OCC 2024

Retrospective · July 27th, 2024

I'm apparently really bad at doing anything in a timely manner these days. There's never enough hours in the day to accomplish the things that I'd like to do, such is life sometimes. But that doesn't mean I shouldn't try! How thematic for this year's OCC, no? Just because something is obsolete, doesn't mean it's useless, or unworthy of effort. Sure it might be harder, the obstacles greater, but the victory is all the sweeter once they're overcome! In essence, every year us OCCers dig deep into our junk drawers and take the "obsolete" and face a difficult world that has grown away from its capabilities, and we charge forth despite the odds that we may never return!

Okay I jest, we sometimes return to newer more powerful machines at the end of the challenge, like I'm typing this post from my trust Droid4 which has a 2c armv7 cpu, and a whole gig of ram. Literally twice the limits of the challenge! And I'm still being cheeky, because there's a whole swath of modern technology that I use day to day to accomplish daily tasks, and enjoy my hobbies, but these old and slow ones are near and dear to my heart. And worth every bit of the effort we put into them to keep them running.

Okay I jest, we sometimes return to newer more powerful machines at the end of the challenge, like I'm typing this post from my trusty Droid4 which has a 2c armv7 cpu, and a whole gig of ram. Literally twice the limits of the challenge! And I'm still being cheeky, because there's a whole swath of modern technology that I use day to day to accomplish daily tasks, and enjoy my hobbies, but these old and slow ones are near and dear to my heart. And worth every bit of the effort we put into them to keep them running.

Anyways, OCC for me this year was a little unrefined, or rather, it felt like I was just doing what I would normally have done anyways. I focused on doing real world practical tasks that I would have tried to work on from a more powerful system. And it more or less just worked. That's not a surprise to me at this point, but I think it should cause everyone a moment of pause, at least when we're considering the tools we use for the task at hand.

For me, most of what I use a computer for doesn't require a top notch spec. I need no GPUs, nor an i9 11th gen CPU. I don't need a massive high res screen, though my eyes thank me when I do use one. But those requirements, or lack thereof, are a direct correlation of the types of things I do with my computer. I don't really game, and if I do it's with my son on a dedicated system. I develop software for fun, and help maintain a Linux distro. Most of what is needed to do both of these things grew out of the same world that brought me these small x86 netbooks that I love so much, so they just fit right in.

But those systems do not run my blog, that's a nice little VPS on Digital Ocean. And I have a homelab filled with salvaged hardware, most of it far more powerful than the computers I use day to day. And these systems all have their purposes as well, little incus clusters, or custom monitoring servers. I kind of need it all, because honestly all of this old tech takes time and effort to keep together, and it's so much easier to bootstrap from more powerful systems. I'm very privileged in that manner.

Like my SBCL builds, using ECL took ~8hrs a pop on my x86, during with my system wasn't really usable due to resource load. The same build on a modern x64 system is only a couple of hours. Between the time I started this post, and its publishing now, my droid messed up again. We pushed gcc 14 in Alpine, rebuilt our firmware, and well my linux kernel for my droid was built against an older gcc and now I can't properly load those firmware blobs. They just fail to load, so no wifi for me, and no "posting" this blog post from my droid. Nope I need my Chuwi netbook and its built in micro sd card slot to pull this file off, just to git commit and push it up to the server.

And to fix my kernel issue? I need an arm SBC of some sort, an RPI or Pineboard comes to mind. But AWS sells their t4g.medium aarch64 servers for $0.81/hr, and a kernel build on those specs only takes 2 hours, so I might as well use the more powerful resource for the job that needs it.

All of this is to say that I am an incredibly stubborn person at times, and love the weird little niche I put myself into with my quirky old systems. And for that weird hobby I am willing to go well out of my way to do what it takes to ensure I have the tools I want and need available to me so I can enjoy them. And because of that effort, I can easily tell you that you can be productive on last years junk systems.

This probably doesn't jive with everyone, and for things like photo editing I do actually need a more powerful system, so I sit firmly in the middle with some sort of quasi physical maximalism enabling my digital minimalism. It's weird, but it's this and cameras folks, if I downsize it'll be because I finally bought a boat and have that burning a hole in my wallet instead!