nodalpoint.org - How do you work? - Comments

Echo my thoughts

Animesh — Tue, 11 Jul 2006 11:24:33 -0400

Pedro looks like we are similar in the way we work... Long time back I read somewhere, Chaos breeds Creativity ... I have been successful in creating former without breeding later... I am waiting!
Anyways I feel a researcher should definitely read the Hamming's article on this subject ( http://fuzzylife.org/animesh/hammingOnResearch.pdf ).

Work ?

PedroBeltrao — Mon, 10 Jul 2006 11:12:09 -0400

I come in around 10:30-11:00 that is considered late for most of the people I know in the institute. I mostly go through the blogs/emails/content alerts in the morning. One of my problems is that there is always too much I want to read. I actually really only start working after lunch and my productivity rises to a peek at around 5 to 7pm. I listen to music all day long with headphones, it helps me to focus by removing the outside world :). I do a lot small breaks during the day even if just to stretch my legs for a while.
I have very poor code practises. I think Perl (and my laziness) are probably to blame. Most of the time I am cutting and pasting code instead of doing proper reusable packages. I use editplus to code and I do everything in windows (yes it is possible :).
I am lucky that in my lab most people do a bit of both (wet lab and programming) so most people in the lab and in the institute in general understand the usefulness of proper statistical/computational analysis.
More generally I find it really hard to keep the same productivity over long periods of time (in days). I tend to work at a snail pace with several projects (2-3 projects) at the same time until something really interests me, then I drop everything and work on that project in a very focused mode. This part is usually the part I love, getting consumed for one or two months discovering something.

constant self-appraisal

Neil — Sun, 09 Jul 2006 03:28:14 -0400

Yes, these are the big questions. I'll address your points first and then make some general points.

- email and blogs
I've never really understood the notion of having "a time" for these. Surely the point is that by their nature, email and blog posts appear throughout the day. That's why we leave our email client and feed readers open throughout the day. The question really is, when do you deal with them? When my RSS feed update icon lights up, I normally glance through straight away. If something looks interesting, I can open it in a new Firefox tab with one click and go back to look at it at my leisure. New emails, I read straight away and simply decide whether to act now or later, depending on the message. For me, it's more efficient to keep track of things as they happen rather than store it all up and spend an hour or two wading through it later.

- programming
I try to program all day, but I'm definitely best from early morning through to about 1 pm. If I'm "in the zone", I can't talk to anyone at all. I take 2-5 minute breaks about every half hour, when my legs seize up in pain. Yes, I need a better chair. Like Chris I find that there's a time when my alertness dips, usually mid-late afternoon, which is a good time for reading papers or blogs, blogging myself, organising references, playing with website design or something not requiring creative input.
It is most certainly not true that beer makes you more productive. When researchers drink heavily, it's to numb the pain, not to stimulate their creativity.

- editors and plotting
I do most coding in emacs, for the syntax highlighting and easy commit to CVS. If it's something short, quick, dirty and not requiring thought, I might use nano. If it's an actual text document as opposed to code, I'd also go with nano or perhaps something like kate, if I need 80 column line wrapping without pain.
I plot mostly using gnuplot or occasionally OpenOffice. I do very little editing or graphical analysis of single sequences if I can help it. I'm not sure what you mean by "plotting along sequences", but if I had to go there I'd be using Bio::Graphics from BioPerl.

- various
My own office is a far-off future dream. I listen to music only to drown out other people. My problem is that I like music, so I end up concentrating on that instead of the work. If you must listen to music, choose something gentle and soothing by a female singer-songwriter. Some weeks ago I tried listening to an album at work, appropriately named "Source Tags and Codes", by "And You
Will Know Us by the Trail of Dead". From one review: "Source Tags and Codes will take you in, rip you to shreds, piece you together, lick your wounds clean, and send you back into the world with a concurrent sense of loss and hope. And you will never, ever be the same." It was not a productive 45 minutes.

More generally. For me, the key is not so much how you spend your working day as how you organise your non-work time. Let's face it, even if you were mentally/physically capable of working every waking hour (which you are not), you would still feel that you were not achieving all that you wanted. This is the nature of research. So you need to make the best of your working day and of your time off. You want your home life to be as uncomplicated and stress-free as possible. Otherwise silly things will make you depressed - like getting to the end of the day and realising that the fridge is empty. So make sure that most days, you're in a position to come home, eat and relax straight away. Then you can decide how to spend the rest of the evening - work or otherwise.
Sleep is very important. If you're constantly thinking "I'll just try one more thing" and suddenly it's 1 am, stop it. Make a note, go to bed, start afresh tomorrow. It would be nice to work our own hours, but very few of us can truly do so. You may believe that your best time is from 2-5 am, but this will mean that you are asleep during the day. In most jobs, we have to interact with other people who are not asleep during the day. So complain all you like about the tyranny of the 9-5, but get your body into a routine that can deal with it.
Besides, humans are supposed to sleep at night and the days are nice. The sun comes up, birds sing, flowers open. Learn to appreciate these things.
Chris also mentions self-discipline and this is perhaps the hardest one to conquer, it certainly has been for me. You can only achieve it by constantly defining your goals and knowingly working towards them. If there's a paper that you need to write, you know when you are deliberately avoiding that task by messing about with code. Say at the start of every week "what am I trying to achieve?" and at the end of every week "what did I do towards my goals?" We all feel bad when it seems that we're not progressing, but you feel worse when you know that you're deliberately not progressing.

People are different and you do need to find what works best for you, but I think there are some commonalities. I think the key thing is continual self-appraisal. If you feel bad about work, don't just keep banging away at it in the hope that it will get better. Make time to ask why you feel bad - what contribution comes from you, your peers or the work environment and then try to figure out how to improve things. Science does involve a fair degree of continual pain and mental anguish, unfortunately, but like all mental illness, it can be managed! And publications really help ease the pain.

Damn, I feel ancient

chris — Sat, 08 Jul 2006 01:11:19 -0400

Very well, my son. Since you have hauled yourself over lava flows and razor-sharp rocks to receive wisdom at my feet, behold that which is mine to give :-) The wizened old programmers around here (some of them are in their thirties already - the rest of us not too far behind) tend to have different routines, I would guess. Here's mine...

I generally get in around 10-10:30 am. That's fairly normal for my environment, at least for the people whose lives aren't constrained by day-care schedules. I generally rip through my email, then start working asap. I find that if I don't push myself to start immediately, I waste most of the morning and afternoon, and have to push far into the evening to get anything done. One routine I have is getting things straight in my head on the train to work. Plan the things I'd like to get done in the day ahead, map out a rough time allocation, and hit the ground running as soon as I get to my office.

There's a recurring theme in my work habits: organisation. I find that I spend a great deal of time figuring out what to do next, both intellectually and practically. Although systems such as Getting Things Done work for some people (have a look at Merlin's version), I find them too cumbersome as a whole. The principle is sound, though. Figure out what the next action is, break large tasks down to small ones, make lists, AND THEN DO THEM.

My own battle has primarily been with efficiency: if it takes you 12 hours to do 8 hours' work, you need to work smarter. Develop your own systems, discipline yourself, make copious notes, and write documentation. It's tedious, but saves so much time later. If you're anything like me, you'll find a host of busy-work with which to procrastinate: re-index your pdfs, re-label your cds, make miniscule tweaks to your tex templates... the list goes on. Don't do it: it's not real work, and it's a time sink. Although you feel as if you've achieved something at the end, you have done nothing you were supposed to be doing. A little discipline is required!

I usually take a 15-30 minute break for lunch, read blogs, surf the net, or more usually catch up on my HubMed RSS feeds and peripheral reading. My concentration dips in the early afternoon, so I save mindless tasks like pseudo-code, OS upgrades, bibliography updates, etc for this point. Then it's back to the task at hand: grant writing, coding, playing with data. I try to stop work between 7 and 8 pm, and go home, relax for a couple of hours over dinner/TV, and decide if I'm working later or not.

I'd suggest knocking off work about an hour before you go to sleep at the latest, and relaxing. It's good for your sanity. Although it's tempting to work during creative spurts, as we all do, regular hours of bashing will get you further in the long run.

re: tools. I use emacs for most things (ie latex, perl and R, which tend to comprise my universe). I find Sweave most useful. zsh is a friend, as are awk and sed on occasion. I avoid listeing to music as I work, as I find it too distracting.

As always, YMMV. Good luck!

How do you work?

maximilianh — Fri, 07 Jul 2006 18:28:33 -0400

Hi, before posting this to the site as a story, I want to know how you work. No, seriously, how do you spend the day with the kind of "work" that we do? After some sudden hearing and hair loss recently, I figured that trying to hack perl/python/C-code together during the weekend at night is not a good idea in the long run so I'd like the advice of some of you wise programmers out there. I guess you are not averse to sharing your wisdom with us youngsters. Here are some ideas what you could comment on:

- Do you start your day with your emails (I guess: no)
- When do you read Blogs? (I assume: never... :-)