NB This is NOT generic advice. It only applies to my Ph.D. students and not others - everyone has their own style of supervising, and some may disagree with aspects of mine. So, this advice is specifically for you...

Ben's Ph.D. Students

Ideally, you will get your Ph.D. in your third year. There are some things you can apply for, (before 21st Sept of the penultimate academic year of your PhD) eg the KITP graduate fellowship programme should you fancy a 6 month (financed) stay in Santa Barbara, US

You'll typically be meeting me once a week during our first year, unless I'm away. This is so we can catch up, you can ask questions, and I can see how the work is progressing (or not). You should come to the Cambridge Pheno Working Group meeting. It's got an international reputation, you will absorb a lot of useful knowledge from listening to the conversations there, and you will benefit from developing experimental contacts/collaborators. 1030am in the conference room behind the coffee room of the Rutherford Building on Wednesdays. We show plots, discuss results and the reason why various students are stuck etc. We often all go to lunch after with the theorists there. There is a DAMTP pheno lunch outing to Churchill college just nearby every Friday lunchtime at 12:30 as well, which you should come to (you can bring sandwiches if you prefer). Get into the habit of writing up your research findings, or even stuff you're learning that will be relevant, in LaTeX. Then put it on the Pheno working group pages so we can discuss it during the meeting. It will force you to organise your thoughts, and will prove very useful to paste into your thesis and/or a paper later! If you have stuff for me to sign (eg expenses requests etc), this meeting is the time to bring it. Otherwise, you can usually catch me at coffee (although not on Wednesdays). If it is more convenient for you, you can arrange a visit with me at home if I happen to be working there (happens occasionally): 58 Hemingford Road, Cambridge, CB1 3BZ. Otherwise, just catch me another day at DAMTP.

You should come to the phenomenology seminars, even if you don't get much out of them to begin with. At the least, you'll start absorbing the jargon "by osmosis", and that is useful in itself. If you find yourself in an unintelligible talk, try to get something out of the initial discussion and the conclusions. Also, watch the way people give talks: you'll need to do it before your Ph.D. finishes, and technique is really important (more on this below). You should also pick and choose other talks to go to: for example the HEP/GR seminar, or one of the cosmology ones (keep an eye out for interesting particle physics ones at the Cavendish too). You can organise the advertised talks by using talks.cam.ac.uk.

Go to a part III course that you fancy - there's a bewildering array, and it's good for your education to go to one (a maximum of one) a term, if you can find something interesting or relevant. But on the whole they are good, and you may as well take advantage of having them on your doorstep while you still have time.

In Sept of your first year, it's usually a good idea to go to BUSSTEPP if you're a UK student - it's a good summer school, you'll meet everyone in your Ph.D. year, and the courses should be some good revision and some new stuff too. After that, you should be aiming to either go to a summer school or a conference/workshop every year. A couple, if funds can be raised. There are plenty of events organised by IPPP which are re levant for you (e.g. YETI) and I will generally be supportive, if it's somewhat relevant. You have to worry about funding: DAMTP usually will fund half of your travel, and you're encouraged to get the other half from college. The NATO Summer school at Cargese, the one in Corfu town in September or TASI in the US can also be nice options.

Keep a record of your "transferable skills": courses, teaching/tutoring, talks, helping out with open days etc etc. You'll be required to give evidence of this to bureaucrats at some stage. It's likely I'll ask you to take the problems classes for my Part III course (currently SFP). This is a good way of learning the material really deeply, and also getting paid decent money from the colleges in the process. Every two years, I organise the HEP display for the Open day. I'll ask you to write a poster on your research, and to man it for an hour of two on the day, chatting to members of the general public (along with me, staff and other students etc) about what particle physics is all about.

Papers

There is a golden rule: you must always ask my permission before submitting a paper to the arxiv or in a journal. This is important to avoid any potential research misconduct, but applies even to single author papers because we as supervisors must OK the quality.

Papers are written in the text programming language LaTeX. Get used from the start to writing your findings up in a LaTeX document: this will also help you later when you come to write your thesis: the document can form a basis for your introductory chapter, for instance. Your Ph.D's value, as far as I'm concerned, is in the papers that you write and get published. Ideally, we'll get one in our first year. If this happens, then you're on course to be finished in three, which is a fair way below the average in our field. If things are going really well in this way, I'm very happy for you to work on a project with someone else - in fact, I encourage you to do so. If you want to get post-docs, a publication with someone that's not your supervisor looks good (ideally, a sole author publication in your final year's the best - but that is very rare). Usually I'll get you to write the method/results type sections, and I'll provide the introduction and conclusions, which require a more broad overview (and therefore experience). Before you send anything in writing, please get into the habit of spell-checking it automatically - for example, in emacs the command is ispell (or flyspell). This will save my (increasingly short) time...

Writing to a professional level is a skill that many students have to learn during their PhD, and you should as well. The golden rule here for papers and your thesis is clarity. You want it to be clear and to the point. Extraneous words that do not add to the clarity should probably best be omitted. Side issues that don't add to the main thrust might be omitted. Other things such as writing style take a back seat. We will both write different sections of the paper, and it doesn't matter too much if the style is different in each. Mathematical notation should be consistent throughout the paper, though. You should think about writing style a bit. Before you start writing, ask to borrow my book "Economist Style Guide" for Economist journalists and read the first half or so: it has good advice on clear written communication, before becoming more Economist-magazine specific (you should ignore that bit). You don't have to use fancy language or ultra-technical terms: in fact, it may aid communication if you don't. The main thing is to treat effective writing as a skill, to think about it and improve it. There is a rough pro-forma to papers: in the introduction, introduce the topic motivating its study, then say what others have done (with references), then say what you're going to do and detail the different/new parts. After the introduction, you have some sections on the substance of the paper, before summary and conclusions. In plots, make sure your font sizes are large. This is because you will likely be using the plot giving a talk on a dodgy projector at some point. You will notice the frustrating tendency in other people's talks to not be able to read the axes on their plots because they are too small and the projector's resolution is too bad. So if you use large fonts, the plot will be good for several purposes and save you time. Fill the caption in underneath the plot in full. Some people get tempted to write "see caption to previous figure" or something here. Don't do this: it is lazy writing and will put the lazy readers off. It's better to repeat oneself under every plot so the information is there directly, in the place where you need it.

Programming/calculations

Are you doing any programming or long analytical calculations for your research? If you are, then be assured: it is wrong. Everything anyone programs or calculates is automatically wrong on the first attempt. You have to check it by putting test cases where you know what the answer should be. Constantly ask yourself: do your results look physically reasonable? Can you check any of the limits against someone else's independent program or calculation? When writing programs, only write small modular independent bits to do things, and run your numbers through each one to check it's doing what you think it is. Even when you think it is, it's still probably wrong. If you don't do things this way, you'll spend at least ten times as long to debug your program as it took to actually write. Test, test and test again - while writing, and at the end. What about limiting cases where you know the answer? Look at every plot critically, and think "does this make physical sense?" and "why does the plot look that way?". Try to find inconsistencies in it... this will help you route out bugs.

Talks

Talks are disproportionally important for your career. All of my post-doc positions were awarded on the basis of talks I gave. It's a handy skill in general: don't be terrified, it can be learned. Ideally, we'll get you to give a talk to our local, friendly phenomenology series in your second year about your research: please remind me if I forget. Practising first in the student-only series might be a good idea. The HEP group at DAMTP will ask you to give a shorter one in your second year anyway: if I forget, remind me. I have a few pointers that helped me to learn to give talks:

When writing, pick 3-5 main topics, or themes, for your talk. These should be listed on the first transparency (with your name and institution) - they should provide a map of the talk. Then, write each topic in turn. At the end in your summary, you should again reinforce the main topics. To pick them, ask "If someone is completely bamboozled in my talk but I want them to go away with just a few points in their heads, what should they be?" This "tell them what they're going to hear, tell them in detail then tell them what they've just heard" should help this to work. One long narrative is far too easy to get lost in/lose the will to breathe/forget what the original point was.
Motivation and context of your work (for example, in terms of other's) is important and should probably be near the start of your talk, so people know why they should listen.
Only put a small amount of words/equations on each slide, otherwise they get overloaded. You should give words enough to remind you what you're going to say - not verbatim what you're actually going to say (that's boring to listen to). Talking about figures etc is also good.
If you tend to suffer from the symptoms of nerves (butterflies, brain freeze, nausea or light headedness etc), eat sugary stuff 10 minutes before the start of your talk. The symptoms are similar to those of low blood sugar, and although you still might feel nervous, you can reduce the symptoms this way. In fact, you can use your nerves to give your talk energy. It's too easy to be dead-pan, and that can be boring for the audience.
Psyche yourself up just beforehand. Try to go somewhere (the back of the room?) where you can do a few subtle stretches and actually phsyically warm yourself up, to get yourself in the right frame of mind and the blood flowing. Ham it up. It'll help people to listen and stay engaged. Changing tone of voice, holding a stick and waving it at the projector screen, turning to address the audience sometimes and looking one of them in the eye are all good techniques.
Time your talk. It will annoy the audience if you go over time. Therefore don't: run through your talk a day or two before on your own in a room, timing it and actually going through the slides and saying the words etc. You may notice that things could be ordered or presented better, which you can make a note of. If you're worried about over running because things are going slow, skip some non essential slides. This has to be done on the fly, so you'll have to think on your feet. It's handy to have a few extra slides beyond the summary containing additional non essential interesting stuff, or things backing up some statements you made during the talk. It can be interesting to have material to answer questions that people often ask here too.
I like to encourage questions during the talk, which gives everyone a break, and improves the dynamic.
I'm no master at giving talks, however when I started, I was truly horrific. Now I find it a little hit and miss, which is still a vast improvement!
Ask for feedback about talks from your peers. If I'm in the audience, ask me - I'll try to give you constructive criticism, trying to find stuff which you might improve or try differently.
Typically, you'll have 5 minutes for questions at the end, even if you take questions during the talk.
What do you do if someone asks you a question you don't know the answer to? It really is OK (even for professors) for anyone to say "I don't know", or "I'll have to think more about that". You'll be surprised though: you're the world's expert on what you've just done - you're more likely to be tripped up on questions about the research context. But that's normal for young researchers, and everyone understands.

Watch this video on giving talks for some other good advice.

Taking Much of the Pain Out of Writing Up

I have the slightly lacsadaisical approach to writing up: presuming you have a few, electronically staple your papers together, paper over the cracks, and add an introduction and conclusions. The thesis is supposed to be on a unifying theme. But you'll see plenty entitled with several topics: eg "Neutrino masses and supersymmetric phenoemenology" etc. There are some nice LaTeX style files out there for theses. Get another student's thesis draft, delete all of their content and replace it with your own. To be honest, except in rare circumstances, it's unlikely that anyone else will read your thesis bar your examiners and perhaps future Ph.D. students of me. Like I say above, I think that the worth of your work is contained in your papers, and people will read and use those, rather than your thesis. My attitude is then that the thesis is a necessary piece of bureaucracy to get your doctorate. As such, you should do enough for the thesis to get it passed safely, but it's not worth finessing too much. On the other hand, theses can legitimately contain more detail than papers, as well as additional stuff that wasn't quite to-the-point enough to make it into the paper (but is nonetheless relevant). Also, put it on D Space after it's been accepted, for prosperity. You can reproduce your papers verbatim: one in a chapter. But you'll have a statement somewhere saying which pieces are your work/writing. Following this strategy, writing the thesis took me literally two weeks and was stress-free. If you want to rewrite the whole thing yourself, of course you can: it's your work and you should produce what you want to.

Outreach

I encourage you to do outreach. Aside from being a duty of ours (we're paid for by public money after all), it is important for getting a job: Universities and the STFC are very keen on people that do it. There's a maths OPEN DAY in March every two years, and you could get involved in twitter, talking to journalists etc etc. The university run media training courses, and periodically so do STFC/The Royal Society. If you get a chance to go on one, it's worth it!

Applying For Post-Docs

NB STFC post-docs have very early deadlines - in October. Unfortunately, you're too inexperienced to apply for CERN post-docs yet: you need to wait until you've got past your first job, at least. Aside from that, most places hire between December and January 7. This means that you should be preparing applications in October of the last year of your Ph.D, ready for starting work the following October. Obviously this relies on a realistic prospect of finishing your Ph.D. by then. You'll need a decent list of published papers: the more the merrier!

You need to apply to a good number of places: say 40, or more. Most/nearly all of these will result in rejections. You're playing a numbers game though, like all of the other applicants. It's normal to apply to places even when you're not sure whether they can offer post-docs or not. Often, funding arrives late. Send a CV (example - max 2 sides plus a list of 3 referees with addresses/phone numbers etc...get their permission to use them as referees first though!), a letter of application and a list of publications. Here's a list where the people know me, so that may (hopefully, anyway) help your chances to apply there. Here's a tarball of handy latex styles for multiple application letters and labels etc. Write directly to a relevant member of the faculty, unless the website of the group requires some centralised application. These are a pain, but are becoming more common, and may require you to fill in an online form, and/or send much of the aforementioned material electronically. Keep an eye on online jobs announcements - eg from the arxiv, to see if there are any jobs in your area coming up that catch your eye.

You must give your referees plenty of time to write the references. It's risky pissing off a referee by giving them a last minute deadline! If I write you a post-doc job reference, I'll give it to our HEP secretary in electronic form. You must then supply the secretary with a list of places you're applying, where the references should be sent, as well as URL's/email addresses of places where a reference should go. We'll probably have to discuss who else should be your referee. But it's handy if you've worked with some other researcher, to get their letter...

You should go round in the Michelmas term at the time you're applying, giving talks around the UK - it's a good way of getting your face known, and if you do a decent jobs, vastly boosts your chances of getting a post-doc. You've probably picked up from my tone that the first post-doc job is quite difficult to get, since you've not got many publications compared to foreign Ph.D. finishers (which often have more years of research). The UK is the best chance for you to get your first job: UK departments realise this difficulty, and are more lenient accordingly, to some degree.

Write and apply to faculty you've discussed with, or talked in front of, or that I know and collaborate with. These are all tricks to try to increase the chances of that job offer. I'm afraid that it can get a bit depressing when the rejections start rolling in. You just have to bear in mind that you're playing a numbers game, and that everyone makes many applications - so the average number of people applying for a post-doc is say 100 or so, all with Ph.Ds and a publications list.

I will take an active interest in how jobs work out for you. If you want advice with negotiations or which jobs to accept/when etc, I will be happy to discuss with you.

OFFICIAL homepage