alcinnz

A browser developer posting mostly about how free software projects work, and occasionally about climate change.

Though I do enjoy german board games given an opponent.

Header picture is of Mordecai from Lackadaisy by Tracy Butler.

Pronouns: he/him/whatever
#noindex

Liberapay: https://liberapay.com/alcinnz
Rhapsode: https://rhapsode-web.org/
Homepage: https://adrian.geek.nz
Argonaut Constellation: https://argonaut-constellation.org/

Public Key

npub186lgefknfnkgychnfjancqnv4k4yrdzxal486lw0p80z6vmley9qgsnt98

NIP-05 Address

alcinnz_at_floss.social@momostr.pink

Profile Code

nprofile1qqsra05v5mf5emyzvte5eweuqfk2m2jpk3rwl6na0h8snh3dxdlujzsprpmhxue69uhhyetvv9ujumt0d4hhxarj9ecxjmntqyg8wumn8ghj7mn0wd68ytnhd9hx2rec070

Publishing to

relay.momostr.pink

nostr.wine

relay.damus.io

Author Public Key

npub186lgefknfnkgychnfjancqnv4k4yrdzxal486lw0p80z6vmley9qgsnt98

Show more details

Last Notes

2026-10-21 18:22:42 UTC

- reply

I make the note about definitions because I don't see that much in common between different arms of "AI", & I think we'd be better doing away with the term.
That said... Everything there falls under Machine Learning, which I consider to be a very mixed bag.

2026-10-21 17:56:56 UTC

A useful resource for resisting "AI": https://untrainable.org/en/
Note: This appears to use a broader definition of "AI", & discusses ways to resist different harmful arms of the field.

2026-10-13 19:30:02 UTC

I found this site humorous!
https://openfatigue.ai/

2026-10-08 16:20:12 UTC

RE: https://mstdn.social/@kathimmel/117400891287011442
Do any scots follow me?
#note1x2l…khwq

2026-09-27 16:59:15 UTC

Become worthless (to tech companies) - Cory Dransfeldt:
https://www.coryd.dev/posts/2026/become-worthless-to-tech-companies

2026-09-18 18:46:13 UTC

Digital Ocean has $3 million for DHH and nothing for CSS-Tricks - Keith:
https://keith.is/post/digitalocean-and-css-tricks

2026-08-22 20:37:11 UTC

- reply

To beat the cost of sorting Duan, et al's algorithm takes advantage of the fact that splitting a subnetwork can only give you smaller distances, so it can prepend those results to a priority queue.
You can read the paper for yourself: https://arxiv.org/abs/2504.17033
Previously extensive preprocessing was done to streetmaps searching for shortcuts to be taken at runtime. Duan, et al's algorithm doesn't need that. It efficiently handles any network without ever being slow than Dijkstra's Algorithm!
3/3

2026-08-22 20:30:44 UTC

- reply

The way it works involves recursively splitting the network up into smaller chunks by scouting ahead a finite distance, then stitching the results back together.
Once that subnetwork is small enough we can use our classic algorithms such as Dijkstra's, confident that we won't be considering too many intersections at once.
The trick is in stitching those results back together without cutting into the performance wins! For this we need clever datastructures.
2/3?

2026-08-22 20:24:50 UTC

About a year ago Duan, Mao, Mao, Shu, & Yin published a performance breakthrough in how to calculate a route between 2 points in a network. A breakthrough we've previously suspected to be impossible!
This is useful to me if I want to include street-maps in the feature-phone I'm planning to build, as it greatly simplifies the code I'd need to write in order for it to give you directions on how to get somewhere! Potentially enabling it to run on the cheap microcontroller I've ordered.
1/4?

2026-07-05 19:17:41 UTC

- reply

As long as you've got a GPU capable of running the same program involving multiplies & texture fetches across all pixels of your texture, which describes any 21st-century GPU,then convolution filters are super cheap! Unless you want larger or more shapely blurs...
Then again @npub1vdt…shc2 just mentioned a cheaper way to get shapely blurs.
Include a need for interpolation, matrix transforms, etc... And there's huge demand to include multiply-sum circuits in any GPU I might design!
2/2 Fin for today!

2026-07-05 19:07:01 UTC

Plenty of interesting visual effects & image analysis such as blurring can be achieved using "Convolution Filters", by computing weighted averages over the surrounding pixels of every pixel.
Often its cheaper to split this into 2 passes for the 2 dimensions of the image, if the algebra works out. Which it does for blurs.
Once we've got blurs we get shadows (blur an equivalent background shape), bloom (blur the lights), edges (subtract different blurs), & more!
1/2!

2026-06-12 21:40:06 UTC

"A lot of hype and noise is pressing us to extend the cowboy, wild-west phase of the software industry’s expansion. We should set that aside, and take ourselves seriously as a mature industry."
How building an HTML-first site doubled our users overnight - Alistair Davidson:
https://www.mohkohn.co.uk/writing/html-first/

2026-06-02 03:44:25 UTC

- reply

Hmmmm... A brief look at KaiOS... It appears to both be more & less than what I'm describing!

2026-06-02 03:31:49 UTC

Imagining the UI I'd want for a somewhat-modernized dumbphone...
I'd want something single-handed & power-efficient which promises a minimalist on-the-go featureset, without sacrificing ease of texting! That's my goal.
You'd operate it using a joystick (doubles as a D-Pad) & 2 additional buttons (typically for "Back" & "New"). The joystick would serve as a keyboard, when you move it in loops according to onscreen labels.
The UI would be menus of your texts, contacts, files, etc.
1/?

2026-05-15 20:18:29 UTC

- reply

'But “AI”, even more so than any other tech, is contingent on political clout. It’s what forces through data centres, lets companies infringe on copyright and violate software licences, renders them at least temporarily immune to all kinds of consumer protections and wrongful death suits, and results in the political collaboration where “AI” systems provide authoritarian states with “accountability sinks” and algorithmic cover for institutional racism.'
- Baldur Bjarnason

2026-05-15 20:16:58 UTC

The old world of tech is dying and the new cannot be born - Baldur Bjarnason:
https://www.baldurbjarnason.com/2026/the-old-world-of-tech-is-dying/
CW: Includes "AI" discussion.
But is *much* broader than that!

2026-05-13 17:24:52 UTC

CSS or BS - Keith Cirkel:
https://www.keithcirkel.co.uk/css-or-bs/
via https://roytang.net/links/2026/#8d524b46818a9925c5fa5288a17356ac

2026-04-24 17:29:19 UTC

Rejoice - Devine Lu Linvega:
https://wiki.xxiivv.com/site/rejoice.html

2026-04-18 22:38:59 UTC

What's the best numbers I can cite for how insane the level of investment is?
How much money has been sunk into LLMs? How much training data have they harvested? How much of the available hardware have they bought up?

2026-04-13 18:21:25 UTC

Over/Under #61 with Max Glenister - Lazybear:
https://lazybea.rs/ovr-061/

2026-04-06 18:59:42 UTC

- reply

And the thing is...
In this group I can easily counter talk of "its here anyways & not going way" with "so are cars, doesn't stop us from running our Drop One compaign!".
Then there's the issue of optics...

2026-04-06 18:55:00 UTC

Sigh... Today's the day for me to request that environmentalists I know don't pressure others into using "AI".
I'd prefer that they didn't use it themselves, but I don't have the patience for that argument. So I'm saving my spoons for fighting it on an institutional level!

2026-03-20 17:19:33 UTC

Universal stack operations using Uxn primitives - Wim Vanderbauwhede:
https://limited.systems/articles/universal-stack-operations-uxn/

2026-03-18 20:15:31 UTC

What sort of processor would we want for running the flash-memory wear-leveling firmware I spent the past few days describing?
Primarily we'd a decent of RAM in which to store virtual-addressing tables, freelists, tentative output, etc. And to make this fast... We'll need a cache hierarchy!
That is large memory is slow & fast memory is small, so to avoid the compromise include both! With the fast memory storing copies of the data in the large memory.
1/3?

2026-03-09 15:38:50 UTC

Accessibility people are a superpower - Jared Cunha:
https://jaredcunha.com/blog/accessibility-people-are-a-superpower
You Know What? Just Don’t Split Words into Letters - Adrian Roselli:
https://adrianroselli.com/2026/02/you-know-what-just-dont-split-words-into-letters.html
CW: "AI"-related footnote
The Ultimate Mobile Accessibility Resource Guide - Mark Steadman @ Dev.To:
https://blog.pope.tech/2026/02/10/does-your-navigation-need-an-aria-menu-probably-not/
Does your navigation need an ARIA menu? Probably not. - Elle Smith @ PopeTech:
https://blog.pope.tech/2026/02/10/does-your-navigation-need-an-aria-menu-probably-not/
Quick Tip: Do Not Replicate OS Behavior - Nat Tarnoff:
https://tarnoff.info/2026/02/05/quick-tip-do-not-replicate-os-behavior/

2026-02-20 19:30:44 UTC

Announcing Interop 2026 - WebKit:
https://webkit.org/blog/17818/announcing-interop-2026/

2026-02-11 19:43:10 UTC

- reply

I'm jealous! You get to meet Ink & Switch!

2026-01-10 20:53:31 UTC

I (like presumably many of you) don't like DRM & the "anti-circumvention" laws backing them up.
Though to be clear I don't oppose copyright, even if how long it lasts is dubious.
As far as I'm concerned granting major studios (individual artists don't have the sway to exercise DRM) say in the software I use to enjoy their work is far too much power for them to hold! Not to mention all the other (anti-repair) ways DRM has been abused.
1/2!

2025-12-16 18:29:40 UTC

It seems timely to restate this:
I want a future where the app-web & document-web splits in 2! Thus allowing for simpler browser engines.
Personally I don't particularly care about the app-web, but others do.
I consider the DOM an abomination of an API, inspired by Java-hype, that frustrates everyone.
If we discard the DOM (and in turn JS) whilst leaning on CSS's & HTML's built-in graceful degradation... I reckon a small team should be able to build a browser engine without much trouble!

2025-12-10 17:30:42 UTC

Pre-Teaching Programming Languages To Visually Impaired Students - Veroniiiica Lewis:
https://veroniiiica.com/pre-teaching-programming-vi-students/

2025-12-06 20:37:53 UTC

You know... If I were building a new backwards-incompatible version of HTML...
I'd remove self-closing tags!
Text can always be rendered in a way any given person can perceive it, so everything else should be treated as enhancements.

2025-12-03 17:13:28 UTC

Resisting software driven hardware obsolescence - Wim Vanderbauwhede:
https://limited.systems/articles/resisting-hardware-obsolescence/

2025-11-23 14:55:24 UTC

Open Source Has Too Many Parasocial Relationships - Justin Warren @ Pivot Nine:
https://pivotnine.com/blog/open-source-has-too-many-parasocial-relationships/

2025-11-22 19:42:18 UTC

With how many people appear to be yet again upset at Mozilla recently... I have to ask!
Which browser (engine) do you main?

2025-09-18 23:07:12 UTC

Oh, are we boycotting Disney now? Already on it! For years now.
I'm enjoying audio shows from the open web, and *loving* it.
If you want an alternative to the MCU I can't recommend The Red Panda Adventures highly enough! And the same studio also does Deck Gibson: Far Reach Commander, which could serve as a Star Wars alternative.
Tonight though... I'm listening to a small isolated crew orbiting a remote red dwarf star... Wolf359!
https://decoderringtheatre.com
https://wolf359.fm

2025-09-12 03:31:35 UTC

- reply

I was confused about which toot of mine you were speaking to when I replied. I had ATMEGA AVR (as found in Arduino & many of our peripherals) on mind. I was surprised by how low-powered I found that to be when I looked into it, but it delivers the I/O its market demands!
Nevertheless I agree with you!
We could in theory design, rather than evolve, a much nicer hardware platform for us to build upon. But that'd be a massive investment noone wants to make, when we already have *something*.

2025-09-10 18:58:14 UTC

- reply

Yeah, that could explain why such a 1990-ish processor design remains popular today! Its got solid I/O hardware!
Oh, wait. Thought you were responding to a different toot: https://floss.social/@alcinnz/115181162003263671

2025-09-10 17:51:15 UTC

ARM is great, ARM is terrible (and so is RISC-V) - John Goerzen:
https://changelog.complete.org/archives/10858-arm-is-great-arm-is-terrible-and-so-is-risc-v

2025-08-27 23:07:16 UTC

- reply

Yes! While I may not like these rules... Its even worse when they're enforced on the powerless & not the powerful!

2025-08-18 19:00:12 UTC

CSS-only scrollspy effect using scroll-marker-group and :target-current - Sara Soueidan:
https://www.sarasoueidan.com/blog/css-scrollspy/

2025-08-15 22:33:35 UTC

Programming language tournament Round 1 Match 3:
Nominations by: [@angelthorns](https://social.treehouse.systems/@angelthorns ) , @npub1z8r…mrq6 , @npub1cp3…39ww , & me!

2025-08-12 21:43:37 UTC

Who has some good, under-hyped Computer Science news from the past decade or so for me?
I knew a few, but I'd love to know more!

2025-08-07 21:23:58 UTC

What's the best programming language? Debate!
Yes, I want to stir up trouble...

2025-08-04 21:24:23 UTC

3 dynamics which I think has shaped most of computing history:
1. Worker power in computing lay not where it should with unions, but in how difficult the job traditionally was.
2. We reached "sufficiency" in the late 20'00s, early 2010s (when having fully-optimized CPU cores, we turned out attention to the real hardware bottlenecks) which didn't play well with point 1.
3. Hardware engineers are in general very divorced from the needs they're fulfilling. I know one!

2025-07-29 19:03:27 UTC

- reply

On the otherhand, by keeping my setup as trivial as I can... I find home-hosting very un-involved!
This seems to clash with some people's experience!

2025-07-11 22:25:22 UTC

I value understanding how technology works, not *just* for its own sake. But because knowledge is power!
It lets us fix things. It lets us change things!
It gives us confidence in imagining different futures!

2025-07-08 21:56:05 UTC

- reply

Incidentally the features which I think makes sufficiently fast computers:
1. SSDs - Entered laptop market 2006, Wikipedia suggests saturated market 2010 but my dad recollects a few years later.
2. Fast CPUs - fuzzy line to draw, but lets say early-20'00s, judging by Id Software's struggles I can't say 1990s.
3. Fully-programmable GPUs (for some artists & gamers) - XBox 360 was 2005, early.
So I propose late 20'00s, early 2010s as my "sufficiency" line!
2/1!

2025-07-08 21:46:53 UTC

Its my impression that the people designing & building computers are alienated enough from how computers are used that they don't really understand what's needed. They chase higher specs assuming that's empowering, without understanding what these numbers actually mean.
I reckon home computers have been sufficiently powerful enough for over a decade now, but predictions of sufficiency have led us astray before. So skepticism is perfectly understandable!

2025-06-30 20:49:56 UTC

- reply

My understanding is that ZLib is a core part of PNG compression?

2025-06-27 00:23:49 UTC

Writing a basic Linux device driver when you know nothing about Linux drivers or USB - crescentrose:
https://crescentro.se/posts/writing-drivers/

2025-06-04 16:39:15 UTC

Looks like a good directory! I've checked several random sites it listed!
https://url.town/

2025-05-29 19:37:55 UTC

Color in CSS - Aubrey Sambor @ A11y Talks:
https://a11ytalks.com/posts/2025-may (embedded YouTube)

2025-05-28 19:20:24 UTC

Its probably worth asking here: I'm looking for work!
I'm an incredibly knowledgable programmer interested in a wide range of technologies both new & old (though I have little patience for hype).
I've also got extensive volunteer experience in electronic repair!
I'm open to Nelson, NZ or remote work. I prefer part-time or environmental jobs.
#fedihire :boost_love:

2025-05-26 19:25:43 UTC

I'm now looking at https://github.com/drhelius/Gearsystem (which was promoted by [@Le_bottin_des_jeux_linux](https://floss.social/@Le_bottin_des_jeux_linux ) ) to figure out how to study it. I'll start discussing it tomorrow!

2025-05-24 20:33:24 UTC

- reply

Funny that...

2025-05-24 20:21:55 UTC

Building an IPv6-focused OpenBSD home router - Alex Haydock:
https://blog.infected.systems/posts/2024-12-07-building-an-ipv6-focused-openbsd-home-router/

2025-05-20 03:39:05 UTC

- reply

Presumably the argument is that blind people can enjoy audiobooks but not normal ebooks.
(Which ignores the fact that speech synthesis is old tech which I hope they were using already)

2025-05-04 21:02:54 UTC

Do you approve of the state of the software industry?

2025-04-15 19:20:20 UTC

We're in for a rough ride as an industry - Baldur Bjarnason:
https://www.baldurbjarnason.com/notes/2025/we-are-in-for-a-rough-ride/

2025-04-09 03:37:42 UTC

- reply

Talking of which... What's your favourite NES/Famicom emulator?
Which one should I study?
If I don't get an answer, I'll study the one [@Le_bottin_des_jeux_linux](https://floss.social/@Le_bottin_des_jeux_linux ) recently promoted...

2025-04-09 03:09:55 UTC

My upcoming threads:
On my Ink&Switch schedule (relating their writeups chronologically to my hardware-communicator hypothetical)... Tomorrow: schema versioning "Cambrian", next: exchanging crypto keys "BackChannel", then: richtext "PeriText", later: drawing dynamic models "CrossCut".
On my Linux From Scratch schedule.... Next couple days: Finish Groff (getting close), to be followed up with GRUB.
Between which I'll finish studying LibShumate, followed up with emulators for game consoles!

2025-04-03 16:39:38 UTC

Huion Inspiroy H610X - review on GNU/Linux - David Revoy:
https://www.davidrevoy.com/article1072/huion-inspiroy-h610x-review-on-gnulinux

2025-02-08 20:37:38 UTC

How much are you using products/services by US tech oligarchs? By Alphabet/Google, Amazon, Meta/Facebook, Apple, Microsoft, etc?
Don't worry about whether you're using their services indirectly, because you are.
I'd encourage you to not, since they've lined up for fascism with no respect for consent! But... Where we at?
(I'm sure my response will be biased, :boost_ok:)

2025-02-06 19:35:01 UTC

Its a widespread meme that modern computing is always changing, & we must keep pace with it.
For 1: To me it doesn't look like its changing near as rapidly as it was in, say, the 1990s! Lots of computing history happened back then!
For 2: I find the meme tends to inhibit deep thinking, acknowledging what has remained static or what *should* change. Instead it puts us on the treadmill chasing distractions.
Yes, computing is changing. But its not that fast! And not always in good ways!

2025-01-28 01:27:13 UTC

My principles for boosting generative AI discussion:
I avoid doing so unless its newsworthy that the author is taking a particular stance, or if I think the writing is valuable beyond the field of generative AI.
I will boost advice on how to avoid generative AI.
DeepSeek is newsworthy as is NVidia's stock-price drop, but most posts I see have hot-takes. Which I believe is premature. We'll see what comes from it!
(NOTE: I'm specifically saying "generative AI")

2025-01-25 16:53:01 UTC

FORTRAN, an “infantile disorder”? - Wim Vanderbauwhede:
https://wimvanderbauwhede.github.io/articles/dijkstra-fortran/

2024-12-28 04:17:43 UTC

In cleaning up my street-mapping threads for publication on my personal site...
I've just folded in [@RL_Dane](https://fosstodon.org/@RL_Dane ) 's quip that its no wonder Dijkstra hated GOTOs if he had a path-finding algorithm!

2024-12-26 19:55:30 UTC

I think I know which way this poll will lean...
AI features?

2024-12-26 17:09:35 UTC

The Cost of Self-Hosting - Kev Quirk:
https://kevquirk.com/blog/the-cost-of-self-hosting
I'm happily self-hosting, doesn't take me much effort. In large part because I insist on using the simplest tech, that doesn't take me much effort!

2024-12-22 19:15:18 UTC

Why aren't there more visual programming languages? (An ode to DRAKON) - Terence Eden:
https://shkspr.mobi/blog/2023/04/why-arent-there-more-visual-programming-languages-an-ode-to-drakon/
Starting from scratch - Samir Talwar "Functional Computer":
https://functional.computer/blog/starting-from-scratch

2024-12-16 20:45:40 UTC

- reply

Nice initiative!

2024-12-13 03:55:33 UTC

- reply

The mention of astrology & LLMs in the same toot brought my mind to @npub1hr7…jxyc 's blogpost providing a theory as to why so many people find them so convincing: https://softwarecrisis.dev/letters/llmentalist/

2024-12-09 20:33:32 UTC

- reply

Ouch, the article's painful for me... Reads to me like The Guardian doesn't know what they're talking about...

2024-12-07 15:00:36 UTC

- reply

Personally the way I phrase it is: Possible, but don't count on it!

2024-12-03 03:57:55 UTC

- reply

Will do!

2024-12-02 19:46:08 UTC

Question to help me shape some upcoming threads:
Say I'm bootstrapping from an assembly language to a (relatively) high-level language, without using any software I didn't write myself?
Any tips?
(I won't actually do this, I'll just describe how in detail!)

2024-11-30 19:07:06 UTC

I wonder... Could we call 2005 the year computing power reached sufficiency?
Is the only reason *basically* any of us need more computing power down to inefficient software?
I'm sure some of you will want to draw the line earlier, but I made sure to include fully-programmable GPUs (like the XBox 360's) in my cut-off!
This is before smartphones, but only just!

2024-11-29 19:33:06 UTC

- reply

Some "AI" related links...
Devs gaining little (if anything) from AI coding assistants - Grant Gross:
https://www.cio.com/article/3540579/devs-gaining-little-if-anything-from-ai-coding-assistants.html
Why We Don’t Know AI's True Water Footprint - Miranda Gabbott @ Tech Policy:
https://www.techpolicy.press/why-we-dont-know-ais-true-water-footprint/
TV Writers Found 139,000 of Their Scripts Trained AI - Elian Low "The Ankler":
https://theankler.com/p/tv-writers-scripts-ai-training-controversy-chatbot (Substack, partially-paywalled)
Labelers training AI say they're overworked, underpaid & exploited by big tech companies - Lesley Stahl @ CBS:
https://www.cbsnews.com/news/labelers-training-ai-say-theyre-overworked-underpaid-and-exploited-60-minutes-transcript/

2024-11-29 18:51:05 UTC

Links (29 November 2024) - Bladur Bjarnason:
https://www.baldurbjarnason.com/notes/2024/links-42/
Some links from it...
Care work - Mandy Brown "everything changes":
https://everythingchanges.us/blog/care-work/
Yes, progressive enhancement is a fucking moral argument - awfulwoman:
https://awfulwoman.com/posts/2016/progressive-enhancement/
Personally I don't even budge to suggest that progressive-enhancement makes webdev harder, I find it makes it easier!
Elements of creativity - Annie Mueller:
https://anniemueller.com/posts/elements-of-creativity

2024-11-14 18:36:47 UTC

- reply

A couple links from Own Your Web's newsletter...
Aggregating my distributed self - Miriam Eric Suzanne:
https://www.miriamsuzanne.com/2024/09/22/agregate-self/
Fighting for our web - Molly White:
https://www.citationneeded.news/fighting-for-our-web/
I love this talk!
Robb Knight's personal site:
https://rknight.me/

2024-11-14 18:24:58 UTC

Own Your Web – Issue 16: Under Construction - Own Your Web:
https://buttondown.com/ownyourweb/archive/issue-16/
Some links from it...
Flexible Typesetting is now yours, for free - Tim Brown:
https://tbrown.org/notes/2024/10/26/flexible-typesetting/
Under-construction signs - Jason Scott (sourced from Geocities via Internet Archive):
http://textfiles.com/underconstruction/
Blog Posts vs. Social Posts - Jim Nielsen:
https://blog.jim-nielsen.com/2024/blog-vs-social-posts/
Coming home - Mandy Brown "A Working Library":
https://aworkinglibrary.com/writing/coming-home

2024-11-11 20:56:17 UTC

Whose heard of boids?
Its an early (1986) technology to be branded as "AI", which has had huge influence over later hypecycles like object-oriented programming & self-driving cars.
It shows that with only a handful of rules & without any central coordination we can obtain fancy patterns! Resembling flocks of birds or schools of fish!

2024-10-28 23:10:40 UTC

- reply

Hang on, we were already using machine learning to do speech transcription before these LLMs. Why did we introduce hallucinations into them?

2024-10-26 19:50:38 UTC

Vector graphics allows you to describe the geometry of an image, or a letter ("glyph" from a font), such that it'll look just as sharp when you zoom in or out. I'm talking formats like PostScript, open/free-type fonts, or SVG.
Once the caller has provided all this data, a shape is rendered by using a "transformation matrix" to run the position of all its points through some algebraic formula. Then we discard erroneous details too fine to see.
1/5?

2024-10-24 18:03:54 UTC

As I stated yesterday: It strikes me that the devices we use for interacting with computers are as much defined by software as they are by hardware!
Optical mice? They compare how far to move the previous video frame it captured to match its new frame, using an algebraic shortcut known as an FFT. Still they have a lot of numbers to crunch in the mouse!
Touchscreens? They collapse the 2D runs of pixels your fingers obscure to extract touches it can clean & interpret as gestures.
1/2!

2024-10-22 21:11:55 UTC

- reply

To improve confidentiality we use cryptography. For this we find "trapdoor functions" in mathematics, using them to build ciphers which can't be reversed unless you know a certain "key". Modulo-exponentiation or XOR are common trapdoor functions.
To improve reliability we use "error detection" & "error correction", reducing the likelihood of interference producing valid data. A "parity bit" indicating whether there's an odd number of 1 bits is a basic "error detection" circuit.
3/4!

2024-10-22 20:57:27 UTC

- reply

Serializing data to a bytestream is relatively trivial, even without pulling in dependencies.
Parsing from a bytestream I find tedious, with a large portion of devtools existing to help you with it in some way or other. Its typically aided by a "scanner" to ease backtracking & "lexer" to abstract the raw bytes into "tokens".
To address a lack of bandwidth we use "compression". A simple strategy being to look up each byte in a "codebook", which characterizes retro graphics!
2/4?

2024-10-22 20:49:20 UTC

What does it take to transfer data between different pieces of software? Or between software & hardware?
To start: Most of the channels for doing so transmit the data byte by byte in sequence. If your data isn't trivial enough to already fit in that structure, you need a "parser" & "serializer" to make it fit!
Then there's the question: What if that channel doesn't enough bandwidth? Enough confidentiality? Enough reliability? We can fix it in software!
1/4?

2024-10-17 16:44:48 UTC

The Blockquote Element - Haydon Pickering:
https://heydonworks.com/article/the-blockquote-element/
I’ve Been Doing Blockquotes Wrong - Michelle Barker "CSS IRL":
https://css-irl.info/ive-been-doing-blockquotes-wrong/
Boosted by Robin Rendle "The Cascade":
https://csscade.com/i-ve-been-doing-blockquotes-wrong/

2024-10-14 17:05:35 UTC

ARIA DevTools - Mateusz Zieliński:
https://chromewebstore.google.com/detail/aria-devtools/dneemiigcbbgbdjlcdjjnianlikimpck (Chrome plugin)
A11y at Work: When Team Opinions Differ - Just Melanie From The Internet:
https://melsumner.github.io/a11y-at-work
Reminder – automatic accessibility testing can detect inaccessibility but can’t detect accessibility - Bogdan Cerovac:
https://cerovac.com/a11y/2024/10/reminder-automatic-accessibility-testing-can-detect-inaccessibility-but-cant-detect-accessibility/
57/30 - =HTML Accessibility:
https://html5accessibility.com/stuff/2024/10/07/57-30/
What’s the Difference Between HTML’s Dialog Element and Popovers? - Chris Coyier @ Frontend Masters:
https://frontendmasters.com/blog/whats-the-difference-between-htmls-dialog-element-and-popovers/

2024-10-13 17:19:57 UTC

Bufferless e-paper - Larry Bank:
https://bitbanksoftware.blogspot.com/2024/09/bufferless-e-paper.html

2024-10-09 19:06:08 UTC

The NES, like most home computers of its era, had a 6502*. As it turns out the simplest "Arithmetic Core" I could imagine is comparable in power to the 6502!
What I described had some more memory, & less instruction-decoding logic. Yet a full barrel-shifter to aid crypto! How exactly does it compare, & how'd I emulate the 6502 on it?
To start to handle the amount of RAM the NES had I'd need to enable our AC's "harvard-mode".
* Rather Ricoh 2A03, I'll come back to that...
1/?

2024-10-06 17:16:24 UTC

- reply

Context?
I like using Debian for servers & elementary for laptops/desktops!

2024-10-05 01:03:39 UTC

- reply

Yeah, the thing is I don't see *that much* advertising on the indie web...
The surface web seems to serve as its own echo chamber! How do people stand it anymore?

2024-10-04 00:51:53 UTC

- reply

They blogged "we don't care what you think, we're doubling down on advertising!"
https://blog.mozilla.org/en/mozilla/improving-online-advertising/

2024-10-03 23:32:50 UTC

Seeing the Mozilla news...
Yes, they've crossed a line for me! Amongst so many previous grumbles. Though I was already treating FireFox as an at-least-its-not-Chrome fallback.
Now that they've alienated their advocates, I doubt they'll stay around much longer. Who'll they advertise to?
I was hoping that they could remain a good stopgap until we could steer the web into a more sustainable direction, that hope is now dashed!
So what can we do?
1/2!

2024-09-30 00:11:35 UTC

- reply

I strongly suspect the relatively-recent resurgence of interest in physical/analog media is related to all mainstream e-stores shipping their media with DRM. Computing has broken our promises of ownership far too many times to be trusted!
(Don't blame me!)
Sure there's all that pseudoscience about analog having higher audio quality... Its alright, you can admit to enjoying the imperfections!

2024-09-29 20:57:38 UTC

An important perspective that helps me make sense of the push for DRM:
Its less about constraining your use of the media as much as it is about constraining which players make it to market. Thus shaping the technological reality laypeople operate within.
Its there to constrain me as a browserdev! Hence why I hate it.
That's how I can best make sense of the tautological remark "it keeps honest people honest"!

2024-09-28 21:18:53 UTC

- reply

In its normal mode as I've established our hypothetical Arithmetic Core there's only a couple KiBs of RAM, shared with code. Whereas its harvard-mode would have a full 128KiB (64Ki-words), which could be excessive for our needs!
At which point we may add a hashtable to the top of data-RAM to allow us to update existing entries in our min-heap once we've found a faster way to reach that intersection, instead of growing this min-heap.
Still... We can do better than Dijkstra's!
4/4 Fin today!

2024-09-28 21:08:39 UTC

- reply

Dequeueing an entry involves removing the 1st/root entry to return by replacing it with the last entry. Then we repeatedly replace this parent with its smallest child to reestablish the min-heap invariant.
Enqueueing involves adding it to the end of the array, then swapping this child with its parent until the min-heap's invariant's reestablished.
We can store the array's length at its 0'th index. And *maybe* we'd enable our Arithmetic Core's harvard-mode for the extra storage.
3/4!