I increasingly hate computers and the world inside them that my brain is trapped inside of. That said, here are some things about computers. Let’s be practical, with another resource to get you to stop screwing up shell scripting. Once you’ve figured that out, why don’t you train an AI homunculus to reflect your own neuroses back at you, and start a conversation? It’s time for a vibe shift.
The unfortunate truth is that the digital world is the real world is the digital world. Culture and reality do not reside cleanly in a single realm; they slosh back and forth (an ass smashing into an ass, forever). The seemingly rigid schemas and structures in computing are in fact as plastic as our human messiness demands they be. As such: the definitive guide to vaporwave text encoding.
I was a real life, buck-toothed nerd when I was a child. I liked video games, didn’t play outside enough. I spent a lot of time playing around on a computer. But I never built one.
As an adult, I pay for the convenience of not having to be in one place to do computer things. I have been subscribed to Dropbox, iCloud, and many TV and movie streaming services over the years. Those are becoming less and less of a good deal, and the geeky voice at the back of mind is saying louder and louder: why not do it yourself?
TL;DR I’m switching from Goodreads to Oku. Sign up here (referral code).
I got rid of most of my social media accounts. The remaining ones are really services I use to track something I do myself that I share with others: Strava (running and cycling), Duolingo (learning languages), and Goodreads (reading). Of these, the one that I have always been dying to replace is Goodreads. The website and the native app are both terrible, it’s owned by Amazon, and the means to get your reading data out of it and into something else are being made increasingly difficult. But for a long time, Goodreads has dominated. I think the main reason is social inertia: people are on Goodreads because other people are on Goodreads. If you really want to see what your friends are reading, and most people are on Goodreads, you had better be on Goodreads too. For that reason, any challenger to Goodreads had better be a lot better, not just a little.
Ideally all the books in the API should stay in the store even if they haven’t been included on any of the named shelves in the last Goodreads scrape. When a new scrape is run it would add any books that don’t appear, shift any books that are in the store but don’t appear in the latest scrape to a no-shelf status (representing books I know about but have no relation with, I guess). All of that is much easier if Goodreads has a persisting book ID. If the data persists I can add my own columns to the data too, like whether I own the books.
Last month, external accessibility experts certified buzzfeed.com as compliant with the best accessibility practices for the web. That simple statement, ripped straight from the headlines of a boilerplate internal email, does not do justice to the two-year process that brought us to that point. Nor does it embody what the achievement means to our team, especially myself, on a personal level.
I was prompted to write this article by the experience of a friend of mine who was somewhat embarrassed by having his full, legal name called out in front of a room full of people who only knew him by another name. The was the result of a software system suffering from precisely the ailment I described: someone used the ’name’ field for the name displayed to kitchen workers who were preparing an order. There was no need for his legal name to be used there, but either the integration was made before the ‘preferred name’ field was added, or the developer of the integration simply didn’t think to use it.
I wanted to quickly follow up to my recent post about personal infrastructure with some updates I made this week.
I got a warning last week that I was almost at the limit for my allocation of “build minutes” on Netlify. Upon investigation, I found that my personal website had been building too often and for too long on Netlify, and that soon they would start charging me for the overages. Looking at the logs and running the build locally I saw that the vast majority of the build time was down to preprocessing the many images in the “Photo” part of my site to compress and resize them. So, in the short term those have been removed; I wasn’t really presenting them very well anyway.
I can, for example, guarantee that this blog is faster than any Gatsby blog (and much love to the Gatsby team) because there is nothing that a React static site can do that will make it faster than a non-React static site.
Note: There’s a follow up to this because I’ve since made more changes to the infrastructure of the site. Read more.
I’ve been slowly moving over to self-hosting more services and trying to balance that with personal convenience. This post is a quick summary of the current setup I have running to do the following:
My website is built on Hugo, a static suite builder written in Go. I like that all the content on my website can be markdown files with some front-matter, any extra data can be in simple JSON files, and the template system is very simple. I’ve spent a lot of time tweaking and playing with this Hugo site, but really it would work just fine without much work at all.
I spend a lot of my time picking apart how things work, and a lot of time sticking things together to see if they work in the way that I hope. That’s tinkering. I’ve been thinking about how I first started working this way.
I remember when I was a kid, I spent long days in my dad’s office. His office was actually a garage, a separate building from the house, across the back yard. He had his main desk in the corner, three computers lined up underneath with one of those boxy, ninetees monitors on top. His workspace was never clear and minimal; he was always snowed in with papers, and pieces of circuit board. He had slapped together racks of power supplies and oscilloscopes that spouted wires with little plugs and clips on the ends. On the other side of room was a standing workbench with even more electronics equipment: a soldering iron and endless reels of solder that I used to bend into animal shapes, breadboards and bags of loose microchips. Piled up in the warmest corner of the often chilly building was a set of servers, keyboards, and monitors. Somewhere between three and six servers were built into standard PC boxes, linked up by an octopus of Ethernet cables, VGA leads, and KVM switches.