The lab does some computational work to complement all the experimental work we do. However, the prevalent view on computational chemistry seems to be:
- Put structure (or atomic coordinates, what have you) in a black box
- Follow instructions printed on a sheet
- ???
- Profit!
Anyway, given that this seems to be the case, when (one of) my supervisor(s) runs calculations with me, *I* watch like a hawk* (a role reversal, there!). Recently we ran into some problems, but I don't think my supervisor even really noticed (or cared).
Unlike the previous troubles I've had with calculations, this problem actually has a really simple solution. Unfortunately, it's a brute-force solution. Thus, I don't have much motivation to fix it. Nobody likes boring problems.
Somewhat exacerbating the issue is the fact that the calculation still terminates. It's just that the results are all wrong. I don't know how my supervisor does it - he somehow blithely ignores all the error messages and forges gaily on. (Seriously - why does he even read the error messages if he's going to ignore them?)
Since my supervisor doesn't care, I should fix the problem on my own time**. But it's so boring! But if I don't fix it, nobody would! I need motivation.
Sometimes I find myself missing The-Program-That-Shall-Not-Be-Named. There's more room to play with, even if I don't understand what I'm doing sometimes.
Tangentially, is there a good book or review paper on things like comparing choice of basis sets and computational methods? I never really understand why people pick what they pick for their calculations. Sure, I can read the original papers, but y'know, I'm lazy.
On another tangent -
The crystallographer, today: "Your crystal was shitty. Grow me a better one, then come back." *Sad* (Actually, the crystallographer is really nice, but there's no going around a crappy crystal.) For my particular material, getting a polycrystalline solid is trivial; growing *nice* single crystals is the harder part. Sigh.
*The program runs in a Unix environment; it's so easy to make mistakes with the command line, it's not funny. And there's no undo button.
**I have now acquired enough information to access my supervisor's account remotely (yes, yes, legitimately). *Evil smile*
3 comments:
If I had a nickel for every time I heard that from a crystallographer...
Actually, Sean the xtal god usually ran them anyway, and he was able to solve structures for some that totally sucked (fibrous needles! hideous). Failure was always a result of problems with the crystal--too bendy, too small to mount, no diffraction, no long-range order in one direction...
Are you going for your first structure? :)
In my case, the bulk structure is well-known, but I added stuff, so I'm looking for the positions of what I've added.
A crystal can be bendy?
If it's very thin, yes. But the bendy things I submitted were more along the lines of very deceptive, shiny aggregates. They looked just like crystals! Weren't, though.
Post a Comment