Over at Colorblind Chemistry, I came across a quote from Fritz Haber, writing about his thesis work:
The thesis is miserable. One and a half years of new substances prepared like baker’s bread rolls… and in addition, lots of negative results just where I was looking for significant results, and further, results that I cannot even publish because I fear that a competent chemist will find them and prove to me that the camel is missing its humps. One learns to be modest.
Now, Haber was definitely someone to take seriously. He's showing up in "The Chemistry Book", for sure, both for his historic ammonia process and his work in chemical warfare. He was a good enough chemist to know that his doctoral work was not all that great, although he seems to have followed my own recommended path to get that degree as soon as is consistent with honor and not making enemies.
The post's author, MB, wonders what this says about organic synthesis in general. How much of it is just baking bread rolls, and how bad is that? My own take is that the sort of think that Haber was regretting is the lowest form of synthesis. We've all seen the sorts of papers - here is a heterocyclic core, of no particular interest that anyone has ever been able to show. Here it has an amine. Here are twenty-five amides of that amine. Here is our paper telling you about them. Part fourteen in a series. In six months, the sulfonamides. This sort of things gets published, when it does, in the lowest tiers of the journals, and rightly so. There's nothing wrong with it (well, not usually, although this stuff isn't always the most careful work in the world). But there's nothing right with it either. It's reference data. Someone, someday, might stumble into this area of chemical space again, and when they do, they'll find a name scratched onto the wall and below it, a yellowing pile of old spectral data.
I've wondered before about what to do with those sorts of papers. There are so many compounds in the world of organic chemistry that the marginal utility of describing new random ones, while clearly not zero, is very, very close to it, especially if they're not directed towards any known use other than to make a manuscript. So if this is what's meant by baking rolls, then it's not too useful.
But I'm a medicinal chemist. When I start working on a new hit structure, I will most likely turn around and put the biggest pan of bread rolls into the biggest oven I can find. This, though, is chemistry with a purpose - there's some activity that I'm seeking, and if cranking out compounds is the best and/or fastest way to move in on it, then crank away. I'm not going to turn that blast of analogs into a paper; most (maybe all) of them will be tested, found wanting, and make their way into our compound archives. Their marginal utility is pretty low, too, given the numbers of compounds already in there, but it's still by far the best thing to do with them. Any that show activity, though, will get more attention.
I really don't mind that aspect of the synthesis I do. Setting up a row of easy reactions is actually kind of pleasant, because I know that (1) they're likely to work, and (2) they're going to tell me something I really want to know after I send them off for testing. Maybe they aren't bread rolls after all - they're bricks, and I can just possibly build something from them.