Alt text is vast and varied. There’s no way to make it perfect, but there are plenty of things we can do better. Let me show you what I know about that. Buckle up — there’s a lot to cover.

Before we dive in

First, I’m not the almighty knower of everything. And everything is always changing. If you think I got something wrong, let me know. I’d rather learn something new than continue to fail.

Second, don’t forget about braille readers. They’re different than aural readers, which is what folks think of when you say “screen reader.”

The accessibility tree is the same source of truth for both aural and braille readers. The primary difference is that aural readers battle phonics, while braille readers battle real estate.

Aural readers often mispronounce words. But braille readers are literal, albeit with abbreviations and shorthand. (That’s a lengthy topic for another day.)

So when I mention screen readers in this article, I’m primarily talking about aural readers. If it’s braille specific, I’ll call it out.

Who needs alt text?

I know, I know. “It’s for blind folks using screen readers.” This is mostly true. People use screen readers for a variety of reasons. This includes visual impairments, neurodivergence, hearing impairments, and even motor skills.

Alt text also helps folks with no disabilities at all. “Situational accommodations” for scenarios like:

• Preferences: text-only browsers or privacy concerns
• Circumstances: slow/expensive bandwidth or low-end devices
• Technology push-back: missing images or VPNs/firewalls

Alt text is also not for some things. Lying liars will tell you they’re for search engines. Alt text is for people, not capitalism. Tell your friends.

Okay, let’s get into the good stuff…

Decision tree: the when, what, and how

So here’s that ocean. When to use alt text, what to describe, and how to add it.

When to use alt text

First, we have to ask a simple question. Is what we’re describing functional, informative, or decorative?

• Functional images invoke an action. So they need alt text. A common example is social media interaction icons.
• Informative images convey information. They, too, need alt text. One example is a book cover.
• Decorative images are for visual effect. So alt text isn’t needed. One of the most common examples is a button with both a text label and an icon.

When we don’t need alt text, we still need an empty alt attribute. And “empty” means nothing between the value quotes; not even a space.

What to describe

The key considerations are context, efficiency, and avoiding redundancy.

Let’s say we have a gallery of a dog named Earl. It includes an overview that describes him, including his color and breed. With that context we can focus on what’s happening in the photos, instead of repeating his physical description in every one of them.

<img alt="Earl, our golden chihuahua dachshund mix dog, standing on freshly-cut logs in our yard, looking off into the distance.">

Other images might have visual characteristics that don’t need to be described, like this image from an article about ink traps.

<img alt="Source letterforms for the capital N and M, showing unusually-large gaps in their crotches. A printed version shows the correct letterform with ink bleed.">

The label and arrows help visually, but have less relevance as alt text.

An important part of context is the role of an image. Is it there to support content or is it a focal point? Take for example this photo of The Runaways, circa 1977, taken by Michael Ochs.

In the middle of an article, it may be supporting the content. If so, a description that’s too detailed may not be necessary, and can even slow down a reader. We can describe the baseline, then move on.

<img alt="The full band, lying down on a beach, facing the camera with their middle fingers raised.">

If it’s part of Michael’s portfolio, it’s a focal point. The image is the content. It’s showcasing his work, where composition and details are important. This extended alt text tells a story. It amplifies the purpose of the art.

<img alt="Looking head-on at all five members of The Runaways, lying down on a sunny beach with their middle fingers raised. They are facing the camera with the ocean at their backs. All of them are wearing bikinis except for Joan, who is wearing a black wetsuit. Their expressions are wry and rebellious.">

Finally, if there is text in an image, we have to decide what to repeat. Not all text in an image needs repeating in alt text. Lying liars will tell you otherwise.

Take for example, these two photos of signs with reader boards. The board on the top is the entire point of the image. So it should be in the alt text. But the one on the bottom isn’t about the job opening; it’s about the sign’s beautiful design. So that text doesn’t need to be included.

<img alt="Store marquee that reads “if it’s in stock, we have it.”">

<img alt="Colorful neon sign on a painted metal canvas with metal piercings on each side. The name “United Battery” is stacked uniformly in a traditional typeface below a lightning bolt.">

How to add it

Here’s how to add alt text across different image types.

Simple images like the following are easy.

We just roll out the alt attribute because the graffiti is all you need to know:

<img alt="Graffiti that reads “Molotovs melt ice.”">

Detailed images need more work. Such as infographics like this:

Explaining this image in an alt attribute would be massive. Fortunately there’s a better option. Lying liars will tell you that longdesc is the answer. Nope. It’s deprecated and poorly supported by screen readers. Not to mention it’s shitty UX. And no one likes shitty UX.

The answer is a details tag:

<details>
 <summary>Infographic showing how gerrymandering cheats a majority out of a win.</summary>
 <p>Graphic titled “Gerrymandering explained. Two different ways to divide 50 people into 5 districts.” There are three 5 by 10 grids of blue/red squares: 20 red and 30 blue. The third grid shows how creative line drawing can create a victory for the minority of overall blocks. The first, labeled “aggregate votes,” shows them as a single group, concluding with “60% blue, 40%, blue wins.” The second, labeled “compact but unfair,” shows them grouped into 5 equal blocks comprising 4 red squares and 6 blue squares, concluding with “5 blue districts, 0 red districts, blue wins.” The final, labeled “manipulated and inaccurate,” shows them grouped into 5 calculated groups to manipulate the majority, concluding with “2 blue districts, 3 red districts, red wins.” The caption reads “adapted from Washington Post’s adaptation of Stephen Nass’ illustration with a CC license (https://commons.wikimedia.org/wiki/File:How_to_Steal_an_Election_-_Gerrymandering.svg).”</p>
</details>

First, it separates along description into two parts: a summary and its details. And the details can be many paragraphs or even an entire data table. Second, it has built in disclosure functionality.

Some of you might be thinking, “but a details tag is visible on the page.” How observant of you. And thank you for questioning my logic. With a few lines of CSS you can hide it visually while retaining its presence for screen readers.

A blind accessibility engineer, James Scholes, suggested this to me. We had a discussion following a poll I conducted on Mastodon for folks who use screen readers daily.

Another option for long descriptions is figcaption. Whereby the details tag has two parts, the figcaption is a single container for what is literally a caption intended to be visible aside the image. It can be used in place of or in harmony with a description in the alt attribute.

If the figcaption is descriptive, we include an empty alt attribute and an aria-describedby attribute to the image tag. This tells screen readers where the description is.

<figure>
 <img src="earl.jpg" alt="" aria-describedby="caption">
 <figcaption id="caption">Young gold chihuahua, looking intently towards the camera with his ears perked up.</figcaption>
</figure>

If the figcaption doesn’t describe the image, we supplement the figcaption with an alt attribute for the description. And we no longer need aria-describedby.

<figure>
 <img src="earl.jpg" alt="Young gold chihuahua, looking intently into the camera with his ears perked up.">
 <figcaption>My boi, Earl.</figcaption>
</figure>

Grouped images may be considered a single image. Take for example a five-star rating system.

This is a group of similar images intended to convey a meaning as an aggregate. So we don’t need to describe each one. The W3C’s recommendation works perfectly:

<img src="star.svg" alt="4 out of 5 stars">
<img src="star.svg" alt="">
<img src="star.svg" alt="">
<img src="star.svg" alt="">
<img src="star-empty.svg" alt="">

It’s not vital to force a screen reader to hit every image. So the aggregate description is read once, and the others are skipped over.

If you have a group of images, they’re probably already in an HTML group of some sort. So you can take a different approach:

<div role="image" aria-label="4 out of 5 stars">
 <img src="star.svg" alt="">
 <img src="star.svg" alt="">
 <img src="star.svg" alt="">
 <img src="star.svg" alt="">
 <img src="star-empty.svg" alt="">
</div>

Using role="image" tells a screen reader that the div should be treated as an image. And the aria-label attribute tells the screen reader what to say.

SVG images have their own system entirely. If they’re linked using an img tag, you can reference the aforementioned methods. But if the SVG code is embedded you’ll need three components: a role attribute, a title tag, and a desc tag.

<svg role="image">
 <title>Simple summary</title>
 <desc>Longer, detailed description can go here with more details.</desc>
 <path d="…"/>
</svg>

As with the details tag, we have a description broken into two parts: a summary (title tag) and its details (desc) tag. But the desc tag isn’t required. If the alt text only needs to be simple, it can all be placed in the title tag.

If an embedded SVG doesn’t need alt text, it simply needs to be hidden from screen readers with aria-hidden=“true”.

A final note on the how. Use smart quotes instead of straight quotes. It’s generally good practice because in formal writing smart/straight quotes do not mean the same thing. But it’s especially helpful for alt text because in code the results can be unpredictable.

<img alt=""Straight quotes""> 🚫

<img alt="“Smart quotes”"> ✅

Straight quotes nested inside of straight quotes can result in delimiter collision. Who knows where the alt text begins/ends? (My syntax highlighter is even confused by it in the example above.)

Be intentional and formal. Just use smart quotes.

Composition style and depth

The two key considerations here are avoiding verbosity and knowing your audience. And there are four things you can do to help with this.

Use proper punctuation

Think about alt text like proper writing. Commas and periods break up content meaningfully. This helps with longer alt text and keeps alt text segregated from other content. Just remember that input equals output.

<img alt="hello friend it’s lovely to see you how are you"> 🚫

<img alt="Hello, friend. It’s lovely to see you. How are you?"> ✅

Front load important information

This is one of the most important parts of composition. Especially when the alt text is long. Take this photo for example:

There’s a lot going on here. We need to tell the right story. Let’s break some things down.

Is that can important? No. How about the flare? Maybe, but it’s not vital. And that dude in the back? Not really. But it’s probably noteworthy. So how do we put these pieces together to tell the right story? Like this:

<img alt="Trash can flying through the air, about to land on cops with riot gear. They look disorganized and there is a flare on the ground behind them. Ironically, a man directly behind them is seemingly disinterested, looking off into the distance.">

That tells the right story in the right order:

1. Primary: “trash can flying through the air, about to land on cops with riot gear.”
2. Secondary: “they look disorganized and there is a flare on the ground behind them.”
3. Tertiary: “ironically, a man directly behind them is seemingly disinterested, looking off into the distance.”

We lead in with primary info. A reader can get the gist of this image with only this sentence. The secondary info helps shape the narrative, and the tertiary info is only an aside. Those with the gift of sight are able to scan the photo to glean lots of different information. Even the parts that aren’t critical. Tertiary info can give everyone else an opportunity to do the same.

Write for your audience

Knowing who you’re writing for is important. Are they academic? Tech savvy? Knowledgable about the topic at hand? What’s their demographic? You don’t need to know everything, but give it a moment of thought.

The French protest image is a perfect specimen for this. It was taken in Bordeaux, France. When I posted this image on Mastodon, it was for a general audience. And even though I have followers around the globe, most speak English. So I used the word “cops” in my alt text.

But if my primary audience was English-speaking folks in France, I would have used the word “flics.” That’s their slang equivalent for cops. This context switching isn’t critical, but it’s helpful.

Be concise

Avoid non-essential visual information.

<img alt="Quote from Jessica Williams that reads “I like my government how I like my coffee — not in my vagina.”">

In this image we don’t need to mention the brick wall or the diagonal color shape. And we don’t need to begin with “meme of” or “graphic of.” We keep it simple because the quote is all that matters here.

Also, the transcriptions don’t have to be literal. In the image, the order of quote then citation makes sense visually. In the alt text, the order makes sense compositionally.

Phonetics

This is important: no two screen readers sound alike. Even within the same dialect in the same operating system. And no screen reader has perfect pronunciation. But don’t try to hack or override pronunciation. Instead, be mindful and consider alternative ways to compose information.

<img alt="Bag with IV fluids hanging from a stand."> 🚫

<img alt="Bag with intravenous fluids hanging from a stand."> ✅

In this image, the acronym “IV” can be pronounced “four” or “roman four.” It’s tempting to hack the system by swapping the acronym with “eye vee,” for phonetics. But this creates nonsense for folks using braille readers and for those not using screen readers at all.

In this case, using full words instead of the acronym is the best alternative. “Intravenous fluids” reads correctly for everyone.

And this context applies to dimensions, units of measurement, and locations:

<img alt="8′2″L x 12′4″W."> 🚫
<img alt="8 feet, 2 inches long, by 12 feet, 4 inches wide."> ✅

<img alt="5GB."> 🚫
<img alt="5 gigabytes."> ✅

<img alt="100lbs."> 🚫
<img alt="100 pounds."> ✅

<img alt="123 Abolish Ice Dr., St. Paul, MN."> 🚫
<img alt="123 Abolish Ice Drive, Saint Paul, Minnesota."> ✅

We can avoid ambiguity and mispronunciation simply by using whole words instead of symbols and abbreviations.

Social Media

Both practitioners and common folks publish content on social networks. So not everyone gets it right. But as professionals, it’s our duty to use our skills to make sure we do.

Don’t blindly follow “a11y” trends

LinkedIn is filled with these trends. Herd mentality that’s performative for clout. And it hurts people who need better from us.

For example, LinkedIn has an alt-text input. They force you through it when posting an image. But there’s a trend of putting alt text in posts instead of using that tool. Or worse, replicating alt text in both places. Like the following example:

That passage was repeated three times. Three times! As post content, fake alt text, and real alt text. Imagine having to use a screen reader every day, and dealing with this nonsense. And this was posted by a so-called “accessibility expert.”

This is so irresponsible that it makes me angry.

Don’t put hashtags mid-sentence

Hashtags belong at the end of a post, not throughout. There are two important reasons for this. First, they’re converted to links, which are announced as such by screen readers. Second, the symbol itself is read aloud.

Here’s an example of what that sounds like:

Besides hashtag littering, this person also used redundant alt text in the post. Accessibility problems twofold. And it, too, was posted by another “accessibility expert.”

Hashtag formats

On the topic of accessible hashtags…

(I know, I know. This isn’t about alt text. But it’s important. Calm down.)

Lowercase hashtags are mispronounced by screen readers:

• #uxdesign can be read as “ucks design”
• #uidesign can be read as “ooey design”
• #webaccessibility can be read as “weeb accessibility”

CamelCase* hashtags are almost always pronounced correctly:

• #UXDesign will be read as “UX design”
• #UIDesign will be read as “UI design”
• #WebAccessibility will be read as “web accessibility”

And they improve readability for everyone. Especially folks with Dyslexia and other reading challenges.

*Before you invite me to a cage match about camel case vs Pascal case, there’s some history you might want to know about.

Video also needs alt text

There are three different kinds of alt text for video: transcripts, descriptions, and captions. Each serves a different purpose. But they all benefit everyone. Even folks without disabilities. The situational accommodations noted earlier.

Captions and descriptions

For most video, use captions and descriptions. Each helps in a different way, and they support each other.

Videos are generally long-form content, rich with visuals and audio. For sighted people, a title/thumbnail are quick scans for this decision point. Adding a description helps folks with screen readers decide if they want to watch it. It’s baseline alt text.

Captions are alt text of audio in real time. The target is folks with hearing impairments. But this is where those situational accommodations come into play. Captions are a gift for non-native speakers and people in quiet spaces.

Transcripts

Use transcripts for full recaps of video. They’re helpful for people with visual impairments who don’t want to listen to an entire video.

They are either formatted with timestamps or speaker labels, depending on the context:

00:00:50.860 -> 00:00:53.420
Well, maybe not every style. Ref?
00:00:54.740 -> 00:00:55.840
Aw, come on!

Speaker 1: Well, maybe not every style. Ref?
Speaker 2: Aw, come on!

Most video platforms will auto-caption for you (with an edit function) and also let you upload your own transcript files. Some will even translate for you.

A note on live video

Auto-generated captions and/or a sign language interpreter is ideal. But this can be a logistical challenge. It’s costly and there are technical hurdles that may not be available for some publishers.

If it’s simply not possible, commit to publishing an archived version afterward, complete with captions and a transcript. Announce this intention and make accommodations for folks who need captions/transcripts. Then they can forego the live video to watch the archived version at a later time.

In closing

That was a lot. If you’re still here, you deserve a cookie. Now go write some alt text that doesn’t suck. You can do it. I believe in you.

The curated list of things that bring me joy. Plus bytes on design, code, accessibility, privacy, and tools you might find useful. Available via RSS and email. This volume’s goodies…

At the moment

I took a mental-heath break over the past month. I produced almost nothing. Then I went to the coast with my family. Now I’m ready to get back to my giant list. More on my now page.

Mullvad just released an official long-form response to the discovery that their co-founder Daniel Berntsson supporting a fascist party in Sweden. Their response doesn’t help, and neither does Daniel’s essay on why he donated to them.

In Mullvad’s post they consign to being a “political company” who is “careful in not taking political stances outside of our narrow areas of focus. Because we want to cooperate with everyone.” They follow this up by noting “protecting ideas one does not agree with is the only way to have a neutral field where truth can be found.”

The rest of the post is just an argument about how Mullvad fights for privacy. Which they do. And they’re good at it. But dismissing fascist ideals because “societies need a great diversity of ideas” is unequivocally dangerous. Especially when they align with Nazi ideals.

Daniel’s excuse

In his post about why he supports the Örebro Party, he digs a deeper hole by citing reasons that align with how the German Nazis won people over.

After a weak, ambiguous analogy (good luck deciphering it), he cites the reasons the ÖP is special and warrants support. The most nefarious being nationalism with a racist, xenophobic accusation that immigrants are to blame for society’s problems.

This is a core ideology that drives the ÖP economic policies: “class-conscious populism,” as a way to unite “productive” classes against the “transferiat” (people who need social welfare or who hold jobs in “made-up services”). And most of Daniel’s post is about this “deserves encouragement.”

The most offensive part of the ÖP “transferiat” fallacy is the xenophobia. They blame the problem on immigrants (sound familiar?) and promote repatriation policies to deport them. Party leader Markus Allard (who Daniel champions) is cited as saying “Sweden belongs to the Swedes. We have to make sure that we take care of our own damn people and we must deport these damn parasites who sit and live at our expense.”

We have to make sure that we take care of our own damn people and we must deport these damn parasites who sit and live at our expense.

Daniel also spends a lot of time praising a Markus Allard and Malcom Kyeyune podcast. In that podcast Markus stated his opinion on immigrants and deportation: “they will also be forced to leave, even if they are born in Sweden, because they have no natural connection to Sweden. They are not Swedish.”

The Nazi Party began as the German Worker’s Party (DAP). That was their name when Adolf Hitler joined them. He was attracted to the ideals of the DAP founder, Anton Drexler. Those ideals being anti-Semitism, nationalism, anti-capitalism, and anti-Marxism.

Mullvad’s excuse

The Nazis blamed the Jew. ÖP blames the Somalis. Nationalism and xenophobia, founded on racism. And this is the fundamental problem with Mullvad’s excuse. They’re trying to downplay Daniel’s fascism by promoting “diversity of ideas” with “rational debate.”

Debate is okay when we’re talking about pizza toppings. Debate is not okay when one side is trying to inflict suffering and violence on others because of racist lines drawn on a map.

We cannot separate ethics from technology. Ever. Under any circumstances. And we certainly cannot give our money to organizations who support violent ideals like fascism. Mullvad is saying “we have a privacy product and you should support us for that reason because ethics are irrelevant.” Yet the dollars we spend with Mullvad have enabled a fascist party in Sweden to evolve from largely unknown to prime time. There is a direct connection.

Conclusion

I’m gonna let Daniel put a bow on this essay with his own words:

“Do I like most of ÖP’s policies? No. Are they fit to run a country? No. (Neither are the other parties.) Do I want ÖP to get a lot of power? No. Will giving them a leg up in the competition with the other parties, given the local political circumstances, be a net positive? Maybe. Probably. Hopefully.”

Yet he solely contributed 72% of their total 2025 income. Enough for them to become a national party.

Tossing Mullvad and getting my refund was clearly the right call. Mullvad and Daniel tried to clean things up, but they failed. None of this makes me happy. In fact, it all breaks my heart. But the validation that they’re not interested in course correction is a spoon full of sugar.

The curated list of things that bring me joy. Plus bytes on design, code, accessibility, privacy, and tools you might find useful. Available via RSS and email. This volume’s goodies…

At the moment

I’m really trying to focus on my ham radio license test. It’s so hard, but I’m doing my best. I’ve been listening to a lot of 80s/90s death metal, but also DJ Krush’s new record. More on my now page.

Today I read something about one of Mullvad’s founders, Daniel Berntsson, donating to a fascist party in Sweden. Because I have been championing for Mullvad for a long time, I thought I’d do some research of my own. Here’s what I learned.

1. Daniel did indeed donate to the Örebro Party. In fact, his donation was 72% of their total 2025 income.
2. Lots of people on Mastodon called them out. They replied to most if not all of those posts with this same message.
3. The Örebro Party indeed supports racist, xenophobic ideologies and policies (while pretending to be a left-leaning/socialist party).

Mullvad’s canned response that they posted countless times:

Mullvad is a political company fighting for free speech, free information and privacy, with two equal co-founders, co-owners and co-CEOs who fundamentally disagree on many issues. Daniel’s donation to a political party is private and not part of Mullvad’s mission. We protect the right to express and access views we disagree with. We welcome anyone sharing these core values, whatever their other opinions. We are happy to refund others who don’t, where we can.

The key part to call out is their insistence that they are a company comprising “two equal co-founders, co-owners and co-CEOs who fundamentally disagree on many issues.” And that his donation is “private and not part of Mullvad’s mission.”

The problem with this is that fascism promotes violence against others. Many privacy advocates claim “free speech” should be protected, even when it’s hate speech. Including Mullvad, who notes this in their response. They’re all-in on privacy-first.

But hate speech isn’t free speech. It demands violence against marginalized people. There is no “both sides” to it. It’s not okay to “fundamentally disagree” on it. If you are a company who earns income, and that income is funneled to a fascist political party, it is absolutely part of your mission. Even if it’s turning a blind eye to a teammate with whom you disagree.

Örebro party policies

You can read about the fascist policies of the party on their website (which I’m not linking to, because they don’t deserve our traffic) or on Wikipedia. But here are the key problem areas.

1. Remigration is the crown jewel of their platform. They actively campaign to deport all people their founder Markus Allard says are not Swedish.“They will also be forced to leave, even if they are born in Sweden, because they have no natural connection to Sweden. They are not Swedish.”
2. They work to unite “productive” classes of society against people who “suck money out of the state” by receiving welfare help or work in “made-up services.” They even coined a term for it because it means so much to them, which I’m not repeating here. Allard claims most of these people are immigrants and that they are responsible for most violent crime. This is, of course, a favorite trope among fascists.
3. Even beyond fascism, they are proponents of international violence by means of war. But not just any war. War against countries who are protected by NATO, typically because they have resources that support dreams of totalitarianism and world domination. Markus claims that Sweden “cannot trust or rely on NATO and that the country should instead seek to massively increase its weapons production and also become a major arms exporter.”

My takeaway on this

I’ve been a vocal supporter of Mullvad for a long time. Especially because they have one of the most reputable VPN services around. And because I thought that their leadership was unlike Proton’s CEO, Andy Yen, who is a notorious fascist boot-licker, citing his support for America’s current fascist totalitarian. So it’s heart-breaking to learn that Daniel Berntsson is taking Mullvad down the same path.

As seen, Mullvad is offering refunds to folks who want them. This is telling. They are doubling down in support of funneling money to fascists. My money will no longer be party to that. I will be requesting my refund and moving to a new VPN.

Why can’t we have anything nice?

The curated list of things that bring me joy. Plus bytes on design, code, accessibility, privacy, and tools you might find useful. Available via RSS and email. This volume’s goodies…

At the moment

I’m authoring on a companion article for my accessibility talk. I’m also working overtime on building new components for a design system. More on my now page.

The weekly curated list of things that bring me joy. Plus bytes on design, code, accessibility, privacy, and tools you might find useful. Available via RSS and email. This week’s goodies…

At the moment

This week I’ve been busy reviewing portfolios and resumes. I offered to help folks find work, and a wall of people responded. I’m honored that people trust me to review their work. I’m also continuing to work on my upcoming accessibility lecture for the Portland Accessibility and UX Meetup. More on my now page.

The weekly curated list of things that bring me joy. Plus bytes on design, code, accessibility, privacy, and tools you might find useful. Available via RSS and email. This week’s goodies…

At the moment

This week has kept me super busy with design system documentation. I’m also really close to completing my music app for macOS. More on my now page.

The weekly curated list of things that bring me joy. Plus bytes on design, code, accessibility, privacy, and tools you might find useful. Available by email and RSS. This week’s goodies…

At the moment

I’ve been designing and coding so much lately. Good thing I love it. I’m reading The Second Sex, by Simone de Beauvoir. It’s kind of blowing my mind. More on my now page.

The weekly curated list of things that bring me joy. Plus bytes on design, code, accessibility, privacy, and tools you might find useful. Available by email and RSS. This week’s goodies…

Tickle me pink

• Otoboke Beaver is one of my favorite bands. They just released a new EP, “Is the New Album Out Yet?.” It includes a pressing on 4-inch vinyl (which I pre-ordered). Huzzah!
• The Computer History Museum is a cool resource. The oral histories from the museum have been curated on f0lkl0r3.dev, complete with light mode, reminiscent of a terminal.
• Japanese artist Manabu Kosaka makes intricate replicas of machines and electronics out of paper. They are surreal.
• I just won a Rocket League tournament. 9 matches, 2 hours, 1 winner. And I got MVP.
• Meteorites called angrites likely come from a protoplanet that was catastrophically destroyed in the early solar system. Its fragments are falling to Earth.

Recent Mastodon musings

• Portland springs are always an aesthetic delight.
• Poulain de la Barre, laying down gender equality 300 years ago.
• Christopher Rothko’s No. 5 (Untitled) at the Portland Art Museum.

Design bytes

• Swiss designer Wolfgang Weingart was part of an emerging group of designers criticized for “typographic barbarism.” For many of us, their style remains an inspiring era of design. Letterform Archive profiled his series of covers for Typografische Monatsblätter (TM).

Code bytes

• Christian Hofstede-Kuhn created a list of shell tips that “make life easier and save your sanity.” While many of these are familiar, I learned some stellar tricks.
• This week, for the first time ever, I had a use for a shadow DOM. If you ever need it, you’ll be glad to have it.
• Recently I got Linux installed on my old MacBook. When I sat down to build my accessibility app for Linux, I wanted a way to move information between my Linux/macOS MacBooks. Netcat was the solution. What a great command-line tool.

Accessibility insights

• WebAIM recently published their 2026 accessibility evaluation of the top 1 Million websites. Unfortunately, this year’s news is bad. After gradual improvements in recent years, the latest analysis shows notable increases in errors/failures. They cite the likely source as frameworks and AI “vibe coding.” I concur.
• Are you in control of your focus indicators? Sara Soueidan is, and she has an extensive article about it, to help you make sure you do, too.

Privacy lockdown

• Mastodon was recently awarded a giant service agreement by the Sovereign Tech Fund. This will bring a number of needed improvements to the social media platform. The most exciting one is E2EE for direct messaging. This is long overdue and will massively improve the platform overall.
• A California privacy audit by webXray shows that Google, Microsoft, Meta all track you, even when you opt-out. The problem is they don’t care, because their data harvest gains far outweigh the fines.
• My friend Em published a comprehensive guide on keeping yourself safe if you’re an activist. It’s the Activism Toolbox, published on Privacy Guides.

Tools and platforms

• Wakamai Fondue is an incredible tool that parses variable fonts for you. It’s a mainstay for me. They just published a massive update with lots of improvements.
• I’m finding ByteStash to be a savior. It enables me to catalog code snippets, then search/filter them. I’m searching my existing projects less often for “that one thing” I did one time. Instead, I’m retrieving it from my ByteStash (which I’m self-hosting).
• It’s not FOSS, but NocoDB (as in no-code database) is the best replacement I’ve found for Airtable and Baserow. Both of the latter have let me down. So far, NocoDB is easier to use, more intuitive, and faster by far.

At the moment

Having finished my first edition of Nano Bytes, my focus is now on finishing my macOS music player app and porting my Ratio app to Linux. More on my now page.