Greg Crisp

You’ve heard the saying “unable to see the forest for the trees”?  So many projects suffer needless delays on many levels because there is so much work to do, that time can’t be spent up front to reduce the overall effort.

Here’s a small, but specific example I encountered recently, and how I made things better.  A client has a complex, command-line oriented build process, which involves several steps to complete, each executed in a different location in the folder tree.

It’s really not that complicated, but executing all the change directory commands, followed by the build commands for each step is tedious, and also difficult to remember.  It’s a lot of typing.

To streamline this, I spent about 10 minutes writing some XP command scripts to execute these steps in sequence.  Also, to save time, brain cells, and keystrokes I made these command files’ names short [though unfortunately cryptic], and put them into a folder in my system path.

There are still multiple commands to execute [not all steps need to be executed every time, so it's not one script], but each time it saves at least 20-120 seconds of typing time, depending on what I’m working with.  Add that up over a whole day of building and debugging, you’ve got several minutes per day.

If you’re splitting hairs, this seems like a net loss, and insignificant on a single day, but on day two, it breaks even.  Every day after that, it adds up.  Think about what you can do with an extra hour every week?

How many other repetitive tasks could be automated?  What happens on a larger scale?  How much time could be saved over a long term project if you spent two weeks automating your regression testing?

Automation always seems painful initially, but the gains for repeated process are significant over time.