overtowed's recent activity
-
Comment on What programming/technical projects have you been working on? in ~comp
-
Comment on What programming/technical projects have you been working on? in ~comp
overtowed LinkI last posted on this recurring topic about a performance problem I had in Soggy Planet, a map of Earth that roughly visualizes ocean levels at the peak of the last ice age up to the theoretical...I last posted on this recurring topic about a performance problem I had in Soggy Planet, a map of Earth that roughly visualizes ocean levels at the peak of the last ice age up to the theoretical maximum.
Happy to say with Pixi's docs and a nudge from ChatGPT I was able to write a shader to fix the main performance problem, so instead of having 12 images, one for each level of terrain revealed at lower sea levels that get alpha blended, it's now 1 and controlled by this shader frag. It's not reusable and begs for improvement, but my phone no longer crashes! It was my first time writing a shader.
Next is adding pinch-to-zoom so mobile is more usable, and then I think I'll be able to call it "good enough to share". I'm being stubborn about not using any libraries so I'm not sure when I'll get to this.
-
Comment on What advice would you give to someone who has coded in jquery for years and now wants to gracefully switch to modern js? in ~comp
overtowed (edited )Link ParentExcellent explanation, I'll elaborate from a slightly different point of view. With the shift from imperative to declarative came the now well-developed concept of components. These are much more...Excellent explanation, I'll elaborate from a slightly different point of view.
With the shift from imperative to declarative came the now well-developed concept of components. These are much more than templates - they form a tree of objects with lifecycles and affordances for robust abstraction and communication. It's similar to the difference between hygienic macros (Rust/Lisp) and string substitution (C macros).
Before React we had Backbone/AngularJS/Ember/others all taking templates further towards components, but React was the first to popularize components in their modern form, which all the major frameworks use today. This means that if you learn the component model, you can much more easily move between React, Svelte, Vue, Angular, Solid, etc.
On the word "framework", IMO the library/framework distinction can be unhelpful. I consider all of them component frameworks. They organize your code and mind around components.
I'm a biased user but Svelte is my preferred framework. With React you'll be learning more concepts that are React-specific, but it has by far the most helpful ecosystem of resources and libraries.
So-called meta frameworks or app frameworks like Next.js, SvelteKit, and Angular are where I would point beginners, I use SvelteKit which recently hit 1.0. They're both powerful and flexible and make basic usage much easier unless you want to get deep into JS tooling. (most don't, here be dragons)
-
Comment on What programming/technical projects have you been working on? in ~comp
overtowed LinkI recently added a new set of features to a website with a flat map of Earth using NASA imagery: Soggy Planet. (not mobile friendly) It now shows up to 10 elevations below sea level, in addition...I recently added a new set of features to a website with a flat map of Earth using NASA imagery: Soggy Planet. (not mobile friendly) It now shows up to 10 elevations below sea level, in addition to the 3 elevations above sea level I implemented a while ago. I also added a day/night cycle and overlay Earth's lights at night. The controls need some work.
It works as intended but I'm not happy with the performance, especially when zoomed out. I'm using PixiJS and a simplistic rendering strategy: there's just a bunch of large images that get faded in and out over each other. I'm wondering what I could do to improve performance -- maybe shaders and combine images? It'd be my first time using them. The images with below sea level elevations are very simple, just a single color on each. (example)
The previous version of the map doesn't have the same perf problems because it uses fewer images. Here's the code. I'm planning to do some technical and design screencasts, I may do one where I just describe the performance problem and ask for help.
-
Comment on Why I think "Sponsor Only" repositories introduced by Github is a terrible idea in ~comp
overtowed LinkThis opens the doors for more models than just "pay for the code". For example a dev could still publish code to package registries and docs to a website for free, and paywall the repository....This opens the doors for more models than just "pay for the code".
For example a dev could still publish code to package registries and docs to a website for free, and paywall the repository. GitHub could support this usecase by making readmes optionally public, or paywalling only the issues/PRs and not the code.
Many users of open source treat maintainers as personal tech support, and for some projects this means a lot of wasted time and attention and burnout. It's common for some projects to have many issues opened due to a lack of knowledge not specific to the repo, or issues caused by different software or technology altogether.
I'm not saying this is a perfect situation -- paying users may feel more entitled, and there's the important questions of access to people who can't afford it and the effects on the ecosystem -- but it allows alternative models to be explored that I think could work well for some devs, enabling more free software for the commons that we wouldn't otherwise get.
-
Comment on <deleted topic> in ~talk
overtowed (edited )Link ParentWhat mechanisms can reduce LDB? Some thoughts: invite-only paywalls other barriers to creating accounts? account reputation that matters because the identity is valued (e.g. professional networks,...What mechanisms can reduce LDB? Some thoughts:
- invite-only
- paywalls
- other barriers to creating accounts?
- account reputation that matters because the identity is valued (e.g. professional networks, smaller communities)
- account reputation that matters because it gives you abilities on the platform
All of those imply there's stakes to LDB, something to lose. Wondering what else there is here, and other ways to look at it. (like the text medium vs others)
Edit: here's a relevant idea, Freenet's web of trust: https://www.draketo.de/english/freenet/friendly-communication-with-anonymity
-
Comment on After ruining Android messaging, Google says iMessage is too powerful in ~tech
overtowed Link ParentThis comment on Lobste.rs goes into a lot of technical detail about XMPP's shortcomings.This comment on Lobste.rs goes into a lot of technical detail about XMPP's shortcomings.
-
Comment on How do you actually meditate? in ~talk
overtowed (edited )LinkI've meditated very infrequently for about 15 years, occasionally intense, maybe a thousand hours total? So as an amateur with bits of many traditions, a few major lessons/insights come to mind,...I've meditated very infrequently for about 15 years, occasionally intense, maybe a thousand hours total? So as an amateur with bits of many traditions, a few major lessons/insights come to mind, big shifts in thinking that took a long time and a lot of observation/boredom/repetition to sink in, and they're still sinking in, or maybe floating away, idk.
- watching the mind generate thoughts constantly whether "you" "want" to "think" them or not -- many practices try to demonstrate this right away, because it's so accessible and obvious when demonstrated -- and, because you're now observing things, you start getting language handles on things, wrapping your head around your mind and body, with some genuinely fascinating (but still extremely limited) insight into how their continuous processes are interrelated
- watching and practicing the process of mentally detaching from the body's emotional reaction to stimuli (e.g. "pulling yourself out of your self-fixated story") while simultaneously embracing the physical sensations, so you process emotions rather than suppress them, and practice this as a skill to apply in difficult situations -- so it's self-conscious deliberate emotional regulation, and it helped me take responsibility for an important aspect of my mood more often -- also, something I think that gets too much attention is taking emotional transcendence to some idealistic endgoal, IME it's mostly counterproductive and driven by flawed ideas or ego
- there's many metaphors for the mind and body and the system and "you", like "taming a monkey" or "riding an elephant" or whatever, the idea is that you're just more aware of what's going on in there, and you have practiced tools with well-worn neural pathways for more often pulling yourself out of reactive semiconscious patterns of thinking and behavior, so you start seeing everything as process that's always happening, that can always benefit from a guiding hand, rather than getting bogged down on the past or future or unhelpful emotions or thoughts
Also in my experience, intellectualizing meditation is unhelpful -- my mind and body only got what feels like actual insight with a lot of experience. (lots of boredom, lots of quitting for long periods of time, never liking any particular resources or teachers for long) Then there's the problem that the more I practice the less motivated I am to practice, I guess because I need it less? So I'm always a beginner, but not in the humble way, in the incompetent dilettante way, and I can accept that.
I think meditation is great: there are many interesting traditions and it has helped me a lot. I'd say just beware dogma, overconfidence, sluggish states, power fantasies, and charismatic gurus.
-
Comment on Roguelike Celebration: Call for presenters in ~games
overtowed LinkRoguelike Celebration (roguelike.club) is one of my favorite things, and their call for talk proposals is open for the October 2021 online conference: here's their YouTube archive -- it has many...Roguelike Celebration (roguelike.club) is one of my favorite things, and their call for talk proposals is open for the October 2021 online conference:
Technical talks, game design lessons, creative uses of procedural generation, interactive roguelike performances - we want to hear about it all!
We get more proposals for Roguelike Celebration every year than we can accept - and we decided it would be helpful to provide some guidelines for what is more likely to get your proposal accepted.
here's their YouTube archive -- it has many many gems for people interested in games, game design, design in general, weird creative technical things, procedural generation/generative systems, with a more-passionate-than-average community
-
Roguelike Celebration: Call for presenters
5 votes -
Comment on What programming/technical projects have you been working on? in ~comp
overtowed (edited )LinkI made the first usable experience of what's probably going to be a big project: it lets you browse and draw tarot cards. https://www.dealt.dev/tarot You can try out some divination for yourself,...I made the first usable experience of what's probably going to be a big project: it lets you browse and draw tarot cards. https://www.dealt.dev/tarot
You can try out some divination for yourself, like rogue_cricket recently wrote about in a pseudoscience discussion.
There's a number of enhancements I want to make, like putting state in the url, so you can share links. Here's the source code. I'm excited to use its datasets in various generative experiments and tools. (this open data project was my starting point)
-
Comment on Linux bans the University of Minnesota for sending intentionally buggy patches in the name of research in ~comp
overtowed (edited )Link ParentI think that's speculative given Greg's analysis: Reminder we're talking about the Linux kernel here, not some random C project. Thoughtless patches are inexcusable and unlikely IMO, but I'm not...Assuming I haven't missed anything and we take everyone at their word, Pakki is just an innocent bystander here: he did not intend to send vulnerable patches to the kernel mailing list, and the wildly disproportionate reaction was due to prior bad faith actions he was immediately adjacent to but not actually directly involved in.
I think that's speculative given Greg's analysis:
A few minutes with anyone with the semblance of knowledge of C can see
that your submissions do NOT do anything at all, so to think that a tool
created them, and then that you thought they were a valid "fix" is
totally negligent on your part, not ours. You are the one at fault, it
is not our job to be the test subjects of a tool you create.Reminder we're talking about the Linux kernel here, not some random C project. Thoughtless patches are inexcusable and unlikely IMO, but I'm not on the inside here.
-
Comment on New groups and site mechanics - 2021 edition in ~tildes
overtowed (edited )LinkHas a feature like "private replies" been discussed as a moderation tool? So lighter-touch feedback can be given privately and in context? I can see it being a can of worms without sufficient...Has a feature like "private replies" been discussed as a moderation tool? So lighter-touch feedback can be given privately and in context? I can see it being a can of worms without sufficient trust and training. And it's hard labor. But maybe helpful to reach some new local maximum. Maybe it doesn't have to be an admin-level privilege? Maybe it already exists and a ghost will appear:
edit: reflecting on this a bit more, I believe it's a can of worms because it allows for very low-friction interactions with random people on the internet in what can be heated and serious exchanges. Sometimes things will spiral, people will get hurt, and nobody else will even know unless it's shared; we'll feel the aftermath though. And sometimes Deimos will wade into a disaster.
-
Comment on What programming/technical projects have you been working on? in ~comp
overtowed (edited )Link ParentMy experience has been similar to yours and I have some of the same questions. The basic principle is that you "don't break on unexpected data", but that says nothing about useful presentation of...My experience has been similar to yours and I have some of the same questions. The basic principle is that you "don't break on unexpected data", but that says nothing about useful presentation of data or functioning interactions. I'm still new to the space but as an implementer I see a few options:
- prioritize getting something that works that's spec-ish
- interpret the spec as best you can and fill in the gaps in its spirit
- do what Mastodon does
It seems a lot of implementers do #3. But Mastodon is a microblogging platform inspired by Twitter, just one of many possible forms it could take, too small a box for where I'm starting, and it makes some choices that don't fit my needs.
Given my primary goal of shipping something this leaves me doing #1. Interop is more of an eventual goal than something you get for free day 1. (until the ecosystem matures!) Getting a common set of data structures, with a rich base
Object, tied together with JSON-LD (webscale), and some common system language semantics, is a huge win, regardless of how much we might veer off-spec or extend it from here. I'd recommend understanding the basic design of the Activity Streams vocabulary to anyone building an app: the federation/protocol/ActivityPub stuff is a different part of the spec, and more domain specific, while the vocabulary is widely applicable.I think this all reflects how hard the problem is. Like you said the Activity Streams vocab can be used to construct arbitrarily complex nested objects, whatever "sentences" we want to say, so yeah, I think the spec authors know this and they're just saying "good luck, in the future we'll know more, go implement!". So uh have you tried throwing AI at it?
-
Comment on What programming/technical projects have you been working on? in ~comp
overtowed (edited )LinkI've been learning Fediverse/ActivityPub protocol stuff and made alternative docs for the Activity Streams 2.0 vocabulary. (here's the official ones) I like the vocabulary's design, my main...I've been learning Fediverse/ActivityPub protocol stuff and made alternative docs for the Activity Streams 2.0 vocabulary. (here's the official ones) I like the vocabulary's design, my main criticism being that it's too open ended, which for my exploring is great. Next is generating JSON schemas for the terms - it's partially implemented on a branch. From there you can do a whole lot. I couldn't find existing projects with good data.
-
Comment on How do we combat mass global misinformation? How about making the internet a little harder to use in ~tech
overtowed LinkFor more on the subject of dis/misinfo there's researcher Claire Wardle's work, like the article Information disorder: ‘The techniques we saw in 2016 have evolved’ .For more on the subject of dis/misinfo there's researcher Claire Wardle's work, like the article Information disorder: ‘The techniques we saw in 2016 have evolved’ .
Seven types of mis- and disinformation
fabricated content: New content that is 100% false, designed to deceive and do harm.
manipulated content: When genuine information or imagery is manipulated to deceive.
imposter content: When genuine sources are impersonated.
false context: When genuine context is shared with false contextual information.
misleading content: Misleading use of information to frame an issue or individual.
false connection: When headlines, visuals, or captions don't support the content.
satire or parody: No intention to cause harm but has potential to fool.
-
Comment on A twenty-year-old man was fatally shot while filming a YouTube "prank" robbery in ~tech
overtowed Link ParentGrowing up in the US I heard so many times how litigious people are, how they'll sue you for anything, with huge pain and suffering settlements and judgments in civil court. I don't know how to...Growing up in the US I heard so many times how litigious people are, how they'll sue you for anything, with huge pain and suffering settlements and judgments in civil court. I don't know how to reconcile this with the apparent abundance of traumatic "pranks" on social media.
-
Comment on <deleted topic> in ~tech
overtowed Link ParentI agree. To spin it more negatively, these articles point the finger at practitioners and the field, lamenting its perceived decline, and give a pass to the systems and leadership that give UX its...I agree. To spin it more negatively, these articles point the finger at practitioners and the field, lamenting its perceived decline, and give a pass to the systems and leadership that give UX its marching orders. The buck can only stop at UX professionals if they're the ones in power - otherwise, look higher up the food chain and more broadly at the systems.
So under that lens I think the articles look somewhat toxic, but raising awareness is good.
-
Comment on <deleted topic> in ~tech
overtowed (edited )LinkSeems UX professionals who want non-exploitative work can either: work for a non-exploitative company hope for lucky assignments on products where the business wants actually good UX spend...Seems UX professionals who want non-exploitative work can either:
- work for a non-exploitative company
- hope for lucky assignments on products where the business wants actually good UX
- spend political capital for small-picture wins
- organize
Are there other options?
edit: There's also activism (for public policy, awareness, etc) like humanetech.com of the Social Dilemma Netflix documentary. Their podcast is heavy on the UX perspective from former Facebook employees, going into a lot of detail about existing exploitative practices and possible improvements.
-
Comment on Daily thread - United States 2021 transition of power - January 15 in ~news
overtowed Link Parent"freedom of speech is not freedom of reach""freedom of speech is not freedom of reach"
I've been making an ear training tool that's close to its first release. (source on GitHub) It supports Web MIDI and it's made with Svelte and SvelteKit. There's also a basic virtual piano.
The gameplay is simple - it plays a sequence of notes and then has the player play it back. If the sequence is two notes long, it trains individual intervals, and at longer sequences you get a combination of relative pitch and memory training. Currently a piano is the only input.
I just finished the basic features and I need to add the initial training challenges. The challenges it has right now are a quick proof of concept, and I'm not a musician so I'll need help to fill out the content. It's designed so that you can create custom challenges for your needs, and then share them if you want. Right now, sharing data means importing/exporting JSON that defines challenges and groups of challenges. There's no server behind it, it's just static files. After I spend more time with it, it might become clear how the challenges should be grouped, but right now it seems really open ended and subjective.