What I Want From Technology

30 May, 2024

Lately I've been working on a draft of a /uses page, as I keep getting asked at the point of using a piece of technology what I use it with. I'm still working on it, and I might not ever publish it but as a general philosophy, my approach is:

1. Keep things in files, plaintext where possible.
2. On systems I own/rent.
3. Under self-hosted services where practical, and;
4. Where not practical, be a paying customer.

But there's also what I want from technology, which is rather different. I've used so many different types of technology, with different purposes and in different use cases but I've noticed that since the introduction of app stores and pricing models, and subscription services that everything in technology has been getting worse.

I saw a Fediverse post the other day showing someone unable to play WipeOut on their Playstation from a disc because they hadn't signed in to Playstation Network. This is where we are now. This isn't the only instance of this kind of activity where you're reduced to crofting on other peoples' platforms. It's everywhere, from media to operating systems.

Enshittification isn't something we're going to avoid by continuing to reward this behaviour. To deshittify technology we need to use, reward, and if need be build better technology. To do that, each of us needs to actively consider what we truly want from the software we use. This isn't which software we should use - it's what we want from that software.

What I Need From Software

For now, what I consider non-negotiable for me to use a new piece of software for personal use is:

1. The source must be available. I must be able to build it and use my own build on whatever OS I deem fit to use.
2. I must be able to own the software and my data, with all the rights that implies.
3. It must be usable offline, unless it specifically relates to a remote service (e.g. Mealie, Syncthing, OpenSSH etc).
4. It must work on machines that are at least 10 years old, and ideally on appropriately configured machines up to 20 years old.
5. Cross-compatibility is important, I use a little of most Operating Systems at some point. I should be able to transfer data between systems running different versions and Operating Systems with some degree of portability.
6. The data must be exportable, portable, and easy to migrate to something else. No proprietary formats without common parsers.

These are things that are important to me. They may not be important to you. I use everything from CP/M (mostly for Wordstar) to Amiga Workbench, various versions of MacOS from 7.5 - Mojave, DOS and even sometimes Windows, so cross-compatibility is important. Not all of these machines have TCP/IP Stacks, so offline usability is important. Some of these machines are on the wrong side of their 30s.

I'm open to being more relaxed for tools I use in work, mostly because work has different priorities and I'm not always in control of the tools I have to use. But for personal situations, fuck it. If it's not meeting the above, I'm not sticking around.

Existing Services and Software I Should Probably Move Away From

Some of the existing software I use fails to meet this, most notably Obsidian, Signal Messenger, DuckDuckGo and Protonmail (none of which meet my criteria for source being fully available). I'll take a view on these, but I'm hoping to move away from them over time.

I've used Obsidian for nearly two years. I'm mostly happy with it, although searching is becoming difficult (even with plugins) due to my use of daily notes. Obsidian is Electron-based and also not Open Source. This means that I can't use Obsidian on OpenBSD or Haiku OS. As such I've started moving my permanent notes away to Vimwiki and MDWiki¹, and I'll probably move my daily notes to something separate so I can publish permanent notes online, hopefully via Gopher and Gemini.

I spent years trying to convince people to use Signal, and I'm extremely cautious about jumping to something worse. It's hard to run bots on Signal, I've tried repeatedly and given up each time. It's hard to run unofficial clients on Signal. I've tried repeatedly and given up each time. Signal has so many red flags it's hard to stay (MobileCoin, the Google integration thing for desktop and mobile, not syncing public repos with production for over a year while they implemented the MobileCoin crypto bullshit, begging for in-app donations while sitting on top of a huge loan, oh and that time Moxie went on a power trip against Wire). There are more, but I lack the spoons to go into them. Signal is also Electron based so doesn't work on Haiku or OpenBSD. I'll probably keep Signal for work stuff (same with Telegram), but I'm going to look into IRCv3 and take another look at XMPP when I get the chance.

Bing went down, taking Ecosia, Yahoo and DuckDuckGo with it. I've used DuckDuckGo for years but have noticed a decline in search result quality, and always thought they were supposed to be more than just Bing. Apparently not. I've had a SearxNG installation for a while and decided I'll tune it for a better experience.

Protonmail is another odd duck. You would think mail services touting themselves to be Open Source would be Open Source, but you'd be wrong. Only the client side is Open Source. That might be enough for some people, but not me. Frustratingly Protonmail requires a bridge to allow standard tools to access it, but this is Open Source and there's more than one implementation. I'm also a little put off by how keys are managed in Protonmail. When a service generates the keys - it owns them. Protonmail does this for mail keys and an 'Account Key'. I can import a GPG key I've generated myself, but not the account key. I use Protonmail at work, and I need to think more about how I use it in a way that will prioritize customer safety. I'm not against using it for work, but it's clear it doesn't meet my bar for personal use. I just want to be clear at this point that I don't think Protonmail is inherently wrong in their choices, although their marketing is borderline disingenuous.

The enshittification of the web and AI crawling has made me even less inclined to publish my notes as a digital garden, which saddens me a little. The solution is probably to use Gopher and Gemini, but that requires some more work and thinking.

#philosophy #software #tools