Jamie's Blog

Lessons from a life of startups, coding, countryside, and kids

Things I learnt during, and about, my PhD

This post might be written in a tone that gives you the impression that this is advice about your PhD. It’s not. It’s overly pessimistic and specific to my situation. These are (largely unedited) notes I collected between 2003-2006 when I was working on my PhD in Computer Science. I eventually admitted to myself that I wasn’t going to finish it (due to a combination of research/financial/personal/career pressures) and that, in fact, I didn’t want to finish it. This post represents the advice I wish I could have given to myself when I was thinking about applying for a PhD. The short version of this advice is: Don’t.

The devil is in the details — and a PhD is Hell

I came to my PhD with a fairly clear idea about what I wanted to investigate. In fact, I’d been thinking about it and working on it for the previous 2 years. However, many people start their PhD with a vague interest in an area and spend the first 12 months figuring out what novel aspect they want to pursue. Initially, it can feel as if finding out what you are going to research is the main hurdle. Alas, that is not the case. The main problem is that once you’ve narrowed your research area down, you need to keep focusing. And again. And again. In the end you’re left looking at boring equations, graphs and theories that are the complete opposite of the interesting and practical idea you started with. In many ways this is like starting a business: everyone can have a grand business idea, some people can tease out a feasible business plan, but the successful businesses are run by the people with the big vision and the attention to the smallest details.

Know your audience

It is important to understand, from the outset of the PhD, who your target audience is: it’s you. I remember hearing that, on average, 1.6 people will read your PhD thesis. I’m pretty sure that includes yourself, your spouse, your supervisor, your second supervisor and your examiner (yeah, that’s technically 5 people. If someone says they’ve read your thesis, they’re probably lying — they read page 9). You have to accept, that no one in the world will want to wade through this document. Ever. You might start your PhD with the intention of making a discovery crucial to the future of the world and winning the Nobel prize before you’ve even graduated. You will be very disappointed. No one in the world will care about your work. Repeat after me: No one cares.

Supervisors: a curious species, rarely sighted in their expected habitat

Supervisors are strange creatures. Some are like ghosts, appearing occasionally for a fleeting moment, and you’re more likely to meet them at a conference than at the University. Others are always around but they’re too busy running around like demented hamsters on a wheel — all motion and no progress. They’re disorganised. All of them will, at some point, forget what your project is about — and some will even forget who you are. I made an interesting discovery half way through my PhD: the number of good/useful/interesting/brilliant things that your supervisor will say to you is not proportional to the amount of contact you have with them — it’s constant. Yep, that right. You can have weekly meetings with your supervisor but you’ll only get three good suggestions a year out of them. Oh, and on the subject of meetings, there are only about five types of meeting that you’ll ever have during your PhD: The Big Picture, The Progress Update, The Paper Writing Enslaving (a.k.a. My Research Review Is Approaching So I Need To Get You To Write Something), The Thesis Word Count and The Pub (usually accompanied by beer). Do not make the mistake of going into a meeting and expecting it to be any different to last week’s. And try not to get them confused: even if you supervisor is plying you with beer, watch out for the sudden switch to Paper Writing Enslaving. Supervisors also participate in a little-known game which can catch out the naive student: Hunt the Supervisor. This involves the PhD student attempting to locate their supervisor during the agreed meeting slot. And, no, they are definitely not going to be in their office. You’ll be lucky if they’re in the right country.

The loneliness of the PhD student

A PhD is a completely solo effort. There is no one you can ask advice of. No one can help you. There are no books in the bookshop that will shed light on your problems. Magazine articles are even more pointless. You are alone. Think Frodo without the Fellowship.

Even PhD Students can live a normal life. Sometimes.

Being a PhD student isn’t like being an undergraduate. There a very, very few lectures you have to attend and very few regular assignments. There are no grades either. There’s also no timetable. Essentially you can work (or more often, not) whenever you please. So, it’s not like having a proper job. Even if you work regular hours (say 9am-6pm), you’ll be reading papers, writing papers, running experiments and any number of other pointless things during your free time. I tried sticking to a regular working day and it didn’t work — or, rather, I didn’t. If you’re like me, when it gets down to the nitty-gritty, boring detailed work of the PhD, you need to remove as many distractions as you can because, at this stage, just about anything is going to be preferable to your PhD. Computer games, good fiction and the Internet are all obvious distractions that can be minimised. Washing up was one of my favourite distractions, which I never found a way to avoid. The stipends available to a PhD student are actually very good, especially if you’ve come directly from the pasta & baked bean-eating life of an undergraduate. On the other hand, if you’ve been working for a couple of years, the drop in disposable income is a fair shock. Still, it’s not that bad; I managed to get married during my PhD (something good had to come out of it!).

A PhD is great because… well, just because.

I attended this wonderful introductory course run by my department, in which they presented this slide of great points (completely unedited) about why you might want to do a PhD:

  • Qualification
  • Dr.
  • Curious
  • etc.

Yep, that’s right. These are the only reasons he could think of to do a PhD! According to this professor, you should do a PhD if: you need to to get that academic position you’ve always dreamt of since childhood; you want to show off your title on mortgage applications and get called to medical emergencies; you’re “curious” (just remember what happened to the cat); or some other, unspecified, and entirely unthinkable, reason.

I can actually think of a few more positive reasons. If you’ve just finished your degree and aren’t quite sure what to do next, then a PhD isn’t the worst thing in the world. You have plenty of time. If you want to start your own business then doing a PhD will give you access to the latest research results upon which to base your commercial enterprise. And undertaking a PhD requires many of the same skills as starting a new business: self-motivation, attention to detail, unhealthy work hours, forward thinking, pitching your vision to sceptics and laymen, etc. On the downside, for every job that a PhD will help you get, there are a thousand which it will over-qualify you for. For some jobs (admittedly, probably not your ideal job), you might be better to pretend you were in prison for those years of your life. In any case, a PhD is unlikely to get you a higher salary.

It’s not the Result, it’s the Road that counts

I discovered at the start of my final year that it doesn’t really matter what you produce as an end result of the PhD. What matters are the experiments, trials, results, observations and evaluation you conduct. This makes sense when you consider the PhD for what it is: a qualification to conduct individual research. Producing something interesting, useful, wonderful and absolutely cool is not part of your PhD. Get over it. For me, this made my final year an absolute nightmare of doing things I wasn’t interested in and, frankly, didn’t care about. I started the PhD as a way of shutting myself away from the world for 3 years whilst I worked on an interesting idea. I categorically and absolutely did not care about the qualification. I didn’t need it and I had no desire to work in academia. Unfortunately, after 2 years I had to accept that, unless I modified my approach, I actually wasn’t going to have anything to show for 3 years effort. And despite my initial intentions, I really couldn’t waste 3 years of my life without anything in return.

(Not) Getting Things Done

Like any large long-term project, you’re going to need to learn how to organise yourself. So you might start reading about various personal productivity methods: Getting Things Done, Gantt charts, to-do lists… but don’t bother, they don’t apply. If you do make a list, it’s basically going to come down to five types of tasks: Reading, Inventing, Comprehending, Implementing, Writing. Now, reading is an easy task to complete: just read the book/paper. Unfortunately, sometimes you don’t know what to read. Implementing (say, software) isn’t too difficulty either. Neither is writing. The problem is the two core tasks of your PhD: Inventing (an algorithm, for example) and Comprehending (how/why it works and explaining it to others, typically through graphs and experiments). Inventing something is an open-ended task which can’t be estimated or controlled. In management-speak, it’s highly risky. Comprehending is also quite difficult if you’re as mathematically-illiterate as me.

The Other Year

Myth: A PhD is 3 years long.

Actually, you’ll find that the PhD funding is for 3 years but the university is quite happy for you to take 4 years to complete. After 4 years, they get difficult because the funding organisations will fine them for every PhD student that hasn’t graduated. The solution is naturally to finish during the 3 years of funding. It isn’t going to happen (and if it does, you were destined for academia and shouldn’t be reading this — your fate is sealed). 3 years sounds like a long time but it isn’t. But what happens between the end of the funding and the completion of your PhD? I don’t know (yet) but starvation, alternate employment, debt and non-completion are all on the (credit) cards. It’s worth bearing in mind that the university may “strongly recommend” that you don’t get a job whilst trying to write up. This makes sense in some ways: those that take on jobs are less likely to actually complete — possibly because they’ve already got a job and don’t care any more. The students that don’t take job tend to get very skinny and tremble at the words “bank manager” and “credit card balance”. Of course, the university has their own motivations for wanting you to complete (namely avoiding that fine and looking after your academic interests). On the other hand, you will probably have a motivation to eat and remain part of society — which may outweigh your waning motivation for finishing the PhD. It’s also worth remembering that the university will still expect you to pay fees during this missing, unfunded year (although, reduced when you get a full draft thesis written).

The two most frequent PhD questions

There are really just two questions that you’ll be frequently required to answer:

What’s it about? Enjoy this phase as it only lasts for about 6 months. Once someone has asked the question, and listened to the largely incomprehensible drivel that you’ll reply with, they’re highly unlikely to ever ask again.

How’s it going? The true purpose of this question is revealed after about 2-2.5 years: what they really want to know is “When will you be finished?”. The subtext is that a PhD is something to finish, not something to do. Unlike the first question, this will be asked repeatedly by the same people, regardless of whatever negative, vague, dismissive, or generally cranky response you give them. I’m thinking of other responses including swearing, violent outbursts, “when I finally give up”, “when I finally admit that the algorithm (and general idea) is seriously flawed and I am not able to unflaw it”, “when I shrink to 4 foot nothing”, “when the earth’s magnetic field flips over”, and so on …

Your supervisor is here to stay

A university can be a challenging place for an undergraduate — plenty of them drop out or fail to make the grades… some don’t even get the grades to attend in the first place. For the postgrad, the university is a place of indifference: no one really cares whether you’re there or what you’re doing. It’s slightly more… stressful… as a postdoc since your employment is dependant on the whims of a funding organisation and the bullshitting presentation abilities of your manager. I think new, probationary lecturers might need to make some effort to impress but it soon fades after probation (or, frankly, before).

Academics, on the other hand, lead a fairly stress-free life. Of course, they run around like headless chickens, never have time to attend your meeting, are always behind on deadlines and generally exude an air of stress, mild panic and approaching heart attack. However, this is not the whole story. You see, academics don’t get fired. Undergrads drop out, post-grads give up, post-docs leave but academics… retire. Academia is a place where incompetence can hide from the rest of the world. There are some good academics who have mastered their art of reading, thinking, writing and teaching, but there’s also a good dose of incompetent ones too. In the commercial world, there’s always rumours, quiet whispering and cryptic emails about “Jake Smith no longer works for this company”, but it doesn’t happen in academia. I reckon that the only thing that would get an academic fired is if they slept with a student. One of their own students. Whom they gave top marks to. And everyone else failed. I don’t think many frustrated PhD students will consider sleeping with their supervisor just to get them fired. So, no matter how much you dislike them, or incompetent/abusive/annoying/offensive they are, be under no illusions that your supervisor is going to be there long after you leave.

Here’s what I wrote in response to a post (on the Business of Software forum at Joel on Software) from someone considering a part-time PhD:

I had 5 years commercial experience and then decided to start a PhD in Computer Science (full-time) because I wanted to get some ideas off my chest. I’m almost finished now (well, the funding is almost finished) and absolutely hate it. First, I’ve no idea how you think a PhD is possible part-time. It’s an every-minute, in-the-shower, at-the-weekend, in-the-middle-of-the-night sort of thing. Part time, it’s at least a 6 year commitment, which makes me shudder just thinking about it. I’ve written quite a lot of software during the course of my PhD but that doesn’t get you a PhD and no employer is going to care about this code because none of it is to commercial quality (it’s just enough to get things done). A PhD is not an extension of a Masters, or a super-super-BSc. It’s a qualification to conduct research. I’ll repeat that again because its important: it’s a qualification to conduct research. Therefore, the whole thing is not about actually solving a problem but the process you go through.

It’s about equations, not code; about graphs, not screenshots; about field trials and user experiences, not unit testing; it’s about quick-and-dirty, not smart and professional. In short, it’s about things that aren’t (often) required in the commercial world. If you’re not interested in academia, you probably have no business doing a PhD. I tried to convince myself that I wasn’t interested in the qualification, just the 3 years to work on my ideas. But after 2 years and 10 months, it’s very hard to walk away from the qualification. And you are on your own. There is no team, no one to work with. It’s particularly lonely and, in the end, no one will actually care about your research except yourself (and, by the end, you probably won’t care either). Of course, I’ll be earning less after the PhD then I did when I started it. There aren’t many reasons to be using J2EE (or some other in-demand tech) in a PhD: think Perl, Python, LaTeX, etc. That’s if I can even find a job now: it’s well known that a PhD opens a few employment doors… and slams a hundred more. Sorry to sound so depressing, but a lot of people here have been telling you to jump and I just thought I’d better present the negatives too.

Update: I’ve just read Seth Godin’s “The Dip” and I wish I’d read it during my PhD. It’s a short, easy-to-read book about the basic (and important) decision that you’ll eventually face during the tough times of your PhD: “Do I quit or do I stick it out?”. I can’t recommend this book enough for PhD students — it’ll focus your mind on actually making the choice. As Seth points out, winners quit all the time but they quit the right things… and they stick with others. The loser’s choice (and the one I took), is to not make that choice at all. To neither quit nor stick with it, meant I got the disadvantages of quitting (no PhD/failure) and of sticking with it (wasted time/effort/money) but the advantages of neither. Committing to that choice earlier is possibly my one regret.