“Flutter 3.38.0: How an accessibility upgrade broke Android headings”
abra.ai/blog/flutter-3-38-0-ho…

PR that should fix it:
github.com/flutter/flutter/pul…

from Jan Jaap de Groot (@jjdg.bsky.social).

reshared this

AI shopping assistants now pick sizes and fill carts for you.

WebExpo 2023 speaker @tink asked a bold question. If agents browse for us, what even is the web now?

She zoomed in on what this means for accessibility and why the next UX challenge may involve non-human users.
tetralogical.com/blog/2025/08/…

Krátce po jmenování Andreje Babiše premiérem došlo v Ústí nad Labem k tragédii. Pětapadesátiletý muž se stal první obětí nové éry blahobytu, když zemřel na následky udušení poté, co mu dokřupava upečený holub vletěl rovnou do huby a zasekl se mu v krku. az247.cz/muz-v-usti-s...

Muž v Ústí se udusil, když mu ...

Was just thinking realistically, are you going to pull out your DotPad, Monarch, other larger graphic Braille display at the airport to check your Uber ride status? Or look at a transit /navigation app at directions? Was looking at the dimensions for the DotPad, the smallest of them all, and it's still the size of an iPad 11-inch, slightly larger and thicker. Good to pull out at your gate or in-flight, bad when walking around as it's too large.
Some of you have asked if Cadence uses a membrane or film over dots, and the answer is no. I'm told this was a deliberate choice, to make the cells feel more crisp and sharp. Indeed, their rigidity reminds me of Orbit cells, but you can push down on them and when you lift your finger they spring back up. While you lose out on uninterrupted graphics (since Cadence uses Braille spacing, meaning there's a gap every 2 columns,) you get a pocketable reading device that's prioritized for that task. I was thinking to myself how most of these companies approached the multiline Braille concept purely from the graphic side, not so much making reading portable. For many of these companies Braille is another product alongside magnifiers or screen readers, or big-name medical groups sponsor them. (Dot Inc is an exception here.) So they didn't have the creativity to think, "what if we made smaller versions of this? 4 lines of 12 or 20?" Still find this a bit astounding and sad, until Cadence came along. They took a lost opportunity and made it real.
This entry was edited (1 week ago)

heads-up ArcaneChat users:

the previous poll is over, the winner was:
"allow to set subject in classic email chats" next poll 👇

📊 POLL: What should be done next?

you can influence what is the next thing that should happen in the ArcaneChat ecosystem!

#ArcaneChat supporters can vote here:

ko-fi.com/polls/ArcaneChat-Wha…

🔮choose your own adventure! ✨️

This entry was edited (1 week ago)

"Do not introduce more ARIA into your code in an effort to please the knowledge ghouls of OpenAI and their new browser. Although you wouldn't know it from the bumpf OpenAI published—It is not needed." @SteveFaulkner on the HTMHell advent calendar: htmhell.dev/adventcalendar/202….

#фалыстыннаш
я вот тут ругаю европу за то что они за хамас. а ведь они по факту заткнули дыру утечки мозгов. Я прямо знаю не одного и не двух прямо топовых чуваков которые не приняли предложения в ес из за этой вот всей ебанины.

так что возможно пользы от этого Израилю больше чем вреда.

health issues... again

Sensitive content

in reply to Estelle

health issues... again

Sensitive content

Informal Masto poll: What do y’all think of the name Nayshawn? Pronounced ‘Nay Shawn.’ For context, I was telling my s.o earlier that if we were to have a kid, I would want them to have a name that doesn’t get them picked on in school. And, he was like, if it’s a boy, we should name him Nayshawn. I think this name is ridiculous. He thinks it’s kick-ass. What do you think, fediverse folks?

#deltachat PSAs regarding the last 24 hours:

1) If you messaged someone about 20-30 hours ago, and did not get an expected response/receipt better "resend" (long-tap or right-click the message to resend). Unfortunately, not all queued messages at the crashing nine.testrun.org server made it to the new one.

2) 2.33+ app releases are rolling out, supporting using "multiple relays" so that failing relays (like happened above) will not cause lost messages anymore. See support.delta.chat/t/how-to-mi…

in reply to devSJR

there is a blog post coming. bear with us :)

but roughly the current first multi-relay release introduces the ability to have more than one relay associated to your chat profile. You can choose what is your primary device, but will receive messages also on all secondaries. This is just an intermediate step towards another release where the outcome will be: if any relay works, the message will get through. single relays failing don't matter.

This entry was edited (1 week ago)
in reply to devSJR

@devSJR yes, that's the plan that you can exit from gmail or other big providers if you started with them. It's just that the last weeks were a lot of work for everyone involved, and we still need to stay attentive/co-ordinate about any issues in conjunction with rolling 2.33 releaeses (it's more complicated than you might think). Originally, we only wanted to publically talk/educate about it when the releases are fully available for people. Maybe just check back next week :)
in reply to Delta Chat (39c3)

You are certainly right when you say “it's more complicated than you might think”. I am fine with waiting.

BTW, I always like these non-noisy updates. The last one included "calls" which is a big deal. However, it seemed to have the same importance as “we fixed this one bug that affect one person in a hypothetical scenario”. In this release, there is yet another big feature.

Keep up the good work.

This entry was edited (1 week ago)

The first article in one of our national newspapers this morning about the launch of the European Digital Infrastructure Consortium (#EDIC) today 🍾🇪🇺

Title: "Europe takes first step towards digital independence. Four European countries are taking on the American tech giants, taking the lead in developing European software." 🇳🇱🇫🇷🇩🇪🇮🇹

#digitalsovereignty #eupolicy #opensource

Seufz, da ziehe ich vor einigen Monaten aus Gründen meine Podcasts endlich zu einem deutschen Podcast-Hoster um, bin aber mit der Oberfläche in vielen Belangen nicht wirklich glücklich. Außerdem ist das Modell so, dass ich den zweithöchsten Tarif buchen musste, ohne wirklich alle Features zu brauchen. Und dann spült mir heute Morgen bei einer Diskussion auf FB ein Bekannter den Tipp zu @letscastfm in meine persönliche Wissensdatenbank. So ein sympathischer erster Eindruck! CC @quentara
in reply to Marco Zehe

Ich bin tariflich jetzt auf jeden Fall bis September kommenden Jahres festgelegt bzw. habe das schon bezahlt, ich werde aber definitiv mal einen Testpodcast bei @letscastfm starten, um zu testen, wie mir die Oberfläche des Backends gefällt und diese mit Screen Readern zusammenarbeitet. Die Website selbst macht schon mal einen guten Eindruck. Und wenn's mir gefällt, ziehe ich die Podcasts eben nochmal um. ;-)
in reply to Marco Zehe

Danke Dir, lieber @marcozehe 😊 Wir freuen uns, dass Dir der erste Eindruck gefällt, und hoffen, dass Dein Test dem in nichts nachstehen wird.

Bislang haben wir positives Feedback von Menschen mit Sehbeeinträchtigungen zur Nutzung mit Screenreadern erhalten. Wir lernen aber gerne weiter dazu. Solltest Du also Feedback oder Fragen haben, lass uns das gerne zukommen.

in reply to Marco Zehe

Ich habe jetzt mal einen Account und einen ersten Podcast bei @letscastfm eingerichtet, aber noch keine Episode hochgeladen. Das kommt im nächsten Schritt. Die Tabs pro Podcast oder so sind zwar technisch noch nicht richtig ausgezeichnet, konzeptionell aber gut erreichbar, und vor allem ist das mit Überschriften alles schön aufgeteilt, sodass ich schnell zu den richtigen Bereichen der Seite komme. Bisher gefällt's mir gut.
in reply to Marco Zehe

So, und das ist meine Nullnummer. ;-) Noch habe ich sie bei keinen anderen Podcast-Diensten eingereicht, das kommt dann im nächsten Schritt. Bisher bin ich auf keine nennenswerten Barrieren gestoßen. marco-erzaehlt.letscast.fm/epi…
in reply to Marco Zehe

Und jetzt ist der Podcast zumindest schon mal bei Apple und Spotify angemeldet. Und natürlich gibt's auf der Seite auch den RSS-Feed zum Abonnieren. marco-erzaehlt.letscast.fm/
in reply to Marco Zehe

Nach viel Kopieren, Einfügen und anderem Jonglieren von Daten habe ich inzwischen diverse weitere Dienste hinzugefügt. Bei keinem Punkt bin ich auf Barrieren gestoßen, die Anleitungen für jeden unterstützten Dienst sind sehr eindeutig und geradlinig. Unten auf der Seite sind jetzt viel mehr Dienste zu finden: marco-erzaehlt.letscast.fm/
in reply to Marco Zehe

gefällt mir so gut, und die @auphonic Integration läuft so prima, dass ich jetzt vorzeitig meine beiden anderen Podcasts umgezogen habe. Ob ich das jetzt mache oder erst kurz vor Ablauf der Laufzeit beim alten Hoster, ist ja gehüpft wie gesprungen, wie man bei uns so sagt. Auch die Verdrahtung mit den Ausspielern scheint soweit gut zu laufen. Nur Spotify und YouTube Music brauchen wirklich Handarbeit, weil sie den 301 Redirect ignorieren. Aber auch das lief gut. Happy Podcaster. :-)

Two things I want for Christmas, but which probably don't exist. 1. A wireless keyboard as much as possible like the Logitech G613, all 104 keys, but it doesn't have the palm rest or the macro keys going down the left side. 2. Sennheiser Ambeo binaural ear/microphones, but rather than working with an iPhone with a Lightning port, they have two mini (3.5 millimeter) plugs, one for the earphones, the other for the microphones.

Jonathan reshared this.

the g in gobject stands for glib, and the g in glib stands for gtk, and the g in gtk stands for gimp, however the g in gimp stands for gnu, so really the g in glib stands for gnu, but you shouldn't confuse it with gnulib, which is developed by the gnu project, who shouldn't be confused as the developers of glib, which is the gnome project, in which the g also stands for gnu

reshared this

Es gibt ja so Songs, mit denen hat man eine Love-Hate-Beziehung. Irgendwie hasst man sie, aber sie bleiben doch hängen, dann greift Stockholm-Syndrom, und dann mag man sie ... also ja, jetzt, wo wir das geklärt haben, hier ist mein heutiger Ohrwurm: song.link/i/1708988631

Jonathan reshared this.

Instagram collects scary amounts of your data 🫣 Mastodon on the other hand - None! ✅

What do we learn? There are great services available that do not abuse your data.

Give your friend the nudge they need - they deserve better than Meta's constant surveillance 👉🏼 tuta.com/blog/how-to-delete-an…

in reply to Kenny

@kbruen well I guess that is a matter of opinion? because this other person is saying totally the opposite:

mastodont.cat/@spla/1156953525…

and besides that, setup is something you do only once, the real struggle comes next, I have never self-hosted matrix but see people complaining or even de-federating due to server burden, while I do host a #chatmail server with +8k users in a 5eur VPS using 1.2gb of ram and 19gb of disk, now try the same with #matrix

I am total noob self-hoster btw

@delta

This entry was edited (1 week ago)
in reply to adb

That article doesn't seem to say much of anything about getting the server up and running.
Self hosting using Conduit is as easy as running an executable as a service, then configuring a reverse proxy to point to it. Very easily done with a systemd service, very easily done with a Docker container.
The only documented method to self host a Chatmail relay is to run a magical deploy script that has SSH access to the root account. Root access for a script is unacceptable unless it's the only thing running on that machine, and I don't want to dedicate a machine to Chatmail only.
in reply to 𝓒𝓱𝓻𝓲𝓼

@chris yes, there is initial support for that rolling out currently, see support.delta.chat/t/how-to-mi… and be sure to first update all devices -- the 2.33 releases are rolling to most stores but #fdroid will probably still be a bit later.
in reply to 𝓒𝓱𝓻𝓲𝓼

@chris with version 2.33, you can go to the settings > Advanced > Relays, and scan the QR code of on of the relays on chatmail.at/relays to create a new address :) as soon as you write to a group, the members will message you at your new address.
in reply to 𝓒𝓱𝓻𝓲𝓼

@chris Yes, it is possible since version 2.33: support.delta.chat/t/how-to-mi…

Giving this morning’s gold medal in U8 gymnastics to the lady, who:
- standing at the edge of the wobbly bit that connects cars
- held a tote bag with her 90° angled arm
- held an apple core in her mouth
- while the train started from a station into a particularly wobbly stretch
- standing on one foot …

…successfully tied the laces of the shoe on her other foot.

🥇

This entry was edited (1 week ago)
in reply to daniel:// stenberg://

Want to Get Free Amazon Gift Cards?
Online shopping has become a part of our daily lives! And what if you could get free gift cards for a platform like Amazon? Sounds amazing, right?
Just follow a few simple steps and you can earn gift cards worth $5, $10, or even $50-totally FREE!
How to get them?
Download and use selected apps
2 Complete short surveys
Refer friends to earn points
Redeem those points for Amazon gift cards
abre.ai/ogOP
in reply to sijmen (now on iceshrimp)

My concern here that LLMs do give worse-than results than something like Sonar - in many cases *static* code analysis was good enough to catch things that LLMs miss.

Ideally we would still use different tools, but GitHub seems to want to push everything into models.

tane.codes/@tanepiper/11567433…


Turned on the new #GitHub "Code Quality" feature that seems to be some extended CodeQL + AI.

It's actually worse than anything #SonarQube gave us. That actually gave really good reports on code smells, regressions and coverage failure. It was reliable.

It really does appear tech companies are replacing everything with expensive, resource intensive systems that give worse results and expect us to pay more @pluralistic #enshitificaton