I think you're probably right that asking a specific (technical) question about the paper would be for the best...Care to take a crack at it, Professor G_2?
]]>Again, if the question is changed from "Is this particular claimed proof that S^6 admits a complex structure correct?" to something like "What are the key ideas of this particular paper [that claims that S^6 admits a complex structure]?" then the combative aspect is lost. I do think that some kind of analysis of one's work is exactly what you want when you put papers on the arxiv. It is possible that the outcome of this will be pointing out a mistake, which should then be done in a very polite and cautious manner. But I don't see that it's impossible to have a constructive discussion of the paper. In this case, I am biased in that I cannot participate in such a discussion but I would be very interested to follow it.
About closing the question as a duplicate: sure, that's reasonable. If one does do that, I think it might be reasonable to modify the earlier question to make explicit mention to this preprint, for otherwise everything moves a level down: the "Hey, what about this paper?" becomes an answer and then responses to that become comments, which would be very tedious.
It strikes me just now that a reasonable idea would be to contact the author of the paper and gauge his opinion on public discussion of the paper. If it's okay with him, it seems like it has to be okay, and probably the converse is also true, no?
]]>That said, this always going to be a bit case by case; in this case, I'm actually tempted to close as a duplicate.
]]>