What programming/technical projects have you been working on?

Learning PyTorch for some deep learning experiments I want to do. I always hate the initial phases of learning a new topic, because it can be hard to tell where you jump in. A lot of tutorials cover very basic things that make it hard to stay interested, but without those it can be hard to learn the syntax and become aware of different capabilities in the library. In any case I'm mostly through that phase and on to more interesting things! My first impression is that this is much easier than Tensorflow, which is what I previously had experience with.

Last year I made a scavenger hunt minigame for my wife’s birthday in python.
It included real life elements, as well as some programming and tech concepts like encryption and cryptography, T9, coordinates and with a healthy dose of video games mixed in.

This year I decided to do something similar, but simplify the setup: more encryption, API and rest requests ,QR codes, NFC tags, looking for clues with UV light and with all the steps, knowledge and abbreviations contained in an Obsidian vault

Made my own internet TV channel to act as a analogue for syndicated television without the ads. I wanted a station showing a rotation of shows that I could throw on whenever that I normally wouldn't put on and watch. Styled it after adult swim with short schedules with bumper music and a quick moment of Zen before an episode starts.

I have a copy of OBS running in a docker container with folders mapped to my plex library for media. I wrote an OBS script in LUA that selects a random line from a text file that contains a path and metadata to an episode. Script also selects a random song, moment of Zen video, updates the schedule text and managed the scene transitions. More recently I updated the LUA to send a get request using curl to a simple single file php script that both keeps and presents a running log of played episodes that are stored in mariadb.

The stream is sent to an OwnCast server (also running in docker) so that the show can be sent in multiple bit rates for web viewing and is in a compatible format for google casting. Owncast doesn't have casting support built in but is simple enough to add with cast.js in the custom JavaScript section of OwnCast. All of that behind an nginx reverse proxy mostly for cleaner subdomains and a way to password protect the stream.

Does it count that I'm learning to make my own game by using an engine? Kinda been messing around with Defold and seeing what I can do with it but I have no prior experience.

I'm trying to learn lua by doing a lot of exercises.
But damn it, trying to print a diamond always gets me... if it's python, c# or whatever.
I never get it right without messing around to much...

I've been playing with movement methods in Pico-8. Mostly I'm trying to learn the system and really be able to develop stupid little puzzles, but want to figure out a sokoban game.

I'm also doing a personal project to live in Immutable Linux distros. I started with Kinoite (Fedora's KDE based immutable project), but was inspired to stick with it by Universal Blue (uBlue). It's based on RPM-OSTree, but a goal of mine is image flexibility and either making a new image if I deem it necessary or containerizing/third-party packaging anything else to avoid "layering." This does two things: Ensures what I do works on many distros, and lets me rebase at a moment's notice if I see something cool.

Tonight I'm riding a high because I got Tidalcycles running on my uBlue Kinoite system. Distrobox didn't work for this, so I decided to try Nix. I had to add the nixgl channel to my nix system, then created a shell.nix file to pull haskell, tidal, and SuperCollider, and got it tied together in vim-tidal, a Vim plugin to use tidalcycles. This led me to being more comfortable in Nix, and gave me a way to try to do all of this declaratively, either by making a new ostree image or building Nix environments for applications. My next goal is to get other native (or near-native) editors working properly, then see about getting this out to the Tidalcycles community.

EDIT: For the broader internet, since it doesn't make sense to bump this, Distrobox worked just fine on the 2012 macbook I was using when I reconfigured it. Not sure what happened, but it wasn't even the distro on the box.

This is super small, but I am really happy with it

Years ago I was watching a lot of movies and kind of burned out on Japanese and Russian cinema. These days when I go to watch a movie, I typically go for something precode yada yada yada.. anyway, I've decided that I need a plan and this plan is to have two days a week where I watch movies

• Tuesday: the remaining unseen movies from the IMDB250 -- not a lot here, so I might as well just finish it
• Thursday: I'm going through the AFI lists (I think... I can't remember where I scraped the lists from)

So I have these lists pulling in to Google Sheets, then I run a script to pull all of the IMDB info. Next up, that goes into my calendar sheet which generates the dates and the ICS file in K:K. I then publish the sheet and have it output to a TSV and set the range (K:K) -- since its only the one column, no tabs.

Dump this embed URL into Calendar and all of my devices sync up in real time. It takes a minute for Sheets to update the publish, but it doesn't take long.

This is wordy... but having something like this almost completely automated is exciting.

quick edit: macos is cool with this, ios is not. To get around this, I set up a cron job w/ wget to pull it all to the .ics file I use. Not as fun, but still works well.

I got TaskRabbit ready to launch! It’s getting reviewed by the Commerce7 team right now, but we should have it on their store soon!

I’ve been pondering my next project, but none of the ideas are fleshed out yet.

I've been modeling Turncoats in F#.

Back to doing more coding work at work again and have been meaning to take a simple game like this and try to use it as a top to bottom test of my skills these days. Both trying to better model from domain, and hopefully actually solve the frontend side of it as i'm woefully lacking there.

Helps that it's a simple game, but one goal is to also show %'s of certain outcomes before actions and also allow for a dynamic sized board/stone count or even adding of additional colors. Not because it would make the game better but more just to test that i'm building my tools right and everything doesn't catch fire and burn down when a change is made.