Virtual HLF 2020 – Dialogue: Robert Endre Tarjan/Donald Ervin Knuth
youtube.com
youtube.com
For what it's worth, most(?) of Knuth's videos are collected here: http://tug.org/interviews/#knuthav — if you find more, let the TUG people know. [Or just submit to HN mentioning "Knuth" and I'll probably find it and pass it on. :-)]
--------
Anyone knows what paper is referred to at 39:50?
> [Knuth] I remember last year, I got really angry when I heard about— I looked at a paper and I thought it was— it had solved a problem that I thought I would never see the solution of in my lifetime. It was a brilliant solution. But the problem involved a generalization of knight's tours and the guy got the most insulting referee reports when he submitted the article to a journal and I showed this— Essentially these high pure academics were saying “nobody would ever stoop to write a paper about something like the generalization of a knight's tour”… There is no accounting for taste and I can’t say that my taste is better than others’. I can look at papers and say this is trash and I can look at something else and I say this is beautiful. I have no algorithm that will— I don’t think machine learning will ever solve that problem.
> [Tarjan] That’s the flip side of the coin: I mean, great papers that were rejected by narrow-minded reviewers and the greatness only emerges later on perhaps.
> [Knuth] D. K. : Right. Plus the… So I finally found a good referee for the guy, but you know, he worked for a journal that… screws the universities by charging too much for this journal. So he didn’t want to publish in that journal. So he left it on the arXiv for the moment.
I do not have a lead on that paper. Will post if I find it.