Pro-work, anti-procrastination and unit tests¶

I collected some nice pointers from 43 folders, which gave me some ideas I'll now give back.

The first is 50 strategies for making yourself work, which is aimed at writers, but I think it's more generic. I modified two of the strategies that resonated most with me:

Set a quota for unit tests per day

Part of my work is programming and unit tests are something I need to do more consistently. Programming is OK, but still something I sometimes need to put myself to. Writing, say, 4 unit tests - plus the needed programming for making those automated tests work - might be just the thing to move my projects forward a bit every day. And if I want to be lazy, I just put in some unit tests for (untested) code I've already written, which is still good!

Write or program while listening to music

Music can help you and it can distract you. But you can use it as a timer. Type or write two or three CDs worth of words or code. I actually used this trick sometimes when studying: listening to one side of a 90min tape, break, other 45min, break, etcetera. The modern equivalent (which I'll try) is to make mp3 (or .ogg) playlists of about two or three hours and just keep on working till it's finished.

Bit programming related is the outline-tip. Making an outline of the entire story before starting actual writing. Programming-wise that's just like putting up a back-to-front thin prototype so that you can do a complete walkthrough of the finished product (with only one dummy item, but still). Then you can start filling in the blanks. Helps great. I got the idea from the pragmatic programmers, who call it tracer bullets .

The second pointer was life hacks, which had a good quote at the end from Ed Dumbill:

Edd Dumbill: Ideas rot if you don't do something with them. Don't hoard them. I blog them or otherwise tell people.

This is a way to look organized, "That guy has lots of ideas, what a genius."

The importance of blogging your ideas and your knowledge will probably only increase with time. When you research something, you'll increasingly turn to google-and-friends. Having all your ideas out there for google to find is a definitive improvement over having it locked up in some hidden-in-Elsevier's-or-Balkema's-vault journal article.

I doubt that there is a lot of risk in doing this. For some people. perhaps. But for the most it is way more important to get found. The chance that somebody else robs all your ideas and manages to sell it instead of you... Not too big, imho. You're probably the best placed to execute on your ideas!

Being findable on internet... I'm in the finishing stages of my PhD thesis (I gave the concept to my second professor today) and I needed some more papers to formally reference. Google. Google. Within my research circle there are two major conferences: ecppm and w78. Ecppm's papers are well-hidden in Balkema's vaults, w78's are wide open on the internet . Oh, and web-based journal itcon also makes citation way easier than the paper-based closed journals... And as most scientists re-visit their research from time to time you can get your information in equal quality from both sources. Guess which one gets quoted most... (By a factor of five the last time I saw it researched...)

Show the world what you're thinking! It is also great for storing your own thoughts: google is handy for searching those bits that I can't quite remember but that must be on my weblog somewhere... And making conference summaries is great way of re-finding that supporting evidence or origin for your theory that you heard once.

Just one last thing: unit testing your PhD thesis. Well, sort-of. I'm writing it in Latex (well, the first stage is python's structured text...), so it's all plain text. Which means you can write little scripts to weed out common errors. Commas with a space before, for instance. All sorts of small things. Better make them into a script right away, just to make sure you don't forget to test for them a month from now. Or filtering "missing citation" messages from Latex' logs and displaying the context in which they occurred so that you can add them easily... Powerful.