The Sunk Costs of living in an Open Source World

[elfs]

One of the things I've noticed is that I have a surprising indifference to a sunk cost of time. If a client and I agree on a project that will take six months, and five months into the project a BSD-licensed version of what I'm building comes out, I'll gladly throw out all of my work, edit and revise and theme the BSD-licensed work, and present it to the client as fait accompli, complete with an explanation for this "interesting new library / toolkit / application service provider" that I chose to use.

When it comes to clients, my ego is in delivering, not in writing code. I reserve that shit for my personal projects. (That said, my ego is weirdly fed by the notion that I've actually kept up the Canvas Experiment series, despite having little time to work on it. It's actually taken me more than a month to get to the current stage, and sometimes more than week goes by before I can get back to it. See The Cult of Done Manifest, Rule #5.)

But that ruthless attitude seems to drive people crazy. "You just threw out five months worth of work?" So? The work wasn't the reward; the client's acceptance was. If I can get to that leveraging Wordpress, or Django, or whatever, good. I have two eyeballs; an active developer commmunity has dozens with which to find the bugs. The less code I develop, the fewer bugs I'm going to introduce. Those five months are sunk, but not wasted; templates, look and feel, and unit tests are all transportable to a new platform.

Comments

[http://users.livejournal.com/_candide_/]

Agreed!

I knew, 13 years ago when I first changed careers to software development, that having months-worth of code go unused would be normal. So I came to terms with that, in exactly the way you said, Elf: Save what you did/learned for future use. Let go of the code itself.