What are you doing this week?

I got the new job! I start soon as a principal engineer writing Elixir full time and I am f***ing STOKED.

I put in my 2 weeks on Friday and will spend the time wrapping stuff up and prepping to switch.

Outside that, I am designing and 3d printing a new silverware caddy for my parents dishwasher using OnShape. It is challenging my CAD skills in a great way. Before someone says it: don’t worry I have ASA filament and a printer that can handle it. Not about to stick PLA thru the dishwasher lol.

Deciding whether I want to continue playing with Joxa and turn it into a proper Thing. On the one hand, it’s fun! In a twisted kinda way. And it’s a nice break from grinding away on Garnet. And I’ve been thinking for a long time that if I ever make another semi-serious programming language, it’ll be dynamically typed and, if not run on BEAM, then at least do very BEAM-y things.

On the other hand, I need another side-project like I need another hole in my head, and apparently Clojerl is a thing that exists, and I feel bad about how little work I’ve done on Garnet this year. And I can’t think of a good name for it. Sooooo idk.

Got a few things I could be working on:

• Designing a watchface for my Watchy firmware in Rust.
• Idiomatic F# bindings to SDL3, which I plan to use in the next game jam I participate in.
• Hacking on the Lua implementation to maybe eventually turn it into a typed “system” scripting language.

I’ll pick whatever is grinding my gears the most at any specific time, and right now the third option is absolutely winning so I’ll stop myself here and start writing a blog post instead of turning this into a 2000 word comment :)

Continuing with my Japanese learning app[0] to try to make it easier to learn what you need in order to watch the content you want in Japanese. I currently have most of the basic flows done:

• You can select one of the already uploaded contents or you can upload an SRT
• There is a short evaluation to find a very basic distribution of what we think may be your current level and what words should be known (I may integrate with WaniKani to know better what words you know if you are already using that)
• Then you select the learning plan, where you can select the number of words to learn each week and the target acquisition percentage), that gives you an N week plan on what you will learn and how long it will take you to get that acquisition.
• Weekly story generation with the words you need to know that week. This adapts constantly to your new information, so the plan you created in the last step is just a bit of a guidance, if we learn you know this other word or that you don’t know that, we recalculate the plan on story generation.

This week I’m working on trying to release a small beta with some friends and interested parties, finalizing the SRS model (there are three different sources of what the user knows and aggregating them and querying them properly is tricky, I think I’ve found a better way to do it that still properly merges what the user thinks they know, what the reviews tells us it knows and the proficiency distribution for words where we don’t have any more info) and making sure the basic happy path works well. I’ve also contacted an illustrator to see if we can make some explanatory and decorative illustrations for the landing and internal pages. Once this is done we’ll go over the app with him and see what has more priority.

[0] https://katarineko.com/

Recovering from waking up with an inflamed knee at the weekend, by resting it and not doing anything too strenuous.

Selling a car on Friday, so pulling together all the paperwork for it and trying to remember where all the spares/things I’ve not fitted to it are so I can include them. Going to a friend, have to drive it ~200 miles to his to sell it, should be a decent enough farewell trip in it.

Updating my old pet project: a tiny CLI tool for searching files that uses SQL-like query language: https://github.com/jhspetersson/fselect

I just published my first Rust crate in years: sittard. It’s probably useless for most people out there, but the idea is cool and I couldn’t resist sharing. I hope to write a blog post about it soon :)

We started an educational application project with a friend who moved to Torino just to start this journey together: excitement and anxiety for all that is to come! For the past ten years I have been working for someone, but I have never worked for myself

Also, I’ve been working hard on $CURRENT_JOB$ with almost no time to devote to anything other than coding, which I love, but I’ve been very intoxicated with AI and my plan is to go offline (with coding tools) for a few days just so I don’t get too lazy with my thoughts

I will be working on my memory-safety framework, idea is to track all references at allocator level, so as to make it easier to detect some types of invalid memory accesses such as use-after-free, or accessing reallocated memory. Initial motivation was to have some assistance while sharing memory across threads, without deep-copying or casting pointers, in pursuit of performance. I am not an expert in writing allocators and such frameworks, just trying to make it work based on my existing knowledge . Still a lot of work, i would consider it useful only if could make it work with my existing codebases!

Editing my MacOS app again to pass App Store Review. They are confusing API keys with “unlock codes,” and based on the timezones they’ve outsourced this review process to India, which has increased the speed but likely introduces errors, and sometimes I get different reviewers that dismiss me for things previous ones have approved.

Adding Ansible 12 support to Mitogen