"There's debugging -- always more than I'd like to see. This week, I've been implementing some new molecular descriptors, which have required the implementation in software of a new method for estimated LogP." Once code gets debugged, it has to pass the test suite, and new tests have to be written for anything created from scratch. He explains, "It's these tests that give us confidence that we can keep this stuff working, and port it from Linux to SGI -- or even to Mac, to keep it usable as hardware requirements change....
In pretty much all of my jobs, I've had a customer base of one or more modelers to satisfy -- Ph.D-level chemist who use the software I write. I've also been able to interact with 'real' chemists who have a comptuational leaning -- physical organic or medicinal chemist types. I've usually interacted with or led scientific or computer-science developers, usually with tightly scheduled projects. I've worked with sales and support scientist, salesmen, and marketeres and [I] spend a lot of time with sore feet in booths at American Chemical Society and other conferences."Joe's path after his Ph.D. was pretty interesting; starting with a industrial post-doc in Idaho, he then moved to Aberdeen Proving Ground in Maryland (working for the Army as a civilian) and then a series of medium-sized companies (Marion Laboratories, Wavefunction, Tripos). How did he land his position at Tripos?
"He knew several people from Tripos (a drug-discovery software company) from tradeshows and conferences, and several of them were familiar with Spartan and the work Joe had done there. When the right job became available, they thought of Joe."His advice to aspiring computational chemists? "Learning how modeling tools are applied", "learn[ing] to write," "learning to present oneself", "[not being] too fond of particular technologies, languages and so on. The bottom lines is that modelers have to help projects succeed." Oh, and one more thing:
"Given the Web, I think young folks should create 'chemistry blogs', particularly those discussing correct coding and development work. The only negative aspect of blogs is that it's way, way, too easy to read and write garbage -- it's all unedited. Blogs from 'somebody' tend to be good reading, but I'm not sure how to correctly transition from random person to 'somebody."I, uh, er, resemble that remark.
*As always, CJ's copy of the book helpfully provided by the author, Dr. Lisa Balbes.