14 Mar 2009
My New Development Strategy
I’ve done quite a bit of work on a few projects over the last few years and now I’m at the point where I’ve started a fulltime job and am looking at mountains of unfinished code and asking myself if this was this all just a waste… Quite a bit of original work, some ideas which I thought were quite good and worst of all an end-game project which I had barely gotten the time to start. And then my personal life - wondering how much longer I can try to slave away at code that’s in my head and push aside other life goals and delay other projects that are more in my reach.
So, I have decided to start applying the following three-step process to all of my open-source development work:
- Document the proposed API/Specification. This way I can start getting feedback early and interest other possible developers in collaborating to reduce my workload.
- Write test/benchmark code. This should actually be done first - it’s the framework for should happen when things are done right.
- Implement. It will be easier to do this with help from others if 1 and 2 are already done.
To my credit, the work I had done wasn’t a complete loss - my Xml and Json modules for Sendero eventually made it into the library Tango and the XML module evolved (with the help of Kris Bell) and was eventually benchmarked as one of the fastest in its class.
But, there’s a lot more code I put into Sendero, and it was never the intention for it to sit and rot. I feel like I wasted so much time. So, after some deep meditation and wrestling with my inner critic, I’ve come up with this strategy. I’m hoping this will take off some of the weight off.