Over at BoingBoing, they're investigating the question: "How long would your PhD have taken if everything worked the first time?" I have to admit, it took me a few minutes to adjust my head to that idea, since God knows, nothing in my PhD ever looked like working the first time.
And it's a hard one to answer, because I had to do some backtracking, as so often happens in total synthesis. This was of the "Dang it all, turns out I can't install that carbon at that step, so I'm going to have to go back, put it in earlier, and hope the downstream stuff still works" variety. (Not all of it did, of course). So how do you account for tactical moves like that? There are several layers.
How long would it have taken if I'd chosen the right move each time, and each reaction worked on the first shot? Even then, that's a tricky one, because one typically runs things on a test scale and then on larger amounts as the ground firms up beneath you. So if things had worked every time, just fine, and I'd scaled up as soon as they did each time. . I'd say around a year. Maybe even nine months; it's hard to say, because the concept of everything working is so alien.
Then one could ask, how long would it take to run through the chemistry in your dissertation, straight through, knowing what there is to know about it? In that case, it would be shorter. Just flogging away at the procedures, nonstop, and having nothing go wrong along the way (hah!), I think you could beat through everything in mine in two or three months. Boy howdy, would I hate doing that.
What does that leave out, then, of a degree that took me four and a half years? (A flippin' short span, I might add, considering some of the other degrees coming out of my old group). Well, there are all those false starts down synthetic routes that ended up painting me into corners. Being carbohydrate-based synthesis, many of those were protecting group problems, but there were a couple of rip-the-whole-sequence-up episodes, too, when things just wouldn't go any further. And there were things like finding out that a base camp of material I'd stored in the freezer had gone to hell anyway, in the dark, under argon. And realizing that a TBDMS group had up and migrated on me, such annoyances as that, which also involve proving that it happened and making sure that I knew where everything was still attached.
And there's an awful lot of time spent just getting each reaction to work - six or eight or ten ways to bring in a methyl group. Four or five different reduction conditions. All those choices, every time: borane THF or borane-dimethylsulfide? Swern or PCC? Hydrogenation catalysts, Lewis acids, finding out that switching from BuLi to KHMDS when making methylene Wittig reagent changed the yield of alkene from 10% to 90%. Chip, chip, chip, at every step along the way. At the time, it seemed as if my legs were mired in not-so-fresh concrete, three feet deep across the lab. Looking back, though, I think I must have been flying. . .