Trying to re-introduce myself to Elixir/Phoenix after not touching it for several year. Thankfully, not a lot has changed. Live view is amazing. Im working on a small scheduling app for teachers.
Also reading. I’ve accumulated a huge backlog at this point. Right now im reading “Nickel Boys” (Colson. Whitehead)
Yes, LiveView is great! I’ve really been enjoying it. I like feeling competent in a framework/technology and not feeling like I will have to reach for a new one to do something else that comes up.
Preparing for he Big Family Visit where my wife’s folks and other relatives descend on us for a week.
Also recuperating because I made a mis-step at work and brought a nerf gun to a knife fight, so I’ve been given a RIDE OR DIE project.
On the up side, I’m doing pretty well with it thus far so my chances are good (early days).
On the down side, this whole thing has been incredibly exhausting and I don’t want to wish my life away but wow oh wow I can’t wait to retire :)
Been really getting into No Man’s Sky on my Xbox lately. Somehow gaming with a controller on our big TV is way more relaxing than sitting where I always sit at my desk and twiddling with a PC keyboard :)
Not to pry, feel free to ignore me if this is too personal, but is this ride or die project something like a “performance improvement plan” or PIP? I’ve been growing a bit uneasy lately around the way some people at $WORK have had PIPs held over their heads, despite a complete lack of any transparency or documentation on what that process looks like for everyone else. It’s sort of this murky boogeyman that we’re left to guess about. I’m relatively young in the industry, so maybe I’m just not used to this sort of thing.
Advice for anyone who gets a PIP: start looking for another job immediately, they’ve already decided they want to get rid of you. Many Human Resources (HR) departments require extensive documentation to fire someone. A PIP is the first step in gathering that documentation. It is unlikely they expect something fruitful to come of it.
Obviously a generalization, not every place is the same, but I’ve never seen a PIP improve someone’s performance. Long before a PIP they should have talked with you about how things were going and given guidance.
If you get PIPped, you’re done. You can try to fight it, but if you lose, you get NOTHING, whereas if you go quietly you get a rather generous package.
High-level pitch: programmatic 3D modeling software in Lisp.
A few weeks ago I wrote a little proof-of-concept toy for composing signed distance functions in a language called Janet, compiling them to OpenGL shader language, and then raytracing them in a WebGL canvas. I’ve continued to hack on it a little bit here and there, and I think it’s pretty close to being something that other people could have fun with too. A big missing piece was figuring out how to do surfacing, but I took a stab at it yesterday and I’m feeling pretty good about the approach I landed on.
This weekend I’m going to try to get it into “alpha” status and publish it somewhere. Which means making it work at all on mobile, making the DSL a little more friendly, and adding more primitives. But mostly it means writing some kind of tutorial…
I have not! Thanks for sharing that; that looks really cool. I especially like the feedback between the rendered image and the source – that’s something that I’ve vaguely wished for, but can’t think of a good way to implement. I’ll have to look through this for shapes and combinators to add :)
Continue playing RE4 on my PS2 on the old retro wood panel CRT in my now cleaned retro room (All hail FreeMcBoot)
Researching about more ways to create AST parsers (I’ve recently been using linting more aggressively to avoid code styling discussions and it’s been excellent… I want to employ this for anything.)
I’ve recently been using linting more aggressively to avoid code styling discussions and it’s been excellent… I want to employ this for anything.
Oh! I’ve just started working on what I’ve been thinking of as a “project linter”. There’s currently more lines of Make than Python in the project, so there’s not much to see, but the README might be worth reading: https://github.com/benji-york/fend
I’m on vacation for two weeks, so I’m checked into a hotel in Barcelona. When I’m not strolling around the city, I’ll be lounging around beside the pool (within ordering-distance of the hotel bar, of course) and trying my hand at the 2nd WASM-4 game jam.
Gonna be taking the k8s route myself very soon as Im starting to deal with it at work. Ill probably do a single node k3s cluster to get acquainted with kubectl then do a multi-node full-blown cluster at some point.
Thanks! There’s a lot to learn. My setup is also based on k3s, and the k8s-at-home GitHub org (and Discord) has been a really valuable resource.
If you’re ever thinking of playing around with it before running it on physical hardware, I’ve been using Kind to experiment - it essentially starts up the cluster inside a docker container and configures kubectl to talk to it. I think some people use Minikube to do something similar in a VM.
Thanks for those resources! I learned about KIND back in June, didn’t even know it was a thing. I’ve yet to try it, but it’s on my list. Gotta get at least a single node up first.
I joined the k8s at home discord and will probably drag a friend of mine in there with me as he’s interested in this stuff too.
Was playing a word game with my wife that we realized would be a fun web game. The premise is one person says “What is X”, where X is the first part of a common “doublet”. Then the second person has to say the second word and propose a new doublet. E.g, a session may look like:
“What is undying?”
“Fame. What is vehicular?”
“Manslaughter. What is …”
We thought this could be fun to have in a Wordle-like interface, where you’re presented with, say, 10 starting words and 10 blanks you need to fill in. I hate timers in word games, so we’ll probably score based on the “rank” of the doublet.
The idea of doublet rank is to take the bigram frequency rank and compare it to the unigram frequency rank of the first word. A “stronger” doublet is one where the bigram frequency is closer to the unigram frequency. E.g., “undying fame” is a stronger doublet than “the house” even though “the house” is more frequent because “the” gets used in lots of places without “house”.
Currently waiting on a script that’s streaming all the Google Ngrams dataset bigrams through Python’s heapq module to collect the most frequent 10 million bigrams. Hopefully that will be enough. I don’t have enough space on my computer to store the whole dataset :-) This weekend we’ll work on the game itself.
Also, sous vide-ing a chuck roast for the first time. Never sous vide’d something for >6 hours before, this one is going for 36-48!
Fun sail, and packing for a two week family trip! Not looking forward to three flights in each direction, but excited to see family and friends in the flesh for the first time since the pandemic. My parents are keen to see their grandchild again as well, of course.
In gaps of free time I’m learning The Wind Cries Mary on guitar, and since the weather is ace I took the boy (10) for ice cream.
I was planning to host 2nd birthday party for my son, but seems that his infection made different plans for me, so it will be fighting with fewer and doing my best so he will not dehydrate during that time.
Initially, enjoying the fact that I got the manager job I applied for internally… but my smartphone refused to wake the fuck up and I may have to be in rescue mode and find a new one instead (maybe it is time to switch from Google Pixel 4a 5g to Iphone 13? who knows) with my 10y nephew coming for the we in the house. FML a little bit.
Besides that keep my growing fascination for J and K implementation in check for when I am going to have enough brain time outside work.
I have end-stage renal disease, and have been doing home peritoneal dialysis for the last nine years. With that, I have daily records of my vitals and the results of my treatment that I have been keeping on paper. It was high time that I had it in electronic form, so I have started the dialysis-data project to write something for data entry and analysis.
Last week I got the basic data types defined and basic persistence to JSON for storing the data complete. This weekend, I plan on cleaning up this code a bit and starting to make something loosely along the lines of Clean Architecture for this project. Then I’m going to start diving into figuring out interactions using the Textual Python library.
I haven’t had much time recently to do development and so my skills are somewhat lacking at the moment.
Watched Medea. I was a bit blown out by the performance and the play (the dialogues). I’m in love with attic tragedy but Medea wasn’t one of my favorite plays. I was low-key avoiding the play, well until yesterday evening.
Gonna start blogging again, go for a run and possibly go to the beach!
On one hand I want to try my hand at my OS project again, but I could also go install and wire up a little photoresistor outside so my LEDs can compensate for clouds passing over my room automatically. I’m thinking of just putting one at the end of a coax cable and running it back to the microcontroller I have in the power supply’s box.
(Also maybe I’ll update some of my private branches for various things to catch up with upstream)
I’ve speed-up the build system of my website. More than you years ago, I wrote cleopatra, as a way to make said build system a literate program I could publish on my website. The thing is, it was slow as hell, and I never really documented the all thing, so it was mostly pages of code at the end.
So yeah, no more cleopatra, I ran it one last time, then committed the tangled file, with a new makefile to replace cleopatra. I also take advantage of my motivation to implement several nice trick to speed things up even more. For instance, I use Emacs for most of my content, but rather than starting one Emacs process per file, I start an Emacs server, and use emacsclient (see this commit).
Overall, I am pretty happy with the result. It builds in 8s on my machine, and most of the time is used by soupault.
Trying to re-introduce myself to Elixir/Phoenix after not touching it for several year. Thankfully, not a lot has changed. Live view is amazing. Im working on a small scheduling app for teachers.
Also reading. I’ve accumulated a huge backlog at this point. Right now im reading “Nickel Boys” (Colson. Whitehead)
Yes, LiveView is great! I’ve really been enjoying it. I like feeling competent in a framework/technology and not feeling like I will have to reach for a new one to do something else that comes up.
Absolutely nothing code related :)
Preparing for he Big Family Visit where my wife’s folks and other relatives descend on us for a week.
Also recuperating because I made a mis-step at work and brought a nerf gun to a knife fight, so I’ve been given a RIDE OR DIE project.
On the up side, I’m doing pretty well with it thus far so my chances are good (early days).
On the down side, this whole thing has been incredibly exhausting and I don’t want to wish my life away but wow oh wow I can’t wait to retire :)
Been really getting into No Man’s Sky on my Xbox lately. Somehow gaming with a controller on our big TV is way more relaxing than sitting where I always sit at my desk and twiddling with a PC keyboard :)
Not to pry, feel free to ignore me if this is too personal, but is this ride or die project something like a “performance improvement plan” or PIP? I’ve been growing a bit uneasy lately around the way some people at $WORK have had PIPs held over their heads, despite a complete lack of any transparency or documentation on what that process looks like for everyone else. It’s sort of this murky boogeyman that we’re left to guess about. I’m relatively young in the industry, so maybe I’m just not used to this sort of thing.
Advice for anyone who gets a PIP: start looking for another job immediately, they’ve already decided they want to get rid of you. Many Human Resources (HR) departments require extensive documentation to fire someone. A PIP is the first step in gathering that documentation. It is unlikely they expect something fruitful to come of it.
Obviously a generalization, not every place is the same, but I’ve never seen a PIP improve someone’s performance. Long before a PIP they should have talked with you about how things were going and given guidance.
At my BigCorp this advice is absolutely valid.
If you get PIPped, you’re done. You can try to fight it, but if you lose, you get NOTHING, whereas if you go quietly you get a rather generous package.
High-level pitch: programmatic 3D modeling software in Lisp.
A few weeks ago I wrote a little proof-of-concept toy for composing signed distance functions in a language called Janet, compiling them to OpenGL shader language, and then raytracing them in a WebGL canvas. I’ve continued to hack on it a little bit here and there, and I think it’s pretty close to being something that other people could have fun with too. A big missing piece was figuring out how to do surfacing, but I took a stab at it yesterday and I’m feeling pretty good about the approach I landed on.
This weekend I’m going to try to get it into “alpha” status and publish it somewhere. Which means making it work at all on mobile, making the DSL a little more friendly, and adding more primitives. But mostly it means writing some kind of tutorial…
Errm, making sure : you’ve seen libfive right? It has this (well, in scheme)
I have not! Thanks for sharing that; that looks really cool. I especially like the feedback between the rendered image and the source – that’s something that I’ve vaguely wished for, but can’t think of a good way to implement. I’ll have to look through this for shapes and combinators to add :)
Oh! I’ve just started working on what I’ve been thinking of as a “project linter”. There’s currently more lines of Make than Python in the project, so there’s not much to see, but the README might be worth reading: https://github.com/benji-york/fend
Re. parsers: I don’t know if this is what you’re looking for, but Lark looks like a really nice parser framework.
I’m on vacation for two weeks, so I’m checked into a hotel in Barcelona. When I’m not strolling around the city, I’ll be lounging around beside the pool (within ordering-distance of the hotel bar, of course) and trying my hand at the 2nd WASM-4 game jam.
Trying (for what feels like the umteenth time) to get my homelab k8s cluster up and running and using either Flux or ArgoCD.
Gonna be taking the k8s route myself very soon as Im starting to deal with it at work. Ill probably do a single node k3s cluster to get acquainted with kubectl then do a multi-node full-blown cluster at some point.
Good luck!!!
Thanks! There’s a lot to learn. My setup is also based on k3s, and the k8s-at-home GitHub org (and Discord) has been a really valuable resource.
If you’re ever thinking of playing around with it before running it on physical hardware, I’ve been using Kind to experiment - it essentially starts up the cluster inside a docker container and configures kubectl to talk to it. I think some people use Minikube to do something similar in a VM.
Thanks for those resources! I learned about KIND back in June, didn’t even know it was a thing. I’ve yet to try it, but it’s on my list. Gotta get at least a single node up first.
I joined the k8s at home discord and will probably drag a friend of mine in there with me as he’s interested in this stuff too.
Thanks again!!!
“kind” made a lot of things … simple to approach I felt
Was playing a word game with my wife that we realized would be a fun web game. The premise is one person says “What is X”, where X is the first part of a common “doublet”. Then the second person has to say the second word and propose a new doublet. E.g, a session may look like:
“What is undying?” “Fame. What is vehicular?” “Manslaughter. What is …”
We thought this could be fun to have in a Wordle-like interface, where you’re presented with, say, 10 starting words and 10 blanks you need to fill in. I hate timers in word games, so we’ll probably score based on the “rank” of the doublet.
The idea of doublet rank is to take the bigram frequency rank and compare it to the unigram frequency rank of the first word. A “stronger” doublet is one where the bigram frequency is closer to the unigram frequency. E.g., “undying fame” is a stronger doublet than “the house” even though “the house” is more frequent because “the” gets used in lots of places without “house”.
Currently waiting on a script that’s streaming all the Google Ngrams dataset bigrams through Python’s
heapq
module to collect the most frequent 10 million bigrams. Hopefully that will be enough. I don’t have enough space on my computer to store the whole dataset :-) This weekend we’ll work on the game itself.Also, sous vide-ing a chuck roast for the first time. Never sous vide’d something for >6 hours before, this one is going for 36-48!
Fun sail, and packing for a two week family trip! Not looking forward to three flights in each direction, but excited to see family and friends in the flesh for the first time since the pandemic. My parents are keen to see their grandchild again as well, of course.
In gaps of free time I’m learning The Wind Cries Mary on guitar, and since the weather is ace I took the boy (10) for ice cream.
I’ve been sick literally all week so I assume this weekend will likely still be recovery at this point
get well soon mate!
I was planning to host 2nd birthday party for my son, but seems that his infection made different plans for me, so it will be fighting with fewer and doing my best so he will not dehydrate during that time.
All the best and I hope your son get well soon!
Hope everything will be fine soon!
Initially, enjoying the fact that I got the manager job I applied for internally… but my smartphone refused to wake the fuck up and I may have to be in rescue mode and find a new one instead (maybe it is time to switch from Google Pixel 4a 5g to Iphone 13? who knows) with my 10y nephew coming for the we in the house. FML a little bit.
Besides that keep my growing fascination for J and K implementation in check for when I am going to have enough brain time outside work.
Smartphone decided to work again after an hour in the fridge, can just enjoy the nephew and the promotion.
shoot a video with my friend. It’s for learn and explore new area :) (video making)
I have end-stage renal disease, and have been doing home peritoneal dialysis for the last nine years. With that, I have daily records of my vitals and the results of my treatment that I have been keeping on paper. It was high time that I had it in electronic form, so I have started the dialysis-data project to write something for data entry and analysis.
Last week I got the basic data types defined and basic persistence to JSON for storing the data complete. This weekend, I plan on cleaning up this code a bit and starting to make something loosely along the lines of Clean Architecture for this project. Then I’m going to start diving into figuring out interactions using the Textual Python library.
I haven’t had much time recently to do development and so my skills are somewhat lacking at the moment.
Touching up my personal site. Breaking ground on my partner’s professional site, following some of the same principles I used for my own.
If I can get up the motivation, doing a bit of reading and cooking.
Last weekend of a fortnight sailing. Cleaning the boat, making sure the logs are up to date and flying home.
Maybe a swim or two first though.
Watched Medea. I was a bit blown out by the performance and the play (the dialogues). I’m in love with attic tragedy but Medea wasn’t one of my favorite plays. I was low-key avoiding the play, well until yesterday evening.
Gonna start blogging again, go for a run and possibly go to the beach!
I’m not sure yet.
On one hand I want to try my hand at my OS project again, but I could also go install and wire up a little photoresistor outside so my LEDs can compensate for clouds passing over my room automatically. I’m thinking of just putting one at the end of a coax cable and running it back to the microcontroller I have in the power supply’s box.
(Also maybe I’ll update some of my
private
branches for various things to catch up with upstream)I have 2 more weeks of vacation starting. We’re down at my parents for a big ole shindig with cousins on my mother’s side.
I have some plans to refactor HN&&LO. The DB queries are … suboptimal.
Stuff needs to be done around the home.
Oh, and Elden Ring.
Getting in 4 hours of overtime tomorrow, and then free for the rest of the weekend.
I am at DEF CON this weekend, the panels are good. Turns out TOTP isn’t great in practice with a patient adversary. Looking forward to seeing more
I’ve speed-up the build system of my website. More than you years ago, I wrote cleopatra, as a way to make said build system a literate program I could publish on my website. The thing is, it was slow as hell, and I never really documented the all thing, so it was mostly pages of code at the end.
So yeah, no more cleopatra, I ran it one last time, then committed the tangled file, with a new
makefile
to replace cleopatra. I also take advantage of my motivation to implement several nice trick to speed things up even more. For instance, I use Emacs for most of my content, but rather than starting one Emacs process per file, I start an Emacs server, and useemacsclient
(see this commit).Overall, I am pretty happy with the result. It builds in 8s on my machine, and most of the time is used by
soupault
.