Shelly Glashow was there in Paris and I had a lot of conversations with him. He was a student of Schwinger's, and partly under Schwinger's influence he suggested models of the weak interaction in which the Yang-Mills theory was applied somehow to the intermediate boson - the positive and negative intermediate bosons, the photon and one other particle, which would be a neutral intermediate boson. How that was to be reconciled with the strangeness changing rules was not so clear, but it was a very interesting idea otherwise. And I formulated it I think a little better than he did, although it was his idea. And when I returned from Europe and Africa in the fall of '60, I gave a report on his idea to the Rochester Conference - which was I believe was in Rochester. And I described it a Yang-Mills theory based on the usual group SU(2) or whatever its called, and a singlet which we would call U1; so it was an SU(2) times U1 theory. And I just described exactly how it would, how it would work. But Shelly... But Shelly had included some wrong stuff. He had said that this type of theory was renormalizable, with brute force mass, and of course that wasn't true, as Salam and Kumar had shown. But he'd already done SU(2) cross U1 by then, by that Rochester meeting? Oh, before that, even. He had written a dissertation perhaps, and anyway a long paper including it and including the mass - the symmetry break and claiming it was renormalizable, which it wasn't. But otherwise it was clever and it was essentially correct. Because Schwinger had done SU(2) prior to that, right? Because he was looking for real unification as distinct from... Yeah, but that didn't work, of course. No, it didn't work, no, absolutely. That didn't work. Also he tried to deform it into a scalar and tensor theory somehow, because at that time some people, most people thought the weak interaction was scalar and tensor. That didn't work either. But anyway, Shelly and I talked a lot. And we talked also about a Yang-Mills theory for the strong interaction and what the charges could be, and whether these charges would then be the same as the ones that would give super multiplets for classifying the hadrons beyond isotopic spin, and if so, what would that system be. Well, I worked out what it would have to be to generalise - the idea was to generalise Yang and Mills', Yang's and Mills' work. What algebraic system could replace isotopic spin? Well, I wrote down the rules for Yang-Mills theory and that would permit a generalised Yang-Mills theory, and I found what the rules were. I had charges that I called F, and what the commutator of Fi with Fj had to be - you had to be able to write it in a suitable representation as little 'i' the square root of -1, times little 'fijk', times capital 'FK' summed over K, where little 'fijk' had have a way of doing it so that little 'fijk' was totally anti-symmetric. And the Fi's were remission. Well, that's certainly true of isotopic spin where little 'fijk' is 'eijk', the totally anti-symmetric tensor with indices running from one to three. But what was the generalisation of this?