I have an ancient blog over at mutelight.org which was the predecessor to this site. It spans the first ~5 years of my professional career, and it’s full of bad ideas. I talked a lot about .NET, and thought that Ruby was a gift from god almighty to save us from heavy enterprise frameworks. I’d forgotten almost everything I wrote on there, and was entertained by this article, which suggests adding a Ruby dependency to your .NET project so you can use Rake. Ahead of its time really.
For the last ten years, Mutelight’s been slowly but surely sinking beneath the waves. It ran on one my old Ruby-based blogging engines called Hekla, and thanks to Heroku’s platform it still runs, but the writing was on the wall. The app is on the
cedar-14 stack, which was put into readonly last year. Before I archived the project, GitHub showed half a dozen major vulnerabilities in dependent gems. I barely remember how to even go about updating the thing.
But I like the idea of keeping URLs around and unbroken as long as possible (cool URIs don’t change), so I decided to mount a rescue operation. I spent last night converting it over to a static site built by Modulir, the same framework that builds this one. The conversion involved:
.rbfor metadata (!!!) to
.mdwith TOML frontmatter.
The job is done, and it only took one evening. Mutelight is now a static site that has a good chance of outliving Heroku, and maybe even Ruby. It’s also noticeably faster now that there’s no dynamic component. It’s still on a NIH stack, but one that has minimal dependencies aside from Go’s standard library. And even if Modulir dies, the HTML static files that it generated will be portable to any system for about as long as we can expect modern computing to live.
Chalk this up to another of “probably not worth the time” lockdown project, but hey, beats Netflix.