I finally quit my day job! This is my last week, and self-employment starts on Friday, when the big experiment begins. This week, I'm revamping the mousehouse.bio landing page so it better expresses what the software is, who it's for, and how to sign up.
MouseHouse is nearly feature complete for a version 1.0, so September will be focused on getting this feature - animal colony management - out to researchers who work with mice. I have one customer right now (whose name ends in "arvard"), which is enough proof for me that there's something here worth pursuing.
I'm working on slop. It's a tiny Markov-chain generator I'm writing in Scheme. My pipe dream is to make it good enough to serve it on my site and tarpit wandering scrapers. (Basically a very much dumbed-down and non-Rust version of iocaine)
It is entirely possible that I won't be able to optimize it enough to make it viable as a tarpit (I'm not very good at Scheme), but I always wanted to make a Markov-chain generator so that I understand how they work, so it's a worthwhile experience either way.
I'm working on zopt. A small, fairly basic (deliberately, but also a bit because I'm a relative beginner at Zig) command line argument parsing library. My goal is to have something schemaless and super simple to use, but the flipside is the lack of validation and autogeneration of help text etc. I suppose there are very capable libraries available already that do just that - I wanted something a bit more bare bones :). It's been a fun way to learn some more Zig, including comptime programming.
I tried it during Zig Day Seattle yesterday and I really like your API choices! The only thing missing for me is a way to iterate on the non-flag positional arguments.
Ah, that's cool to hear! I've been thinking about a good way of working with positional arguments. The current API for it is pretty rough. Being schemaless, I might have to resort to some heuristics to decide which is the first positional argument, or maybe I'll have to use -- to indicate that everything after it is positional.
Work: We're iterating a bunch on our agentic orchestration and memory system - Witan and I'm starting in on doing end to end UI testing with Playwright.
Personal:
Been iterating on a curriculum to teach myself how to build "visual sketches" a-la Zach Liberman using OpenFramework and C++. Still working on it but as part of that I'm actually beginning to work myself through the course.
You can find it here - feedback welcome especially from already C++ and OpenFramework graphics savvy folks.
Clearing the decks at $DAYJOB to begin working on an important new feature that has a (resaonable, discussed, purposeful) deadline attached.
Otherwise, waiting impatiently for UPS to deliver the screen for my DIY phone. It's already been delayed by a day; hoping it'll arrive in time for some weekend tinkering.
At work: I am going to take it easy and try not to stress because I am frustrated with my job.
[venting]Last week I got passed up for a promotion I was sure I'd get. Leadership kind of fucked me by changing a large project I am running. A large vendor management project was suddenly put on hold but just far enough into its progress that my entire contractor team knows they're being fired but doesn't know when. I am being increasingly pressured to use AI and surrounded by people praising low quality AI output.
At home: Continue running, cycling, learning K, studying Kubernetes, applying for jobs.
I am being increasingly pressured to use AI and surrounded by people praising low quality AI output.
This was one of my last straws at work - AI-generated requirements that are impossible to meet, AI generated code that shatters into a million pieces at the slightest touch, AI-generated Slack-announcements...I'm like halfway convinced that some of my colleagues have been fired, but their accounts were kept active and hooked up to an LLM.
Tomorrow I'll try to get some paperwork done from my university. (I left 23 years ago, now I need some certificates from them.) This is to start a master's degree in teaching, which is required for any secondary school (highschool, "trade school") here. Inscription ends on Saturday so hopefully I can get my ducks in a row.
I'm taking holidays until my job ends, but Wednesday I'll do the final "recreation hour" and on Thursday I'll work my last day. That's weird because that might be the last day in something I've been doing since... 2002. I don't know what I'll do if/when I complete the master's degree, but I might never work as a coder again, which is weird. It's been half my life.
I've migrated Uruky and Uruky Site Search into a new, bigger server during the weekend and this week I'm working on improving performance (as much as possible) and increasing the limits on the USS index and crawler a bit more. It reached 300 monthly active accounts, so we're super thrilled about that.
That looks interesting and the proposition to get the source code after a year is an idea I haven’t seen as such before; nice!
As you posted those 2 links: how do those 2 plans differ? I’m a bit confused for whom Site Search is given that it looks like a separate product or plan?
Trying to figure out what to do in my career as I'm bored doing anything CRUD. Would love to work in Rust, ideally something systems related, but it's been a pain finding anything that doesn't require 3+ years of production exp.
Also knocking out Module 2 and 3 of my Cybersecurity of Drones class!
Having and then recovering from gum surgery. I'll use the time I can't do heavy physical stuff to catch up on some virtual chores at the hobbyist hosting company that I co-run.
Continuing work on a Windows 98 display driver in Rust. I'm starting by replacing the normal toolchain (i.e. MASM 5.10) with modified LLVM components. Last week I wrapped up adding OMF, NE, and MZ support to LLD. This week I'm working on replacing ML.EXE with an OMF-aware llvm-ml. I think MASM's resource compiler may also be replaceable with llvm-rc, for full measure.
Still working on Zd, my extensible lightweight document format. I got some feedback from a friend and he pushed me to include a few more features. I'm generally trying to be very conservative about adding things---I ultimately want something that's maybe a weekend project to write a compliant parser---but he had some great points and now it's back to the drawing board a bit.
If you're interested I'd love to talk about it. I had a spec written but I've made a number of significant changes so I'm not publishing that one until I update it all again. My email is in my profile if you'd like to reach out.
The short idea is to recover 80% of Markdown's convenient, appealing surface syntax by layering it on top of an extensible, general purpose layer. Whenever the shorthands work, they're pretty and accessible. Whenever they're too simple, the "longhand" format is right there (and intermixes freely).
And then next time you're wondering if you can somehow contort or hack Markdown to make it work for your blog/obsidian clone/task management tool/book/knowledgebase/bibliography/etc you can instead define a simple Zd extension that supports your features and is already compatible with parsers/lsps/html generators/etc.
I’m back to work on the haskell-debugger [1] and the next major task is support for debugging multi-threaded programs. That means pause/unpause per thread, per-thread step in, step out, and listing all threads and their call stack (of the stopped ones) when stopped. Also fix all concurrency bugs from a debugger originally made for the single threaded case.
ashwinsundar | 13 hours ago
I finally quit my day job! This is my last week, and self-employment starts on Friday, when the big experiment begins. This week, I'm revamping the mousehouse.bio landing page so it better expresses what the software is, who it's for, and how to sign up.
MouseHouse is nearly feature complete for a version 1.0, so September will be focused on getting this feature - animal colony management - out to researchers who work with mice. I have one customer right now (whose name ends in "arvard"), which is enough proof for me that there's something here worth pursuing.
timthelion | 13 hours ago
Poor mice 😔
regulator | 12 hours ago
Congratulations!
brn | 11 hours ago
Congrats! I wish you the best of luck!
nemin | 14 hours ago
I'm working on slop. It's a tiny Markov-chain generator I'm writing in Scheme. My pipe dream is to make it good enough to serve it on my site and tarpit wandering scrapers. (Basically a very much dumbed-down and non-Rust version of iocaine)
It is entirely possible that I won't be able to optimize it enough to make it viable as a tarpit (I'm not very good at Scheme), but I always wanted to make a Markov-chain generator so that I understand how they work, so it's a worthwhile experience either way.
timthelion | 13 hours ago
I am going to therapy again to hopefully become less angry. I am also hoping to be less sick.
hgrsd | 12 hours ago
I'm working on zopt. A small, fairly basic (deliberately, but also a bit because I'm a relative beginner at Zig) command line argument parsing library. My goal is to have something schemaless and super simple to use, but the flipside is the lack of validation and autogeneration of help text etc. I suppose there are very capable libraries available already that do just that - I wanted something a bit more bare bones :). It's been a fun way to learn some more Zig, including comptime programming.
liberty | 6 hours ago
I tried it during Zig Day Seattle yesterday and I really like your API choices! The only thing missing for me is a way to iterate on the non-flag positional arguments.
hgrsd | 20 minutes ago
Ah, that's cool to hear! I've been thinking about a good way of working with positional arguments. The current API for it is pretty rough. Being schemaless, I might have to resort to some heuristics to decide which is the first positional argument, or maybe I'll have to use
--to indicate that everything after it is positional.liberty | 16 minutes ago
It'd have to iterate through the arguments, returning the ones that don't fit the other criteria, I'd think.
--seems subpar to me.I'm sure you'll figure out some solution. Also happy to discuss in the issue tracker if you'd like!
hgrsd | 12 minutes ago
If you wanted to make an issue with your use case that'd be great! Otherwise I'll make one myself later today to think it through :)
feoh | 13 hours ago
Work: We're iterating a bunch on our agentic orchestration and memory system - Witan and I'm starting in on doing end to end UI testing with Playwright.
Personal:
Been iterating on a curriculum to teach myself how to build "visual sketches" a-la Zach Liberman using OpenFramework and C++. Still working on it but as part of that I'm actually beginning to work myself through the course.
You can find it here - feedback welcome especially from already C++ and OpenFramework graphics savvy folks.
duncan_bayne | 4 hours ago
Clearing the decks at $DAYJOB to begin working on an important new feature that has a (resaonable, discussed, purposeful) deadline attached.
Otherwise, waiting impatiently for UPS to deliver the screen for my DIY phone. It's already been delayed by a day; hoping it'll arrive in time for some weekend tinkering.
antonmedv | 12 hours ago
Playing around with a pure PHP database in a single file, always wanted to try to create a such SQL database.
regulator | 12 hours ago
At work: I am going to take it easy and try not to stress because I am frustrated with my job. [venting]Last week I got passed up for a promotion I was sure I'd get. Leadership kind of fucked me by changing a large project I am running. A large vendor management project was suddenly put on hold but just far enough into its progress that my entire contractor team knows they're being fired but doesn't know when. I am being increasingly pressured to use AI and surrounded by people praising low quality AI output.
At home: Continue running, cycling, learning K, studying Kubernetes, applying for jobs.
ashwinsundar | 2 hours ago
This was one of my last straws at work - AI-generated requirements that are impossible to meet, AI generated code that shatters into a million pieces at the slightest touch, AI-generated Slack-announcements...I'm like halfway convinced that some of my colleagues have been fired, but their accounts were kept active and hooked up to an LLM.
koala | 11 hours ago
Tomorrow I'll try to get some paperwork done from my university. (I left 23 years ago, now I need some certificates from them.) This is to start a master's degree in teaching, which is required for any secondary school (highschool, "trade school") here. Inscription ends on Saturday so hopefully I can get my ducks in a row.
I'm taking holidays until my job ends, but Wednesday I'll do the final "recreation hour" and on Thursday I'll work my last day. That's weird because that might be the last day in something I've been doing since... 2002. I don't know what I'll do if/when I complete the master's degree, but I might never work as a coder again, which is weird. It's been half my life.
brn | 11 hours ago
I've migrated Uruky and Uruky Site Search into a new, bigger server during the weekend and this week I'm working on improving performance (as much as possible) and increasing the limits on the USS index and crawler a bit more. It reached 300 monthly active accounts, so we're super thrilled about that.
ur5us | 2 hours ago
That looks interesting and the proposition to get the source code after a year is an idea I haven’t seen as such before; nice!
As you posted those 2 links: how do those 2 plans differ? I’m a bit confused for whom Site Search is given that it looks like a separate product or plan?
blitz | 9 hours ago
Trying to figure out what to do in my career as I'm bored doing anything CRUD. Would love to work in Rust, ideally something systems related, but it's been a pain finding anything that doesn't require 3+ years of production exp.
Also knocking out Module 2 and 3 of my
Cybersecurity of Dronesclass!insom | 7 hours ago
Having and then recovering from gum surgery. I'll use the time I can't do heavy physical stuff to catch up on some virtual chores at the hobbyist hosting company that I co-run.
pkd | 4 hours ago
Day job, mostly as I am oncall and it's a bit of a shitshow lately. Also looking forward to the upcoming long weekend.
junot | 2 hours ago
Continuing work on a Windows 98 display driver in Rust. I'm starting by replacing the normal toolchain (i.e. MASM 5.10) with modified LLVM components. Last week I wrapped up adding OMF, NE, and MZ support to LLD. This week I'm working on replacing ML.EXE with an OMF-aware llvm-ml. I think MASM's resource compiler may also be replaceable with llvm-rc, for full measure.
calvin | an hour ago
Any reason you're starting with a toolchain replacement first?
btp | 12 hours ago
Day job work will be focused on monitoring our Kubernetes environment in Google Cloud and refining a web-walk monitor for our MSP clients.
Personal work will be migrating some of the Fosstodon external sites (status, blog, Mattermost) to a new Hetzner environment.
tel | 11 hours ago
Still working on Zd, my extensible lightweight document format. I got some feedback from a friend and he pushed me to include a few more features. I'm generally trying to be very conservative about adding things---I ultimately want something that's maybe a weekend project to write a compliant parser---but he had some great points and now it's back to the drawing board a bit.
b3n | 9 hours ago
Anywhere I can find the Zd spec? It sounds interesting!
tel | 8 hours ago
If you're interested I'd love to talk about it. I had a spec written but I've made a number of significant changes so I'm not publishing that one until I update it all again. My email is in my profile if you'd like to reach out.
The short idea is to recover 80% of Markdown's convenient, appealing surface syntax by layering it on top of an extensible, general purpose layer. Whenever the shorthands work, they're pretty and accessible. Whenever they're too simple, the "longhand" format is right there (and intermixes freely).
And then next time you're wondering if you can somehow contort or hack Markdown to make it work for your blog/obsidian clone/task management tool/book/knowledgebase/bibliography/etc you can instead define a simple Zd extension that supports your features and is already compatible with parsers/lsps/html generators/etc.
romes | 11 hours ago
I’m back to work on the haskell-debugger [1] and the next major task is support for debugging multi-threaded programs. That means pause/unpause per thread, per-thread step in, step out, and listing all threads and their call stack (of the stopped ones) when stopped. Also fix all concurrency bugs from a debugger originally made for the single threaded case.
[1] https://github.com/well-typed/haskell-debugger