I recently got my PhD in CS at a European university, and I could directly translate many of the points from chemistry to the CS field. The article is 20+ years old, yet it somehow seems universal even today. I did a LOT of overtime, and often "just" to make a decent work that I wouldn't be ashamed to put my name on. I also met other grad students who wasted most of their time at the office on trivia (lunch, coffee, newspapers), people who rarely talked to their supervisors, people only superficially interested in their work, as well as people who did not really like what they were doing, yet they were striving to do their best. During my PhD years I learned at least as much about myself, academia and different kinds of people that I'm going to meet later in life, as about CS. One of the most important lessons was learning how to handle people who treat their work as a 9-5 chore instead of something to excel in.
--
Long time ago, while I was still an undergrad student, I asked one of the professors about grad studies. We had a long talk, from which I remember two things. 1: your PhD topic does not determine the rest of your scientific career (he shifted his area of interests a number of times). 2: whatever you do, always have a pen and small notebook with you. The latter has proven to be a golden piece of advice.
A little Googling suggests that this was written by Paul G. Gassman, a professor of organic chemistry at University of Minnesota, who passed away in 1993.
This probably reads a little harsh, especially to the software hackers among you, but I can say that my thesis advisor espoused most of the same core ideas ideas, even if he would not have written them down in this form. Strip away the snarky comments about re-shelving his books, and this is the core of what is expected by most research advisors at top universities.
I thought this was an excellent article with some solid advice on how to succeed in the field. At the very least it's an honest description, up front, of what's expected so people can decide whether a career in the sciences is worth it.
That said, I'm also struck by the notion that unless you were particularly passionate about the idea of doing scientific research the best thing an intelligent person could do would be to pursue a career in a field where the rewards were higher or the sacrifices demanded lower.
As a Ph.D. Synthetic organic chemist who went on to postdoc in a top group at Harvard in the 80s, I can attest to the tone and content of these memos. Both Gassman and Meyers, though not giants in the field were very well respected researchers and were not famously hard ass. I'm surprised I hadn't seen these memos before but there wasn't anything there that I didn't know intuitively by my second year of grad school. The most profound advice there is that what your thesis advisor thinks of you is the most important thing you get from grad school. Like many fields, it is an old boy network and your future career is entirely dictated by recommendations. For some reason, many of my lab mates didn't get this and led horrible relationships with the boss. Even if you are a competent researcher, when your recommendation contains the phrase "he's a pain in the ass to work with", you are not going far in this line of work. I'm not saying to be a brown noser, but your ability to function well in a group environment is extremely important to employers.
Read the memos! Sending this to someone at Purdue.
"Thus, when you do not perform your research duties with diligence, dedication, and efficiency, is it any wonder why your research advisor seems to always climb all over you?"
"In effect, I am choosing where I think you should go in your professional career. This is an awesome task with considerable control over your own life and I accept this responsibility with the utmost of concern."
Political correctness and emphasis on public image may have created some measure of accountability over the last few years, but more than too often it means that what is actually going on is simply never said aloud. Instead it must be inferred and gathered second hand.
Reading this stark honesty is really rather quite pleasant and refreshing. It's nice to know what somebody actually thinks and what they're going to do instead of a washed out and contrived 'official statement' which requires severe interpretation to understand.
The professor mentions that 10 hours a week should be spent reading and, at a bare minimum, 6 or 7 particular journals should be read.
Is there an equivalent regular reading list for the computer sciences and/or software engineering? I'm aware of things like arXiv and a handful of journals but it can be hard to dig through the often extremely theoretical stuff to the things that better reflect useful modern and progressive thinking in these disciplines (though I admit I might be approaching it wrong).
(a machine learning phd student here) 10 hours a week reading seems about right for me, although often is a lot more or less than that. Usually most of my reading come from: google scholar searches on a topic of interest, following the references cited in a paper I'm trying to study, following the google scholar list of papers that cited an interesting paper I'm trying to gauge the influence of, and (a few times a year) some 10 to 15 papers that are freshly released from one of the top international conferences in machine learning or natural language processing that had interesting titles or abstracts. A very small amount of reading comes from my advisor or colleagues. I do follow some arxiv feeds about research in ML, but it's very rare to find good papers there that I would not find by some of those other means.
I'm not sure how much of this translates to software engineering, but I find it helpful to keep a loosely organized reading list with papers and topics I should browse or read more carefully. The important thing, I guess, is bootstrapping an initial reading list for you to expand. Search for top conferences (ACM, IEEE, etc) on SE and browse the titles/abstracts of papers. Also look for seminal papers on something that interests you, and see who cites it, find interesting abstract/titles, etc. Then find out where were these papers published and search for more papers there. Maybe keep a note of the good publishing spaces in the areas that interest you and check up on them once a year or so.
I've right now got around 20 papers/books open in my notebook, only 1 was got from arxiv, around 10 from conferences in the area and the rest from google scholar or references.
However, a note of warning: most papers I've on software engineering and horribly dull, uninteresting, and unscientific. Take a look, though, at experimental software engineering conferences, because they seem to be producing very interesting things and slightly improving the face of the field.
Thanks for the tips. I'm interested mostly in search and natural language, so there are quite a lot of papers coming through arXiv and Google Scholar. Thanks for the tips and leads, though, I'll give your techniques a try.
Take a look, though, at experimental software engineering conferences, because they seem to be producing very interesting things and slightly improving the face of the field.
Could you suggest a few "experimental software engineering conferences"? These sound like exactly the sort of thing I'd be interested in following, but a scour through the ACM's events list nor asking my Twitter followers didn't give me many good leads for events like that.
So is this supposed to be shocking? This is actually one of the reasons I've always had significantly less respect for the 5-year masters programs. Undergrads typically have neither the maturity nor the time to dedicate themselves properly to learning the material. As an undergrad I was assigned to a lab where RAs were expected to work 40 hours a week, and on Friday afternoon they were standing by the door looking at their watches.
I've always thought that an advanced degree (past a BS) requires a ton of dedication, and the goal should be to know a topic better than anyone else. But I guess the advance of non-thesis masters degrees gives the lie to my belief.
That one point about how "in effect I am choosing where I think you should go in your professional career" reads pretty harsh, but if you think about it most of us do that when giving a recommendation. If I get a call from an R&D shop, I would naturally think about different parameters and behavior than if I got a call from a sales department.
> So is this supposed to be shocking? This is actually one of the reasons I've always had significantly less respect for the 5-year masters programs.
I don't think it's supposed to be shocking. Both the title and comments sound pretty positive. <Freud> Perhaps you wish it were shocking because it would confirm your negative impression of modern research? </Freud>
My own experience is that there are "relax labs", where students (perhaps interested in research but not as a full academic career) grab a diploma before going on to employment, and "serious labs", where students go to become a professor or really develop an expertise. The first kind isn't necessarily unproductive in research, but they tend to work on more applied stuff - cool gadgets motivate these students more than deep theoretical work.
Should you read this - nah, I actually have a positive impression of modern (as well as older) research. I have to admit that I was surprised to see such a strong set of opinions presented by the article, and after some reflection, I was pleased as well. It's nice to see someone saying "You want to be good at this, you better work your ass off to show me that you want it." rather than the typical modern schoolchild approach which engenders the idea that everyone is good enough.
I think you're right about the different kinds of labs, though I would characterize it as more of a continuum, but it seems to me that the "relax labs" ought to be more corporate or at least corporate-sponsored. But that's a whole different topic.
> it seems to me that the "relax labs" ought to be more corporate or at least corporate-sponsored
Oh yes, I completely agree on that. I've worked at one of these places, and what they made was exactly the definition of "cool gadgets". To my eyes, it wasn't research, it was pure development. Stuff like this should be funded by a company to build a product, not by universities.
I can't go too much in the details, but one of the things they built was a blogging robot. It has a PDA and blogs your location for some reason...
I think much of what he wrote is good advice for doing a PhD. That being said,I would recommend avoiding an advisor with this kind of attitude if at all possible. Despite what he says, he sounds like many professors who focus on the reputation game, carefully weeding out any students that might adversely affect their reputation. In other words, education is not their priority.
As far as the RA memo, what the author failed to mention is that the grant-RA relationship is a two way street. PI's need students to do their research for them and they need students simply to receive grants. They are not supporting students just because they are nice, but rather because they need them.
--
Long time ago, while I was still an undergrad student, I asked one of the professors about grad studies. We had a long talk, from which I remember two things. 1: your PhD topic does not determine the rest of your scientific career (he shifted his area of interests a number of times). 2: whatever you do, always have a pen and small notebook with you. The latter has proven to be a golden piece of advice.