In less than 24 hours I’ll be getting on a plane and traveling for a few weeks and then moving, all while continuing to work remotely. Which also means I am going to be without my auxillary monitors for the forseeable future. I considered taking a monitor with me but was so disgusted by the idea of lugging a full size screen across international (okay, Quebec) borders and up and down stairs and through airport security that I couldn’t follow through with it. I hear there are ‘portable laptop monitors’ now which are basically big tablet screens with HDMI in, but I am cheap and it is too late to order something for arrival before I head to the airport. Eventually I’ll be reunited with my big screens. I will survive.

This got me thinking about why I am so reliant on the extra screens anyway? And you already know the answer: most software needed for working a day job absolutely sucks and requires a lot of screen real estate to suck less. I am not fully a developer these days, but I still have to code and spend a lot of time in an IDE, almost as much time as I have to spend doing dogshit spreadsheet and presentation work. Then there are the fucking video meetings, which usually have screen share going on, oh and lets not forget everyone’s favorite mandatory chat application for distributed teams where people have developed weird expectations around availability and responses. It’s all webapps and Electron and integrated video and I hate it. But let’s be clear, I am not saying there was some golden age of software in the past - it’s always sucked, it’s always been a moving target where the future was always supposed to be better and the present was a work in progress. That’s Software (jazz hands)!

Last month before I knew I was moving but after I knew I’d be doing some travel while also working, I bought a cheap Chromebook so I had a second computer with a Linux terminal available for when I wanted to do non-work-related computing. Again, I’m cheap, it’s a cheap Chromebook, so it has comical specs, I think my phone has more horsepower. 64 EMMC disk and 4GB RAM baybeeee! Nobody is stealing this bad lad. But I was like, this is a challenge! A challenge to work within constraints! To effectively leverage a more powerful remote host for real work and treat this little machine as a thin client! No I will NOT repurpose my phone for this, I need at least a 11.3" screen and the ability to run multiple applications at once! I do have some minimal standards and even if the Chromebook isn’t ‘real’ Linux the terminal works well enough to get me to a real Linux host.

You can tell I spend a lot of time re-packaging shit news for palatable consumption. Anyway, the Chromebook is Fine as long as I actually treat it as a thin client. I use the browser and the terminal and that’s it. Now I’ve already spoken about web apps above, I don’t need to discuss them again and anyway 4GB RAM vs React and/or Electron web apps is nobody’s idea of a good time, so lately for demo work I’ve been making SSH TUI applications with Go and Rust and I really wish they were more common in the wild. Consider:

  • security and authentication baked in
  • everyone loves a terminal! okay not everyone but who cares about everyone
  • great retro style
  • terminal is objectively cool
  • impress your friends and enemies by avoiding Electron and Tailwind
  • do you really need to inject video into every activity? no, and terminal application is there for u

An ex pointed out once that it didn’t matter how much money he had, his needs would just flexibly accomodate his resources. If he had it he’d use it and if he didn’t have it, whatever. Hardware is cheap now and software resource usage is like my ex (great analogy): it bloats because it can and because it doesn’t require any extra effort or creativity. The team can fit it into the sprint and the tech debt never includes refactoring until it’s critical. And so my stupid IDE plugin merrily consumes 10+ GB of RAM just doing… something non-critical. And then I, in real life, have to spin this stupid reality as a bug that is in the process of being fixed and will be rolled out in a future release.

There was an era - mid 80s to 90s maybe? - where TUI apps were A Thing and connecting to them over a network was also A Thing but it wasn’t secure and decades of exposure to ever-increasing graphics and RAM bloat and cache-reliance has made people equate that shit with what a ‘real application’ should look like. And just like everything else in neoliberalized hell world, creativity and trying something different is frowned upon because the market analysis does not show that profit is guaranteed.

I’m not saying SSH/TUI apps should be used for like, accounting software, but why not email? Or source control? Or chat? Actually all of these exist today since they’re the standard use cases for software developers who enjoy making stuff for fun but there is no reason they couldn’t be used for a lot of web applications where React descendents are used instead. Especially in environments where the app is accessed frequently but not kept open constantly and is primarily editing and consuming text it feels like it should be the default, maybe with a web app as a fallback or for external sharing purposes.