What always strikes me about the XKCD picture is that there are so many problems before we get to the punchline. The picture paints a picture of complexity thr

What always strikes me about the XKCD picture is that there are so many problems before we get to the punchline. The picture paints a picture of complexity through layering and centralized liability through the one lone leg. Even if that one leg were well staffed and funded, it is still a centralized liability. There are any number of problems that occur when you have a single spot liability, beyond just “we’re depending on a few people’s goodwill.” I wish we lived in a world where when we see that picture we stress less about the paucity of volunteer developers and instead observe, there should be more legs right there, redundancy would be good. Interoperability would be good. I wish we lived in a world where when we saw that complexity we said “it’s nice to reuse, but simplicity of algorithm and design should still trump quick select complexity.” The file formats, protocols, design patterns, languages, that should get people excited on HN are the ones where you can learn/implement/understand it in a day. Not the ones where we build artificial mountains of complexity and then celebrate those that put in the time to become gurus sitting atop of the arbitrary mountains we’ve built.