The key indicator is his reaction to challenge. He never justifies what is challenged, and always distances himself from the challenger whom he calls a snake. The resulting fuss obscures the original issue that was challenged. It is a classic and well-known pattern, and one that usually indicates an overstatement of achievement.
QuoteWhen a scientist makes a claim of such-and-such, I don't think it's necessary to establish a strict chain-of-custody, like one would at a crime scene. The scientist mails something to another scientist, saying, "this is the control and this is from the live run; will you please take a look at it?" This seems quite acceptable to me.
Not when the stakes are high. For example, any scientist making an extraordinary claim might have some experiment that showed this. If the only evidence came from this one experiment (and the scientists claims) and a Nobel or more rested on the result, chain of custody would be essential.
Perhaps what you observe is that scientists are generally polite. They don't accuse colleagues of fraud in this situation. They say: "we think there is probably some error, wait for the replication". If the original experiment proves unreplicable (with integrity) even by the original claimant, and there is no other merit to the results - like some interesting theory that links it to other definitely observed phenomena, they are quietly forgotten. The scientist will not be accused of fraud even if it seems pretty obvious that he concocted false results. But his results will be ignored.
There is good reason for this restraint. It is difficult enough knowing what is good science without adding issues of character and cheating to it. Far better to view all cheating as just "error". Practically this makes little difference since experimental errors are always possible and in fact usual when the claims made are exceptional. Trying to work out which claims are dishonest is an extra burden that inflames feelings but does not progress the science.