Hacker News Reader

Comment by cschmidt | Hacker News Reader

Boy, all the articles about Ph.D.'s on HN are so negative. I had a great time getting mine. I got a Ph.D. in Chemical Engineering from Carnegie Mellon. I guess it must depend a lot on the subject you're studying. We mostly did classes for the first two years. We got an adviser in the second year, and gradually ramped up the research. I finished in 4 years and 3 months. Pretty much everyone in the program would finish in 4-4.5 years. Everyone worked very closely with their adviser, and didn't have the "drift" I read about a lot, where students seem to feel alone. It was do 3 papers, write your thesis, and get out. I really enjoyed the chance to do research, and do a really deep dive on a subject. They funded everyone, enough to live to a "grad student" standard. I'd do it again in a heartbeat.

View on HN

Note that there is a huge difference between the UK PhD and the US PhD process.

In the UK you are left to your own devices for 3 years, and only find out if the external examiner even approves of your research questions in the one and only final verbal exam. It is up to you alone to ensure the research is valid and makes a contribution.

US PhD candidates have a lot more guidance and handholding that includes 4 years of structured training, followed by the presentation of a research plan that is signed off by examiners before any research is begun so there are no surprises in the final examination.

4 years of structured training, followed by the presentation of a research plan that is signed off by examiners before any research is begun so there are no surprises in the final examination.

And ideally, this plan has been discussed at length with the advisor before the formal proposal happens.

Thanks for pointing this out.

Just a few paragraphs into the article, I was thinking the author had done a spectacularly bad job of picking an adviser (or supervisor). In my experience, "meaningful intermediate goals" and feedback are the reason for having an adviser.

Not to say that all advisers are very good at providing goals and feedback, though.

At good uk universities you get excellent guidance, with multiple intermittent milestones and ongoing scrutinisation of your progress. Unless you pick a rubbish supervisor who simply doesn't care. Though I only can speak of Oxford..

It's also important to note that the supervising culture varies a lot between subjects. Computer scientists are usually spoiled by good supervision, but in the humanities you might see your supervisor only twice a year.. (and I do know people in Oxford in that situation).

I think it's also worth saying that phrasing it as "picking a rubbish superviser" puts an unfair blame on the student. By the time you realize your advisor is neglectful, or incompetent as a teacher, or actually neglectful, you might be a year or more down a research path. There might not be another professor at your school qualified and willing to supervise your research. The choice at that point, whether to tough it out or to throw away all your progress up to that point, is an incredibly difficult one.

(FWIW, my advisor in theoretical physics was merely neglectful. Almost all of my friend in an experimental field had advisors who bordered on abusive.)

This is a failing of your supervisor and university if you have this experience in a UK university. I'm currently supervising PhD students at De Montfort University and we are required by the university to have AT LEAST monthly meetings with our students (usually 2-3 supervisors per student too). We support our students with seminar and theoretical workshop series, and encourage and support our particularly promising students to produce papers and submit them to conferences and journals.

There is a LOT of hand holding that goes on in our PhD programme. But your mileage may vary. I've heard that Russell Group universities can be a particularly mixed bag in this regard, but academics at other universities without such big names have more to gain by supporting our students.

I believe what you're referring to is this, taken from the University of York Computer Science PhD pages:

"Please note also that all Research students embarking on a PhD programme are initially enrolled provisionally for that qualification. It is the responsibility of the student’s Thesis Advisory Panel to recommend whether or not the student’s enrolment for the PhD should be confirmed (consideration & confirmation takes place within 2 years for full time students and four years for part time students). Confirmations are approved by the University's Standing Committee on Assessment. For more information please contact"

Which, standing on it's own, sounds a bit frightening, but when read in context, you see that students are not left to their own devices or 2-4 years:

http://www.cs.york.ac.uk/postgraduate/research-degrees/phd/

I did my PhD in Comp.Sci. at the University of Liverpool. It lasted 3.8 months (as I took more time to write my thesis in English). At the end of each year my progress was reviewed in several way (a 1 on 1 interview with a thesis advisor, a departmental presentation, etc).

I had the best supervisors I could ask for (yeah I got two of them). Every week (first year) or every two weeks (second year) I met with them to discuss about my progress. Shit, they even helped me on the "culture shock" of a Mexican living in the UK (and coping with the heavy NW accent).

Moreover, one of my supervisors invited me to join a EPSRC project which led me to work (in my own chosen subject) with great scholars from different parts of the UK ( and form people from the industry.

Although I have a great experience during my PhD and in the 3 years postdoc I did in Germany, I decided academia was not for me. I got tired of the "publish or perish" part , and I decided to pursue my real passion: programming software.

I think it is highly dependent on field. My field, biology, is basically like flipping a coin. Do you graduate in five years or nine? Will you come out with a golden road to tenure or a life destined to work sad, lonely, post-docs? Flip that coin!

No matter how smart you are, the Gods of Biology are fickle. You may end up with a shitty protein, an assay that only works when the moon is aligned with Venus or cells that only grow when you swirl counter-clockwise.

As much as people like to pretend Biology is science...the sad truth is that a lot is practically voodoo incantations to the protocol devised 15 years ago by a post-doc.

It's one of the main reasons I left. You have to be Smart, Persistent AND Lucky. I'm fine with the first two, but when my career is hinged on luck...well, I started exploring fields that didn't quite as much require blind faith in the universe.

I wish there was some way to fix the situation with protocols in biology. Maybe we could attempt to identify the reasons that they suck, and then fix those reasons. There was a stackexchange site for protocols called methodmint, but nobody used that. Biologists either hate the internet or they hate computers because all of the attempts to get informal web communication happening have failed (even by email). It complicates things even more.

I don't think it's either of those things - biologists and biochemists just seem to be ridiculously busy, in part because of spending to long in the lab trying to get your crap to work. Every lab I've been in has an inefficient process by which people share, modify and troubleshoot their protocols amongst themselves - but that's convenient because you can ask the guy on the bench next to you. How to get researchers to invest more time and effort, communicating with people they don't explicitly have to, is the problem.

Dont forget that each machine you use is slightly different, so there are protocols that exist for those labs and those particular pieces of gear set up that way.

I did my PhD in ECE specifically computer vision and had a decent time, but pretty much everyone I encountered during their PhD in biology had some horror story to tell.

It was do 3 papers, write your thesis, and get out.

I'm a CS grad student at Technion, and that's basically how my advisor has described our grad-school process.

Paper 1: your MSc thesis.

Papers 2 and 3: If related to paper 1, combine into your PhD thesis.

3 related papers on one broad topic => thesis, you're done.

http://matt.might.net/articles/phd-school-in-pictures/

There's one important issue with this picture

You don't have the bigger picture. In fact sometimes you don't even know that a person in another country (or why not, maybe in another university in the same city) did what you were trying to do, or already knows something you don't.

Not to mention the Feuds, aah the feuds. Like something showing up in IEEE but in ACM it's gone and forgotten (or the other way round)

Well yeah, this is why we have a couple of things:

* Literature reviews. Check the published literature before thinking you've got an original research proposal.

* Conference/journal reviewers. If something's unoriginal in your paper, they will reject you in an instant.

Of course. But even with this sometimes it's hard

And reviewers sometimes skip things.

This is a very good reason for open-access, searchable publications.

Completely agreed.

Haha, funny there's another chemical engineer on HN. Maybe our major is just a good one for grad school experiences.

I think you make a great point--the thing is that is really depends on the program, and students should meet students currently in the program, chat with them, and see how things are for them.

Reply on news.ycombinator.com