I have a laptop and bunch of 5 min segments throughout my day. I used to think this was simply too short a time to do 'anything'. Though with a little practice and a few minor behavioral changes (like unplugging) I have realized that I can get something done in each segment. So now I am reaching out to all you quick-draw-programmers out there for more tips that will let me get something done in the shortest time segment.
On
I use standby on the laptop instead of Hibernate because it gets me to the IDE faster. I had expected to have battery problems because of this but it seems to work quite well.
On
Massive use of TODOs. When you start on a new class (I'm a java programmer), write all the method names, and TODOs instead of code, this usually takes (the infamous) five minutes. Then, when you have your next five, start with chosing a random TODO, and write the code. I prefer doing them in a random order, but you might find that writing all the TODOs in a method first works better, just try:)
As I'm not a full-time programmer, but a student and hobby-hacker, most of my programming is done at home, the library or a cafe. I'll complete one or two TODOs, and surf the web a bit, watch people walking by, order a new coffee or get some snacks from the kitchen.
This propably won't work in a production-setting, but for personal projects it's king!
On
I think an answer to another question has a good idea. jalf suggests:
A very simple trick might be to subscribe to the RSS feed for C++ questions here on SO.
A wide range of questions get answered here, on every difficulty level, and they generally get very detailed answers.
It won't replace a good book on C++ of course, but it might be a good way to discover a wide range of concepts, pitfalls and solutions you might not have known about otherwise.
So when you have 5 minutes here & there, check out an RSS of a particular tag in Stack Overflow and read (and answer?) questions.
On
Boring back story: I was in a similar situation when I had to look after my ten year old sis for a day. While we were waiting for a friend to come over I really wanted to get some features done on a personal project.
I found that playing a ~20 second acoustic bit of music and clearing my head before I wrote anything was really useful, along with spending the first 5 minutes making a threadbare list of things I wanted to complete.
On
I think the biggest hurdle is knowing what task can fit into 5mins. So the first thing I'd do is break down a bigger piece of work into a bunch of bite sized tasks, each of which will fit into 5mins. That way when you have your spare 5mins, you don't have to context switch to a large problem, then try to work out what needs doing & then try to get something done. Instead you just look at your task/todo list & grab the top item.
On
Adopt Test Driven Development.
A big cost in task switching is figuring out where you were last.
If you always write the test before you start, it's a no-brainer to pick up where you left off. Run the tests, whatever fails, that's what you do next.
On
Believe it or not, there's a website called Five Minute Videos, and they have a Software section.
On
I also sometimes program in short bursts, like while waiting for trains. My method is to throw an exception that says 'TODO: Next step is to return a query here'. When I open my laptop, I run the module I'm working on and it blows up, telling me what to so with a stack trace telling me where in the code to start.
Also, don't ever bother closing your IDE...
On
Programming involves two processes: thinking about your program, and typing the code into the computer. Try defining everything into small projects that should take no more than ten minutes. If you pre-compute what you want to type in and learn how to type well, you can knock some good work out in 5-10 minutes.
On
Project Euler!
Project Euler is a series of challenging mathematical/computer programming problems that will require more than just mathematical insights to solve. Although mathematics will help you arrive at elegant and efficient methods, the use of a computer and programming skills will be required to solve most problems.
Also you can read The Daily WTF, 5 minutes worth spent.
Here's an idea: when you have larger chunks of time, write some unit tests for functionality you plan to implement. Then, when you have a 5 minute increment, choose a test and write code to satisfy it. This way you aren't spending any of your 5 minutes deciding what to work on, you already did that and documented it in the form of unit tests.
And hey, you get TDD and test coverage for free. Bonus.