matt

@matt@toot.cafe

Software developer, formerly at Microsoft, now leader of the AccessKit open-source project (https://accesskit.dev/) and cofounder of Pneuma Solutions (https://pneumasolutions.com/). My current favorite programming language is Rust, but I don't want to make that part of my identity.

Music lover. Karaoke singer. Science fiction fan. Visually impaired (legally blind). Secular humanist

This profile is from a federated server and may be incomplete. Browse more on the original instance.

ekuber, to random
@ekuber@hachyderm.io avatar

Random question of the day: how do TTS services deal with terminal emulators when sixel graphics are used? Do they get super confused? Ignore the graphics and just deal with the text?
If the later, that might be a workable (if super hacky) way of providing simultaneously nicer output in terminals and cleaner text that conveys the underlying semantics to TTS users.

matt,

@ekuber How hard would it be to provide a sample that we can run with various combinations of terminal and screen reader? Thanks for asking.

matt,

@ekuber I haven't yet found a terminal emulator that supports sixel graphics and is accessible with a screen reader. The legacy Windows console host (conhost), the newer Windows Terminal, gnome-terminal, and the Linux console driver all output nothing for the sixel sequence. PuTTY and Apple's Terminal.app ignore the initial escape code and output the rest of the sequence verbatim. Maybe I'm overlooking something.

matt,

@ekuber Which popular terminal emulators support sixel graphics anyway? The only one I've tried so far that does support sixel graphics, if I'm doing the test correctly, is foot, a lightweight Wayland-only terminal emulator that isn't accessible with a screen reader. (I have some usable vision.)

drew, to random

I'm old enough to remember when Chrome rolled out automatic alt text descriptions, in the ancient past of 2019. https://blog.google/outreach-initiatives/accessibility/get-image-descriptions/

So controversial! We faced immediate backlash. "How dare you?" they said. "Google is taking all the good alt text production jobs!" they said. Nevertheless, we persisted, in the face of the slings and arrows of outrageous fortune.

Wait, I'm sorry. Actually, nobody said anything like that. Man! That would've been so stupid, right?

matt,

@drew So what changed? Why is the reaction so much more negative now? And can we collectively change back?

jscholes, to accessibility
@jscholes@dragonscave.space avatar

Headline: "I Don't Understand why we Need This!": Sighted User Furious at Accessibility Feature for Blind People.

Reginald Cavendish, of Castle Combe in Wiltshire, today made an impassioned plea for a technology company to reconsider their new feature for and users, aimed at creating visual descriptions of images in a privacy-preserving way.

Speaking to reporters from his 1,800-acre estate, Reginald communicated his confusion about why the feature was being considered in the first place.

"I'm not clear on exactly who asked for this," he said. "When I need something, I ask one of my staff, and it just tends to get done. I take great Solace in that human touch, and I can't really comprehend why people would want computers doing things for them! My son has some friends who've seen disabled people on the television, and he didn't understand it either."

When asked whether he supported increased independence for people who are unable to see, he responded: "Look, some people have the money to pay people to be at their beck and call, to describe images or pick up their eyedrop bottles or what-have-you. I understand that doesn't go for everyone, but if someone can't afford human help, I'd suggest they'd be much happier making peace with the scraps they're thrown by society."

Editor's note: Reginald was asked to describe the featured image for this article. "Man on lawn," he offered over one shoulder, before moving off for a spot of afternoon shooting.

matt,

@FluidEscence @jscholes Satire, I assume.

attie, to random
@attie@chaos.social avatar

Ever heard of the "millennial pause"? ... it's both making me feel a bit old, and making me dig my heels in because starting audio on the first frame of video is incorrect and broken for technical reasons. 😂

One reason being that some pipelines / devices take any opportunity to sleep, and then take a moment to get going again - like bluetooth headphones... This leaves you missing the first ~500ms or doing the forward-backward dance to start the media again.

https://en.wikipedia.org/wiki/Millennial_pause

matt,

@attie

> some pipelines / devices take any opportunity to sleep, and then take a moment to get going again

This also affects screen reader users. Yet we don't want screen readers to pause for 500+ ms every time before speaking, because then they'd feel awfully laggy. Better to fix the underlying problem, or do a workaround to keep the audio device awake.

dangero, to accessibility

#AccessKit question. I'd like to recommend this to developers so they can make their projects more #accessible. I noticed there's no documentation. Is it easy enough to learn without the need for documentation? If not, I'm hesitant to recommend something that I know developers are going to struggle with. It's tough enough asking them to make their apps accessible when they don't know much about us, let alone asking them to learn a new library with no docs. #Accessibility #Blind #VisuallyImpaired

matt,

@jcsteh @dangero Yeah, we have to work on documentation. Also, AccessKit is useful in specific situations, but not others. For example, it's not useful for most web applications, or for applications that are already using the platform-native controls (e.g. Cocoa on macOS, UIKit or SwiftUI on iOS, or the Android standard widgets).

matt,

@dangero @jcsteh That still might not be a good fit for AccessKit, not just for lack of documentation, but because it sounds like that program might be using a GUI toolkit that already has an accessibility implementation. If I could take a quick look at the program, either on my own machine or via a remote session on yours, I could tell you what kind of advice to give the developer.

matt,

@dangero Removing Jamie from the thread. DM me if you want to do this. I won't be available until after 1 PM US Central time.

glyph, to random
@glyph@mastodon.social avatar

I seem to have accidentally discovered a new Fucking Guy that the leftist podcast/blog/youtube-o-sphere has not yet found, who writes and streams and podcasts about the usual garbage ("IQ", race science, misogyny) it is very frustrating, because he has a tiny audience and I don't want to give him more attention, but the fact that only his grifting victims have discovered his books on Amazon means that his reviews are all 4-5 stars.

matt,

@glyph How is one supposed to respond to something in a way that's discoverable to people who happen to find that thing but don't happen to be in the correct bubble to find the response, or don't happen to watch the response go by in their feed, if one isn't supposed to refer to the thing by its unique, unambiguous identifier, that is, the URL?

matt,

@glyph And of course, on platforms that support comments and reviews, like YouTube or Amazon, leaving a comment directly on the thing that you want to respond to makes the response even more discoverable by those who most need to find out about it. But we're not supposed to do that?

matt,

@glyph And perhaps I'm overly sensitive about this, but it annoys me that he casually suggests an inaccessible way of quoting the source material -- screenshotting -- without noting the accessibility issue.

matt,

@glyph I assume the problem with quote tweets is that, as an official Twitter feature, they're explicitly tracked by Twitter and thus reward the original poster. I wonder if a screenshot with automatic, complete accessibility metadata (something roughly equivalent to the HTML of the thing being screenshotted) would end up being coopted by the adversarial surveillance apparatus.

matt,

@glyph I used to be more hopeful about OCR, and computer vision based on machine learning in general, as a catch-all solution, but it's easy to see its current limitations. Consider the infamous example of "AI" being misrecognized as "Al".

matt,

@glyph Ugh, yeah.

matt,

@glyph Holy shit, that's starting to sound like the circumlocutions that I heard that freedom bloggers in communist China used/use.

matt,

@glyph BTW, I thought, while listening to the video you linked, that maybe the music video itself was just a stand-in for the more serious things that people discuss.

glyph, to random
@glyph@mastodon.social avatar

For OPSEC reasons I cannot disclose whether I recently had, or may soon have, a birthday, but I probably was born at some point. In honor of that fact and inspired by @mcc's excellent #Glitch4Andi hashtag, if you are so inclined, please post a story about or picture of a time (ideally in the last ~year) that a computer did something cool and fun and improved your life in some way, using the hashtag #FunTech4Glyph .

(Bonus points if it looks cool and/or uses Python, but very much not necessary.)

matt,

@davew @mcc As much as I don't want to spoil this positive moment for @glyph, I think I ought to point out that PySimpleGUI, being based on Tk, is not at all accessible to screen reader users. If you're just doing fun stuff for yourself, then carry on.

matt,

@davew @glyph Unfortunately, there is no Python GUI toolkit that I can recommend without reservation across all platforms. If you're developing only for Linux (or free desktops more generally), GTK is probably the best. I'm also working on making GTK accessible on Windows and macOS. For those platforms at the moment, wxPython is adequate, with some caveats. For example, last time I checked, one of wx's two list view controls was accessible on Windows but not macOS, and vice versa.

matt,

@davew @glyph Toga (https://beeware.org/project/projects/libraries/toga/) is also worth a look. It's very wx-like in its approach. I haven't actually tried it yet though. The fact that it brings in .NET as a dependency on Windows (so it can use Windows Forms) isn't ideal, but I understand that working with raw Win32 would be more work.

matt, to random

I'm re-listening to the audiobook of The Three-Body Problem by Liu Cixin, translated by Ken Liu. It's too bad the audiobook producer didn't recruit a Chinese-speaking narrator for this book; such a narrator presumably would have pronounced all of the Chinese names more accurately than the American narrator Luke Daniels, even though he presumably had some coaching on pronunciation. Surely there are competent voice actors out there who are fluent in both Chinese and English.

matt,

Just found out that there's a new audiobook of The Three-Body Problem, released a few months ago, that was narrated by Rosalind Chao, who played one of the lead characters in the Netflix adaptation. Seems like a much better choice of narrator than the previous audiobook. Buying now.

matt,

@drew I actually thought Luke Daniels's portrayal of some of the characters was exaggerated, perhaps going for too much comedic effect. I'm thinking in particular of his portrayal of Von Neumann in the scene with 30 million soldiers forming a computer. It doesn't help that I primarily think of Luke Daniels as the narrator of Off to Be the Wizard and its sequels, which have much more of a light-hearted tone.

cross, to random
@cross@discuss.systems avatar

H/T my friend Austin (who's not on social media).

@irene @dan

matt,

@cross Can you please describe the image, that is, add alt text?

amxmln, to foss
@amxmln@mastodon.design avatar

Apparently the German government paid Microsoft 198 million Euro for software licenses in 2023 alone…and all I can think of is the good that such a large sum could do if instead it was donated to the Linux foundation and various other projects. 😳

matt,

@amxmln In fact, the Sovereign Tech Fund, which I believe is affiliated with the German government, already provided a million in funding to the GNOME Foundation, and that is paying for my current accessibility work among several other things.

  • All
  • Subscribed
  • Moderated
  • Favorites
  • JUstTest
  • mdbf
  • ngwrru68w68
  • cubers
  • magazineikmin
  • thenastyranch
  • rosin
  • khanakhh
  • InstantRegret
  • Youngstown
  • slotface
  • Durango
  • kavyap
  • DreamBathrooms
  • megavids
  • tacticalgear
  • osvaldo12
  • normalnudes
  • tester
  • cisconetworking
  • everett
  • GTA5RPClips
  • ethstaker
  • anitta
  • Leos
  • provamag3
  • modclub
  • lostlight
  • All magazines