The media in this post is not displayed to visitors. To view it, please go to the original post.
#movuary 28 The After Party. I'm gonna miss doing these. This was fun. I hope you guys enjoy this last one for the year.
reshared this
2026-03-01 12:11:33
2026-03-01 12:11:33
2026-02-28 20:49:57
9837888
Peter Vágner reshared this.
#movuary 28 The After Party. I'm gonna miss doing these. This was fun. I hope you guys enjoy this last one for the year.
#movuary 28 The After Party. I'm gonna miss doing these. This was fun. I hope you guys enjoy this last one for the year.
2026-03-01 12:02:19
2026-03-01 12:02:19
2026-02-28 20:59:32
9838330
Peter Vágner reshared this.
Found that orca remote plugin on github , and fixed it up so it is actually useful now. What you guys think? I love having access to this thanks to vibe coding. github.com/serrebi/orca-remote
GitHub - serrebi/orca-remote: Bidirectional communication between Orca (Linux) and NVDA (Windows) screen readers using the NVDA Remote protocol
Bidirectional communication between Orca (Linux) and NVDA (Windows) screen readers using the NVDA Remote protocol - serrebi/orca-remoteGitHub
reshared this
Steffen and Peter Vágner reshared this.
2026-03-03 09:52:51
2026-03-01 11:39:45
2026-03-01 10:33:36
9840369
Peter Vágner reshared this.
in reply to Jonathan
in reply to Jonathan
in reply to Jonathan
in reply to Felix Steindorff
Ok just found a quite nice speech to text app for PC which does actually properly work and is quite accessible.
handy.computer
To clarify: Speech to Text (like dictation on phone), not transcription.
handy.computer
To clarify: Speech to Text (like dictation on phone), not transcription.
reshared this
Peter Vágner and Mike Gorse reshared this.
I know it, but I had problems with large Whispper models. Sometimes parts are not recognized correctly, sometimes only exclamation marks and question marks are transcribed.
Cool, Wee need this for IOS too
Yes, I love this thing. I've been using it for a few weeks now and it really saves a lot of time in some situations.
@felixule Yeah, especially when I'm eating and want to respond to something with a quick message it's awesome because I only need one hand for the shortcut.
2026-02-28 08:46:28
2026-02-28 08:46:28
2026-02-28 02:55:16
9833846
Peter Vágner reshared this.
in reply to Tamas G
in reply to victor tsaran
PSA for Android devs using Jetpack Compose with TalkBack: ExposedDropdownMenuBox breaks linear swipe navigation order. It creates internal Surface nodes with isTraversalGroup = true, which makes TalkBack sort your dropdowns into completely wrong positions in the swipe order — even though explore-by-touch works fine.
The fix: replace ExposedDropdownMenuBox with plain Box + DropdownMenu. You lose the auto-width matching (easy to add
back with onGloballyPositioned) but gain an accessibility tree that actually works.
isTraversalGroup on parent sections didn't help either — the rogue groups inside the Material component override it.
#Android #A11y #JetpackCompose #TalkBack #Accessibility
The fix: replace ExposedDropdownMenuBox with plain Box + DropdownMenu. You lose the auto-width matching (easy to add
back with onGloballyPositioned) but gain an accessibility tree that actually works.
isTraversalGroup on parent sections didn't help either — the rogue groups inside the Material component override it.
#Android #A11y #JetpackCompose #TalkBack #Accessibility
Peter Vágner reshared this.
Wow, what are you doing there?
@vick21 oh, I ran into this oddly for the TGSpeechBox Settings UI :) but all this work especially with getting Jetpack compose stuff to have proper roles and sound good with Talkback will be really useful for work too, no doubt, so I'm not upset at all going a bit more out of my web-centric comfort zone. xD
2026-02-28 08:38:10
2026-02-28 08:38:10
2026-02-28 06:47:29
9834295
Peter Vágner reshared this.
Hurvínek letos slaví 100 let. Jak přišel ke svému jménu? Původně měl být “Spejblík”. Malíř Dvořák prý první vyřezanou verzi označil za “(z)kurvínka”, což se v okruhu přátel hned ujalo. A malou úpravou z hlavy Skupovy manželky bylo hotovo.
This entry was edited (Saturday, February 28, 2026, 7:08 AM)
reshared this
Peter Vágner reshared this.
2026-02-26 19:12:13
2026-02-25 16:51:00
2026-02-25 14:51:04
9816580
Peter Vágner reshared this.
in reply to ondrosik
in reply to ondrosik
in reply to ondrosik
in reply to Andre Louis
in reply to Andre Louis
in reply to Andre Louis
in reply to ondrosik
in reply to ondrosik
I was proud that I can recognize songs created by Suno. But for songs created in Gemini well, that seems to be hard for me in some cases.
Peter Vágner reshared this.
Oh really? I haven't tried that. I can definitely tell Suno for now...
There's nothing attached.
I can tell it's AI. There's something about the aliasing in the high strings, and also it floats around the stereo spectrum a bit. That would never happen in a real Orchestral arrangement unless it was for artistic reasons.
Doesn't mean it isn't pretty good, but I hear AI anyway.
@FreakyFwoof I'll be straight up honest, if I didn't knew, I couldn't. Could also just be a bad mix or vst/instrument for me. Yes of course not the real instrument, but well. Since I know, I can somewhat tell, but this shit is getting way too good for my taste.
After close listening I can hear some little distortion too, It also generated another piece where the distortion is more prominent.
hmm, for some reason attachment doesnt work here, lets see if I can fix it.
I just don't know how to make it longer or start from some kind of intro, it just throws those little pieces at me.
2026-02-25 20:48:47
2026-02-25 16:47:53
2026-02-25 15:11:28
9816800
Peter Vágner reshared this.
in reply to Cullen Gallagher
in reply to aaron
in reply to Quin
in reply to Quin
in reply to Alex Chapman
in reply to Cullen Gallagher
in reply to Cullen Gallagher
in reply to Cullen Gallagher
Well here it is, Selvas is announcing the BrailleSense 7! I’ll be very interested to take a look at this. Here is the full announcement. Introducing the New Smart Braille Tablet Series: BrailleSense 7
While the name may be straightforward, the innovation behind BrailleSense 7 is
anything but. This next-generation platform represents a major leap forward in
performance, usability, and future-focused design.
This series is offered in 3 different sizes:
BrailleSense 7: 40-cell
BrailleSense 7: 32-cell
BrailleSense 7: 20-cell
Planned features across all models include:
Android 15 operating system
Expanded braille-first applications
Integrated AI capabilities
Touch-sensitive braille cells
User-replaceable battery
Included QWERTY keyboard case with secondary battery
Powered by Google Gemini AI
CSUN Launch & Preorder Promotion
BrailleSense 7 will be officially unveiled at CSUN 2026, where we will begin taking pre-orders and continue doing so up until the time we start shipping. While we can’t reveal the price yet, we can say that we will be offering a large discount on all pre-orders. Visit the Selvas BLV booth #508 to place your preorder early, as initial stock will be given on a first-come first-served basis.
#508
reshared this
Peter Vágner, Zach Bennoui and Khronos reshared this.
So... another several thousand dollar device that's launching already 2 years behind. Incredible innovasion.
To be fair to Selvas though, they actually are interested in updating and maintaining their braille note takers for years, unlike certain other companies. The Windows CE-Based BrailleSense U2 got updates into 2020, and they bumped the android version of the BrailleSense 6 at least once.
This entry was edited (Wednesday, February 25, 2026, 3:17 PM)
@alexchapman @TheQuinbox @fireborn I was discussing this in another thread. I still think the main people who will get these devices are the education market. They want a locked down closed system which allows minimal fuss and opertunities for students to mess up the device.
Touch-sensitive braille cells? This sounds like it could be either very great or one of the greatest fails in AT history.
This entry was edited (Wednesday, February 25, 2026, 3:16 PM)
I wish I was interested. Just have been so unimpressed with their support over the last few years. I will not be buying another Hims product.
Ooo nice! And android 15? Wow that's nice! I'm not complaining if this one stays on android while the BrailleNote evolve switches to windows. In my view, I think using android for certain note takers can give people a taste of the android experience. ANd wow this thing has gemini? Wow!
2026-02-26 07:34:45
2026-02-25 16:46:50
2026-02-25 15:22:31
9816890
Peter Vágner reshared this.
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to MostlyBlindGamer
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to MostlyBlindGamer
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to Sami Lehtinen
in reply to MostlyBlindGamer
@MostlyBlindGamer Just to lighten up today. AI slop image based on classic meme.
in reply to 🇨🇦Samuel Proulx🇨🇦
in reply to 🇨🇦Samuel Proulx🇨🇦
AI is both impressive and ridiculous. I have a little command line Python script that does some basic text reformatting stuff for me. I wanted a version that I could use on my phone. I gave Gemini the script, and asked it to make "a version of this program with the same functionality that I can use on my iphone, but with a textbox and some buttons instead of command line options." I expected it to rewrite this extremely simple Python file into JavaScript or something and give me an HTML file. Or if it wanted to do things the hard way, maybe swift. Nope! It installed uvicorn and requests and fastapi and a bunch of stuff and then rewrote the python script so it could import the existing functionality into the web app it just built. To be fair, it worked perfectly the first time, either on the command line or as a web server. But the original script was about 100 lines of Python! Rewriting it in JavaScript would've been much faster. And not required a reverse proxy and a bunch of deployment work. I guess it took "the same functionality" way too literally and decided the best way to achieve that would be to run the exact same code. When we talk about the impact of #AI, nobody seems to be talking about the impact caused by running AI generated code for years to come. If I didn't know better, and was just an AI vibe coder, I'd now be running an entire docker and reverse proxy setup for something that could be done in the browser with a hundred or so lines of JavaScript. That's honestly a tiny impact individually, but over the course of hundreds and thousands of people and companies, it's going to add up.
#AI
reshared this
Peter Vágner, Rui Batista, Mike Gorse and Federico Mena Quintero reshared this.
I dunno, I think I’ve been talking about this for a bit. The risk there, that I’ve also observed at work, is that the LLM will run with it, no questions asked. What I expect from a developer, and insist on doing myself, is to ask “do you mean like serve this as a web app, rewrite in JavaScript or do you just want to run this on excel?”
@MostlyBlindGamer Right. I admit that my problem was a bit underspecified. But if I'd handed you, a real human, a hundred lines of Python that does a bunch of string replacements and uses some regular expressions, and said I wanted to run this on my iPhone, your first reaction would probably not have been "turn it into a web app". Actually, your first reaction would have probably been "so install iSH and stop bothering me" LOL.
and I wouldn’t even have used enough water and power to keep entire families alive!
I have taken the time to play around with Codex and Claude Code. In my experience Claude Code is a little more inclined to think about frameworks and tech stacks. Now, I often supply long prompts trying to encapsulate the details from a programmer's perspective. I'm agreeing with everyone else that this only highlights how tech skills are still non-negotiable. Even if a model can handle the grunt work of writing. This simply makes more developers, project managers.
@MostlyBlindGamer Brython + PWA works.
But I’ve observed same pattern. Always discuss design and solution first. And don’t approve anything before full diff check.
I use Agentic flow for fun hobby stuff. But at work I go with Aider and manual commits after review. My CONVENTIONS.md contains very strict instructions.
This entry was edited (Wednesday, February 25, 2026, 8:02 PM)
@sl at work I talk to people, read docs, and invest time into teaching humans to help them develop as people and professionals, instead of wasting that time with word generators.
🇨🇦Samuel Proulx🇨🇦 likes this.
The media in this post is not displayed to visitors. To view it, please go to the original post.
@MostlyBlindGamer Just to lighten up today. AI slop image based on classic meme.
Yeah the direct way would have been JS given some regex dependency or other, or if it must, python wasm? That's a thing now isn't it? Is that what it did?
This entry was edited (Wednesday, February 25, 2026, 4:36 PM)
That's why #AI is a tool or rather a set of tools you need to learn to wield. And that's why, fortunately or not, vibe coding per se will go away. If you steer it correctly, if you follow what it's doing, it will fulfill your expectations most of the time.
#AI
🇨🇦Samuel Proulx🇨🇦 likes this.
2026-02-25 16:44:36
2026-02-25 16:43:06
2026-02-25 15:50:20
9817003
Peter Vágner reshared this.
The error of my life
in reply to Bogomil Shopov - Бого
in reply to JB
The media in this post is not displayed to visitors. To view it, please go to the original post.
The error of my life
Peter Vágner reshared this.
The other day I had:
"OOM is now expected behaviour"
Server was a bit passive-aggressive..
2026-02-25 16:55:11
2026-02-25 16:42:09
2026-02-25 15:53:28
9817080
Peter Vágner reshared this.
Just saw a german YouTube short in which a music snipped was playing, and thhe singing was literally removed by Googles weird multilingual AI trannslation shit, and replaced by a bad english AI voice. We're all lost.
in reply to Jonathan
The media in this post is not displayed to visitors. To view it, please go to the original post.
Just saw a german YouTube short in which a music snipped was playing, and thhe singing was literally removed by Googles weird multilingual AI trannslation shit, and replaced by a bad english AI voice. We're all lost.
reshared this
Peter Vágner and Zach Bennoui reshared this.
Auto dubbing became even worse now because sometimes it tries to replicate emotions. Spoiler: It sounds terrible.
2026-02-25 16:40:12
2026-02-25 16:40:12
2026-02-25 15:24:03
9817316
Peter Vágner reshared this.
The media in this post is not displayed to visitors. To view it, please go to the original post.
Did you know Joplin supports ABC notation?🎵
Turn your notes into sheet music directly inside your notebook.
📺 Watch now 👉 youtu.be/4AMdwydbFjQ
Feature : ABC Notation
Turn your notes into music with ABC notation in Joplin 🎼In this video, we explore how Joplin supports ABC notation, allowing you to write music directly ins...Joplin (YouTube)
reshared this
Peter Vágner reshared this.
2026-02-25 19:38:29
2026-02-25 14:36:55
2026-02-25 10:03:00
9814907
Peter Vágner reshared this.
in reply to Jake Gross
in reply to Pratik Patel
in reply to Jake Gross
in reply to JamminJerry
in reply to JamminJerry
in reply to Jake Gross
in reply to Jake Gross
Apparently, an NVDA addon for the Atari STSpeech synthesizer now exists. For anyone interested, it can be found here: dectalk.nu/Software%20and%20Ma…
reshared this
Peter Vágner and Zach Bennoui reshared this.
I never heard this before. But, it sounds terrible.
@ppatel It's definitely a novelty synth.
holy cow! and we thought the old echo speach was bad! that is just ouch! but I downloaded it, and will keep it for a just because thing.
@JamminJerry yeah that sounds horrable.
I'm using this synth right now. Only me would be all over this like a fly to feces.
Oh god! It sounds somehow worse than the original. I think the sample rate that the speech synthesizer produces doesn't match that of NVDA's speech output thing. Also, this oughtta be interesting, as StSpeech is the most crash-happy synth I've ever come across.
2026-02-25 14:35:23
2026-02-25 14:35:23
2026-02-17 15:52:40
9814935
Peter Vágner reshared this.
in reply to Cees de Groot
in reply to Cees de Groot
in reply to Cees de Groot
in reply to Cees de Groot
in reply to Paolo Amoroso
in reply to Paolo Amoroso
in reply to Cees de Groot
in reply to Cees de Groot
in reply to cyclical_obsessive
in reply to Cees de Groot
in reply to Cees de Groot
in reply to The ol' tealeg 🐡
in reply to Cees de Groot
in reply to Greg Lloyd
in reply to Cees de Groot
in reply to Cees de Groot
in reply to RooneyMcNibNug
in reply to Cees de Groot
in reply to Cees de Groot
in reply to Cees de Groot
in reply to Cees de Groot
The Kobo EPUB version makes it very difficult to see the formulas in the first few chapter. It's very blurry. The following images is a X2 magnification.
in reply to Northerly Goose
in reply to Cees de Groot
in reply to Ramin Honary
in reply to Cees de Groot
in reply to Cees de Groot
in reply to Cees de Groot
in reply to Cees de Groot
in reply to Reino
in reply to Reino
in reply to Cees de Groot
in reply to Howard the Geek
in reply to Howard the Geek
@howard I can confirm. I purchased from KOBO and I was allowed to download and read on my macOS machine.
in reply to Cees de Groot
look what just arrived!
in reply to Cees de Groot
in reply to Michael Grinder
in reply to Cees de Groot
in reply to Cees de Groot
in reply to Cees de Groot
in reply to David in Tokyo
in reply to Cees de Groot
in reply to Cees de Groot
in reply to Cees de Groot
in reply to rjray
in reply to rjray
The media in this post is not displayed to visitors. To view it, please go to the original post.
Well, today is the day. I'm finally "sorta happy enough to pull the trigger" on publishing the book I've been working on for a very long time. It's a technical history book: by a techie, for techies (although I think that between all the code samples, there is plenty of meat for "tech-adjacent" and "tech-interested" people). It tells the story of the Lisp programming language, invented by a genius called John McCarthy in 1958 and today still going strong (to the extent that many people see it as the most powerful programming language in existence).
And this is a time for shameless self promotion, even if you don't plan on buying the book, please repost 
. Self-publishing is self-marketing, so there we go.
If you do buy and read it, please let me know how you liked it!
The book landing page, berksoft.ca/gol, has links to all outlets where you can buy the book,
reshared this
Peter Vágner reshared this.
Ik heb er eentje besteld. Ben benieuwd!
Just ordered it, should arrive on Monday.
That looks excellent! BTW the link has an extra comma which might need fixing 😉
A great angle for telling the history of Lisp.
Is the ebook DRM-free?
@amoroso Yes. DRM is stupid and I won't have anything to do with it.
I'm also now selling a (DRM-free) PDF (see the book site, berksoft.ca/gol, for a link), I'm not a big fan of technical books in epub and this is, I think, a much better reading experience.
@amoroso says so on the page!
Good luck with the book. Cheers.
“Those who cannot remember the past are destined to repeat it”
I’m not sure we should enshrine LISP into Claude Code, even though I once wrote LISP that flew in a cruise missile, and several “expert systems” for critical applications, it seems likely the “AI”s might choose LISP to enable the singularity.
Emacs and LISP - Live it, love it, remember it!
@cyclical_obsessive Finally someone who spells '"AI"' correctly. In this context, I mean.
This entry was edited (Friday, February 20, 2026, 7:28 PM)
Congrats! Just ordered a physical copy. 👍🏼
fabulous - ordered a physical copy from Lulu. Congratulations. I'm looking forward to reading it!
quotable already, from chapter 8: "Scheme is like a scalpel, and Racket hones that scalpel to a precise tool for language research, Scheme's original focus." #Racketlang
Congratulations and thanks! I’m enjoying the epub edition from Lulu:
“People writing Lisp code to do old-style AI are engaged in a creative act of discovery, an attempt to understand… To discover requires a different sort of language, and it requires a different approach, captured well by the writer Joan Didion:
I write entirely to find out what I’m thinking, what I’m looking at, what I see and what it means.
–Joan Didion, “Why I Write”
Richard P. Gabriel
Foreword
#lisp
#lisp
@Roundtrip
This is much the same as the way Michel de Montaigne wrote his famous essays in 16th century France: to try out his thoughts, from the French word essayer, "to try".
Thus was born the literary form of the essay, much to the chagrin of secondary school students everywhere.
But he was the ur-blogger, and is one of my heroes.
I love technical history books! Purchased a copy.
this looks awesome and lisp has piqued my interest for some time now without having "jumped in" too much - going yo grab a copy of this, thanks!
@rooneymcnibnug it sounds you're the sweet spot of my target audience. Let me know how you like it!
OOOO, that sounds interesting! I go buy now!
this is excellent! I'm so excited for this, it looks awesome!
just ordered! Congrats on the launch 💜
The media in this post is not displayed to visitors. To view it, please go to the original post.
The Kobo EPUB version makes it very difficult to see the formulas in the first few chapter. It's very blurry. The following images is a X2 magnification.
@NortherlyGoose strange. What version are you using? I've tested on my laptop and on my Libra and things looked fine there.
The media in this post is not displayed to visitors. To view it, please go to the original post.
New book: “The Genius of Lisp” by Cees de Groot
Looks like a fascinating read! They have provided chapter 8 as a PDF file downloadable gratis as a sneak-peek into the rest of the book, guess what it’s about:
Chapter 8: Sussman and Steel make Scheme
Awesome! I can’t wait to read that chapter, and then the rest of the book!
Details on the book homepage: berksoft.ca/gol/
#tech #software #Lisp #Scheme #SchemeLang #R7RS
Well, today is the day. I'm finally "sorta happy enough to pull the trigger" on publishing the book I've been working on for a very long time. It's a technical history book: by a techie, for techies (although I think that between all the code samples, there is plenty of meat for "tech-adjacent" and "tech-interested" people). It tells the story of the Lisp programming language, invented by a genius called John McCarthy in 1958 and today still going strong (to the extent that many people see it as the most powerful programming language in existence).And this is a time for shameless self promotion, even if you don't plan on buying the book, please repost :-). Self-publishing is self-marketing, so there we go.
If you do buy and read it, please let me know how you liked it!
The book landing page, berksoft.ca/gol, has links to all outlets where you can buy the book,
@ramin_hal9001 ping! Thanks for the PDF option 👍 I figured it was about time I actually tried to understand the context for LISP, not merely the syntax 😊
And it’s only twelve quid! Just ordered it from Lulu - best of luck with it all.
Just bought from Lulu.
can't wait to read this. I am still a scheme weenie at heart.
Once had a discussion with a fellow student about what would be the most powerful programming language.
He thought it was Forth.
I asked him which would be easier to program (taking e.g. Lisp as the other contender):
A Forth interpreter in Lisp, or
a Lisp interpreter in Forth.
You can guess who 'won'.
BTW my master's thesis was titled "CAR".
Standing for
1. The obvious one !
2. The initials of C.A.R. Hoare
3. Computer Aided Requirements
@reinodeboer They both have interesting ideas that are not fully explored elsewhere. Study both.
I am looking for to reading your book, The Genius of #Lisp.
Sorry to ask a trivial question here (but maybe others would like to know), I'm on the Kobo site to buy the ePub version of your book, but I'd like to confirm I can read it without their poor Kobo Reader app.
Otherwise, I'd love to send you the money directly...
#lisp
@howard It's DRM free, you can read it wherever. You can also buy it through Lulu.com if you don't like Kobo, but to me it doesn't make much difference.
The most direct way is to buy the PDF on Shopify (all the various options are linked from the book site, by the way, berksoft.ca/gol). I personally much prefer reading technical books in PDF format, but then I have the accordingly-sized large format color e-Reader for it :).
But buy wherever you're happy to buy it.
The media in this post is not displayed to visitors. To view it, please go to the original post.
@howard I can confirm. I purchased from KOBO and I was allowed to download and read on my macOS machine.
The media in this post is not displayed to visitors. To view it, please go to the original post.
look what just arrived!
Any possibility of a hardcover from Lulu? I'd like the hardcover version, but not if I have to give money to Amazon.
@mgrinder I'm not sure why, but as far as I can tell Lulu doesn't do hardcovers on the 7x10 format I picked. Sigh. It's on my todo list to ask them though, I hope it's just a mistake.
you got me at "Lisp". Congrats on releasing the book! Looking forward to reading it.
this books sounds amazing, just ordered a paperback from lulu 💜
Hmm. The Amazon blurb mutters about ELIZA, but ELIZA wasn't written in Lisp, it was written in MAD_SLIP. My reading of the original paper was that Joe W. essentially reinvented Lisp.
By the time I arrived at MIT (1972), Joel Moses' introductory CS/programming course was taught in Lisp (and was so good, I got a gig in the Mathlab group by passing a Lisp test), and the AI Lab's PDP-6 had a version of Eliza that was probably rewritten in Lisp.
@djl I came here to make the same comment.
SLIP was interesting because it used bidirectional links—a list can be traversed upwards as well as downwards. It was invented before Lisp had won the AI language wars.
Common Lisp story.
The time, 1987. The place: the 5th generation group in NEC's research lab in Tokyo. We (me and the locals) were frantically programming Common Lisp, me using GLSteele's manual, the locals using the Japanese translation. Every time there'd be a GLS-style joke (which said manual is seriously dense of), I'd check the Japanese translation, and every time, the translation completely missed the point and was complete gibberish.
Humor is hard. GLS humor, even harder.
Apparently this is the book I have been waiting for all the time without knowing. Will buy immediately. Thanks for posting!
I love the book so far! I got an email from Pendora about the updated PDF, but the page it takes me to fails to download the new version. Do you have any insight into their process?
@rjray Try it again, support just pinged me that the broken download issue should have been fixed. If not, private message me with your order number and we'll work something out.
Glad you like the book!
@rjray yes, the insight is that I apparently fatfingered the update. I'm working with Pendora support, they are super helpful, but they also are in India, so first I had to wait out the weekend (teaches me to do updates on Saturday morning) and now it's a back-and-forth with 10.5 hours time difference...
I'm going to mail out everybody an updated PDF later today, it's easier 🙂
2026-02-27 19:13:20
2026-02-25 08:08:53
2026-02-25 07:01:48
9813466
Peter Vágner reshared this.
in reply to Amir
in reply to Amir
in reply to Paweł Masarczyk
in reply to Paweł Masarczyk
in reply to Amir
in reply to Amir
in reply to Paweł Masarczyk
in reply to Amir
in reply to Paweł Masarczyk
in reply to Amir
in reply to Paweł Masarczyk
in reply to Amir
in reply to Paweł Masarczyk
in reply to Amir
in reply to Paweł Masarczyk
in reply to Amir
in reply to Paweł Masarczyk
in reply to Paweł Masarczyk
in reply to Amir
in reply to Amir
in reply to Amir
in reply to James Scholes
in reply to James Scholes
I’m proud to present the third add-on in my four-part NVDA add-on series.
HTML Element Inspector for NVDA is a powerful accessibility debugging tool that reveals detailed information about the current element in Browse Mode. It reports roles, states, attributes, ARIA properties, links, form details, and structural context in a clean, readable format. Designed for accessibility engineers, power users, and testers, it makes inspecting web elements faster, clearer, and more precise directly inside NVDA. Use NVDA+SHIFT+F1.
github.com/amirsol81/html-Elem…
@NVAccess
HTML Element Inspector for NVDA is a powerful accessibility debugging tool that reveals detailed information about the current element in Browse Mode. It reports roles, states, attributes, ARIA properties, links, form details, and structural context in a clean, readable format. Designed for accessibility engineers, power users, and testers, it makes inspecting web elements faster, clearer, and more precise directly inside NVDA. Use NVDA+SHIFT+F1.
github.com/amirsol81/html-Elem…
@NVAccess
GitHub - amirsol81/html-Element-Inspector: HTML Element Inspector for NVDA is a debugging tool for accessibility developers. It reports the current element’s role, states, attributes, and inferred properties in Browse Mode, helping verify ARIA exposure an
HTML Element Inspector for NVDA is a debugging tool for accessibility developers. It reports the current element’s role, states, attributes, and inferred properties in Browse Mode, helping verify A...GitHub
reshared this
Paweł Masarczyk and Peter Vágner reshared this.
This add-on is really Good for #accessibility testing work, accessibility tree inspector within NVDA is certainly can be a long-term goal, the nostalgia of JFWTechnical script by @jcsteh and Mick still lingering in my mind.
Hey! This is a gamechanger for me as an accessibility evaluator. Thank you for creating this. I think even my sighted colleagues will find it useful as we sometimes struggle to locate an element in DOM that is definitely read by NVDA. Could you please add a line about the accessible name of an element, that is what it is and where it comes from? Thanks a lot!
@Piciok I'm so glad you like it. As for your request, please provide me with more information. In particular, what's the element or aspect you want revealed?
@Piciok After a bit of digging, I think I know what you want. I'm taking a stab at it now.
Sorry for not being more explicit about it. I mean, when explicit-name is true, what is this explicit name? Is it an aria-label, aria-labelledby, a title or the native, visible text of the element? What is its value?
Another thing I'd find useful is the ability to explore the elements within the focused element. Sometimes the name read out by NVDA is burried within a span element inside that link or button. Currently there is no way to explore that, am I right? I guess we'd need a heading structure with levels to make it convenient to navigate.
@Piciok All great and useful points! I'm implementing them now.
And you've integrated them all so quickly. Amazing work! One more small thing: could you also make it reveal the source of the accessible name if possible, that is whether it comes from its own text, an aria-label, title etc? I believe the order in which accessible names are searched by the browser's accessibility tree is: aria-labelledby, aria-label, the element's visible text and the title attribute. It could also reveal whether the visible text is put directly on the root of the element or whether it's nested within a child element like a span. Thanks, again, that makes my work so much easier.
@Piciok Thanks a lot, Paweł. Really appreciate the feedback.
The inspector now reports both accessible-name and accessible-name-from, and when applicable, also explicit-name-from. The goal is to reveal the source of the accessible name as accurately as possible (e.g. content, attribute, related-element), but in a conservative way, meaning it avoids guessing the exact ARIA mechanism unless the accessibility layer exposes reliable evidence.
In Advanced mode, you can also explore the exposed subtree of the focused element. If the visible text is nested inside child elements (like spans) and the accessibility API exposes them, they will appear in the Children section. If no subtree is exposed, the report explicitly states that the subtree is not available via the accessibility API.
The idea is to stay faithful to what the accessibility tree actually exposes, rather than reconstructing the full accessible name algorithm from the DOM.
@NVAccess
The inspector now reports both accessible-name and accessible-name-from, and when applicable, also explicit-name-from. The goal is to reveal the source of the accessible name as accurately as possible (e.g. content, attribute, related-element), but in a conservative way, meaning it avoids guessing the exact ARIA mechanism unless the accessibility layer exposes reliable evidence.
In Advanced mode, you can also explore the exposed subtree of the focused element. If the visible text is nested inside child elements (like spans) and the accessibility API exposes them, they will appear in the Children section. If no subtree is exposed, the report explicitly states that the subtree is not available via the accessibility API.
The idea is to stay faithful to what the accessibility tree actually exposes, rather than reconstructing the full accessible name algorithm from the DOM.
@NVAccess
Great! I take it you're referring to the 1.1.1 release. I have tried it on a simple link with regular visible text between the opening and closing a and haven't found the source being mentioned, just the text. I have also tried it on a link which is surely labeled with aria-label and I haven't found the source either. Maybe I misunderstood how it works - I have tried both the simple and advanced mode.
@Piciok
Thanks for testing it so thoroughly.
The add-on reports accessible-name-from only when the source can be reliably determined from what the accessibility layer exposes. In some cases, NVDA provides the computed name but does not expose detailed information about how that name was derived (for example, whether it came from aria-label, visible text, or another mechanism).
If the accessibility API only exposes the final computed name without a clear “name-from” indication, the inspector intentionally avoids guessing the source. This is by design, to prevent misleading information.
For simple links with visible text, if the accessibility layer does not explicitly report the source as “contents”, the tool will only show the accessible name itself. The same applies to aria-label: if the underlying object does not clearly expose that the name comes from aria-label, the inspector will not assume it.
That said, I will review those specific scenarios again to see whether the source can be detected more reliably without introducing heuristics or guesswork. I'd appreciate it if you could share some examples, links, elements, or whatever I can observe.
@NVAccess
Thanks for testing it so thoroughly.
The add-on reports accessible-name-from only when the source can be reliably determined from what the accessibility layer exposes. In some cases, NVDA provides the computed name but does not expose detailed information about how that name was derived (for example, whether it came from aria-label, visible text, or another mechanism).
If the accessibility API only exposes the final computed name without a clear “name-from” indication, the inspector intentionally avoids guessing the source. This is by design, to prevent misleading information.
For simple links with visible text, if the accessibility layer does not explicitly report the source as “contents”, the tool will only show the accessible name itself. The same applies to aria-label: if the underlying object does not clearly expose that the name comes from aria-label, the inspector will not assume it.
That said, I will review those specific scenarios again to see whether the source can be detected more reliably without introducing heuristics or guesswork. I'd appreciate it if you could share some examples, links, elements, or whatever I can observe.
@NVAccess
Oh I see where the problem lies now. Chrome is showing me many more attributes than Firefox does. Do you think the Firefox accessibility tree is composed differently? Nevertheless: Chrome still shows me name-from=attribute without revealing what attribute it is, on a link labelled by aria-label. Also it refers to a description such as aria-describedby as "label" which confused me a bit as label refers to something else, like the labels in forms or aria-label etc Tested on amazon.com and their Keyboard Shortcuts navigation menu links at the top of the page..
@Piciok Yes, Firefox and Chrome do compose their accessibility trees differently. Even though both can expose IA2 on Windows, their internal mappings and which attributes they surface are not identical. So differences in exposed attributes between the two browsers are expected.
Regarding Chrome showing name-from=attribute without specifying which attribute, that is a limitation of the IA2 attribute itself. IA2 typically reports categories such as contents or attribute, but it does not indicate whether the name came specifically from aria-label, aria-labelledby, or another author provided source. So if Chrome reports attribute, that is unfortunately as granular as the API makes it.
As for aria-describedby being referred to as label, that can indeed be confusing. In the accessibility computation model, aria-describedby contributes to the accessible description, not the name. However, some Chromium IA2 mappings use terminology that can blur that distinction when exposed through object attributes. What you are seeing is likely reflecting that browser level mapping rather than a misinterpretation by the Inspector.
Your testing on Amazon is very useful. If you can share a specific element example where the mapping looks particularly odd, I would be happy to look at it more closely.
Regarding Chrome showing name-from=attribute without specifying which attribute, that is a limitation of the IA2 attribute itself. IA2 typically reports categories such as contents or attribute, but it does not indicate whether the name came specifically from aria-label, aria-labelledby, or another author provided source. So if Chrome reports attribute, that is unfortunately as granular as the API makes it.
As for aria-describedby being referred to as label, that can indeed be confusing. In the accessibility computation model, aria-describedby contributes to the accessible description, not the name. However, some Chromium IA2 mappings use terminology that can blur that distinction when exposed through object attributes. What you are seeing is likely reflecting that browser level mapping rather than a misinterpretation by the Inspector.
Your testing on Amazon is very useful. If you can share a specific element example where the mapping looks particularly odd, I would be happy to look at it more closely.
I see. Yes, in that case let's leave it at that - I tried with the "Cart" link but any of those in that navigation will do, actually. Maybe it would be possible to add an accordion in the inspector view to display the complete HTML of the element in that case?
@Piciok Yeah, this is actually something I'm considering. I should do it in a way that it doesn't end up adding a lot of noise.
What is still running around in my head, but this is completely for when you'll have the time to implement less useful features like this, a way of requesting subsequent properties of the focused element with quick hotkeys so that, for example, NVDA+ shift + 1 gives the first property, 2 the second etc. Note that the keystrokes are provisional and may be replaced with something else or even unbound by default.
@Piciok Good idea I'd say! I'm still trying to see if I can find a way to extract stuff from DOM without giving NVDA headaches - LOL!
@Piciok I forgot to mention that right now the inspector already supports heading navigation, so you can jump directly to sections like Name, States, Relations, and so on. That covers most of the same workflow with much less maintenance cost.
I will keep the “quick property hotkeys” idea in mind for the future, but for now I would prefer to keep the inspector simple and rely on the existing heading based navigation, since it already provides a fast way to reach the information you need.
I will keep the “quick property hotkeys” idea in mind for the future, but for now I would prefer to keep the inspector simple and rely on the existing heading based navigation, since it already provides a fast way to reach the information you need.
Yeah, this a great addition! THANKS for all of your work.
@ppatel any way you could have it retrieve current CSS classes affecting the element as well? Harder for sure, but would be very useful
Either GitHub's mobile layout is lying to me, or this repo only contains releases but no code. Given the increasingly miserable state of GH accessibility, it could be either.
@jscholes V1.1.1 does provide the code. In fact, I'll be doing that for all of the add-ons I release from now on.
@jscholes LOL - no. It's not lying to you. As I was working on 4 add-ons simultaneously, I couldn't handle the repo/ code upload due to time constraints. I decided to take care of the releases first, and then upload the code.
2026-02-25 08:08:12
2026-02-25 08:08:12
2026-02-23 19:26:06
9813722
Peter Vágner reshared this.
Colleagues from the European Parliament in Luxembourg are seeking a digital accessibility expert via CTG:
en.moovijob.com/job-offers/ctg…
If you know anyone who might be interested, please let me know!
#Job #Luxembourg #a11y #FediHire
#GetFediHired #hiring
Web Accessibility Consultant at CTG IT Solutions in Bertrange
Discover the Web Accessibility Consultant job ad at Bertrange, and apply online! The company CTG IT Solutions is currently recruiting.Moovijob.com
This entry was edited (Monday, February 23, 2026, 7:29 PM)
reshared this
Peter Vágner reshared this.
2026-02-27 08:23:21
2026-02-25 05:11:14
2026-02-24 22:46:50
9810338
Peter Vágner reshared this.
#movuary 24 Too Many Wires
The media in this post is not displayed to visitors. To view it, please go to the original post.
#movuary 24 Too Many Wires
reshared this
Andre Louis, Peter Vágner and Borris reshared this.
2026-02-25 07:04:23
2026-02-25 05:07:45
2026-02-25 01:31:46
9811296
Peter Vágner reshared this.
in reply to Jamie Teh
in reply to Tamas G
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Amir
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Pratik Patel
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Christopher Duffley
in reply to Jamie Teh
FFS. The GitHub issue list is also just awful now. Half the time, you try to focus or open an issue and it throws focus to or opens some other issue. And don't even try opening issues in a new tab. GitHub honestly just sucks lately and I hate it. I feel like they're just gradually making accessibility shittier and shittier.
Peter Vágner reshared this.
oh no. Reading these kinds of things hurts my soul, I know eventually that kind of thing will make it to our internal enterprise server and then I know my life will truly be thrown into chaos again randomly. Ha.
@Tamasg The issue list has sucked for ages now, but on top of the pull request brokenness, it's just too much. I think I'm going to need to get back into some serious Greasemonkeying.
@Tamasg The open in new tab behavior is ridiculous and I don't even know how they did that. I have to use aplications now.
I've had to train myself to hit the Applications key on the first h3 in the list, close the context menu and then the rest of the issue title links seem to behave. There are probably other ways to achieve the same thing but the whole experience is painful.
This problem with issues and new tabs has been there for a couple of months.
@amir Yeah, I know. But combine that with the pull request clusterzarq and the mounting pile of other GitHub accessibility issues and it just makes using GitHub a real drag. It's a shame because I previously found it reasonably pleasant, but they appear to be doing everything they can to screw that up.
@Amir
Agreed 100 percent! Accessibility is getting worse there.
@ppatel @amir Right, similar thing with the pull request experience. They were trying to make it possible to navigate lines with the arrow keys, but they made managing comments an utterly awful experience in the process. There's no point in trying to do something cool with accessibility if you're going to screw up the basics.
And this is why I love GH on the command line. I use it every time I want to look at PR's and issues and things. I will only use the web option in a command to comment and the like. That's about it.
@ChrisDuffley Fair. The problem is that this is all stuff I'd need to do via the UI: hunting through huge sets of issues (kinda hard to do from a bunch of command line output), code review (lots of comments and threads), etc.
Ah yes, that makes sense.
2026-02-24 23:00:52
2026-02-24 23:00:52
2026-02-24 10:54:51
9807111
Peter Vágner reshared this.
in reply to Pachli
in reply to Pachli
in reply to Pachli
I've started looking into the issue reported by @dhamlinmusic in dragonscave.space/@dhamlinmusi…
If you use Pachli with #accessibility services like #Talkback enabled, what would be the most useful thing to read out when reading a post with a quote?
A few options I'm considering, but could well be wrong.
1/2
D.Hamlin.Music (@dhamlinmusic@dragonscave.space)
@pachli@mastodon.social @pixelate@tweesecake.social So quoted posts are appearing as what talkback reads as "detected image post content" below the post by the person quoting them, but have no indication they are a quote, they just show up like a pos…D.Hamlin.Music (The Dragon's Cave)
reshared this
Peter Vágner reshared this.
1. Read out the post's content first, then read out something like "They quoted a post from X at Y". You would have to open the quoted post for its content.
2. Like 1, but read out the quoted post's content too.
3. Read out the quoted post's content first, then read out the content of the post that quoted it.
4. Something else?
2/2
i think 2 is the most balanced. for example, aryan, post content. quoted from quote author, quoted content. I hope it makes sence.
I'm leaning towards 2, but 1 would work for me as well.
2026-02-24 22:57:29
2026-02-24 22:57:29
2026-02-24 09:30:16
9807897
Peter Vágner reshared this.
in reply to Neil Brown
in reply to Henryk Plötz
in reply to Neil Brown
in reply to Henryk Plötz
in reply to Neil Brown
in reply to Neil Brown
in reply to Neil Brown
in reply to Colman Reilly
in reply to Vysogota
in reply to Vysogota
in reply to Neil Brown
in reply to ahnlak
in reply to Neil Brown
in reply to Neil Brown
in reply to Neil Brown
in reply to Andrew
in reply to Neil Brown
in reply to Neil Brown
in reply to Neil Brown
in reply to Neil Brown
in reply to Dave
in reply to Neil Brown
in reply to v
in reply to Neil Brown
in reply to Chao-c'
in reply to Neil Brown
in reply to unqualifiedtechbros
in reply to Neil Brown
in reply to Neil Brown
in reply to Neil Brown
in reply to Neil Brown
in reply to Windy city
in reply to Windy city
in reply to Neil Brown
in reply to Neil Brown
in reply to Neil Brown
in reply to *|FNAME|*
in reply to Neil Brown
in reply to Neil Brown
in reply to Neil Brown
in reply to Neil Brown
in reply to Neil Brown
in reply to Neil Brown
in reply to Neil Brown
in reply to Neil Brown
in reply to Neil Brown
in reply to Neil Brown
in reply to Neil Brown
in reply to Neil Brown
in reply to Neil Brown
in reply to Neil Brown
I self-host a lot of stuff. Nearly everything that I use. FOSS and self-hosting is a massive part of my computing experience.
I love reading about people enjoying / exploring self-hosting stuff.
I struggle when people advocate "just self-host it", without giving due consideration to the costs, risks, security considerations, and so on.
I know that I've posted this a few times now, but this discussion seems to pop up quite a lot. So:
neilzone.co.uk/2022/07/self-ho…
#SelfHosting #FOSS #blog
'Self-host it' is not the answer
I’m going out on a limb a bit here. I love self-hosting my own systems.neilzone.co.uk
This entry was edited (Tuesday, February 24, 2026, 9:40 AM)
reshared this
Hubert Figuière, Mudb0y and Peter Vágner reshared this.
Me too. The best comparison I heard (was that you?) was: backyard chickens.
Some people have these. It gives them certain benefits, not the least having something to care for. But they *need* to be cared for.
Some people have these. It gives them certain benefits, not the least having something to care for. But they *need* to be cared for.
@henryk Interesting analogy!
@henryk the chicken analogy is great, when it becomes apparent that a system of distribution is needed to sell or provide eggs to others. And also trust and confidence on the part of the egg recipients outside one's household - that too is no trivial thing. There's psychology and relationality there, which goes beyond the sheer technical aspects of raising the chickens.
@henryk I was just thinking about gardening. I grow veg because I want to, not to save money or because I don't trust the shop.
I self host some things (but mostly only on my LAN) and use a mix of self managed and provided services for other things. I'd rather have a healthy ecosystem of good providers people can choose instead of using Google and Microsoft for everything, rather than expect everyone to self host.
as a parellel: "quit YouTube to move to Peertube" misses one of the main value proposal of YouTube: you don't have to pay hosting cost. Non-withstanding the audience.
I'm in the same boat, and I always try to clarify that it is one of my hobbies, so like others love to spend money and time on, for example, skiing, I do the same for self hosting. And I don't expect of anyone that they should just do the same, ESPECIALLY if they lack technical knowledge or joy of debugging
This entry was edited (Tuesday, February 24, 2026, 11:56 AM)
“just host it yourself” is techno-libertarian nonsense. It’s like the Linux crowd’s “all you need to do is recompile the flux capacitor with the flibertygibbet flag and reverse polarity on the ant hill and if you can’t do that you don’t deserve to have a computer“ stuff.
It’s also notably atomising. Community level hosting is one thing — and has its own problems when confidentiality is an issue — but self hosting is just out of reach of most people.
This entry was edited (Tuesday, February 24, 2026, 9:47 AM)
@Colman
As opposed to 'just pay the faceless corp instead of self-hosting.' This is notably not techno-libertarian at all, very inclusive, and the opposite of atomizing --
just look at Facebook and Twitter after all, they are the *Threads* that knit together our nations!!!
If self-hosting shit is anything like a political statement, I'd say it's closer to anarchy than libertarianism.
This entry was edited (Tuesday, February 24, 2026, 11:55 AM)
@petko maybe “host it yourselves” is the techno-anarchist version.
@petko @Colman Agreed. If you need something, like some black-box to share files with other people — you could just make something yourself with existing technical skills. Or organize with some other people with necessary skills and make a thing. This looks like some kind of anarchy and definitely not like "techno-libertarianism" lol.
The "how-to" knowledge is freely acessible and no one hides it with paywall — you can learn or get help from other people (for beer or money, etc) if you don't have a time/courage to learn.
I think self-hosting is a bit like doing your own car servicing.
Sure, you can do it and if you know what you're doing you'll probably be happy and saving yourself some money - but unless you *actually* know what you're doing, many things can and will come back to bite you in the ass 😉
Hell, I *do* know what I'm doing and I'm very happy to pay someone else to do quite a lot of things I could, theoretically, do myself.
@ahnlak I like this metaphor! I don't have a car, but I have a bike and like to make a lot of maintenance by myself. And this definitely looks like self-hosting — if you able to have fun with wrenches and grease, then you will do it yourself and will have fun. Otherwise, no fun 
This is a very thoughtful and balanced article.
It's great though that there are things like #YUNOhost, lowering a lot of the technical barriers.
Still, I have had to learn a lot and put in quite a lot of time. And I haven't exactly done it for fun - more for ethical reasons.
It would be nice if there were more providers of online services to choose from, and you could simply choose a trustworthy one for a reasonable fee. Not everyone should have to host but we need a certain number to.
It's great though that there are things like #YUNOhost, lowering a lot of the technical barriers.
Still, I have had to learn a lot and put in quite a lot of time. And I haven't exactly done it for fun - more for ethical reasons.
It would be nice if there were more providers of online services to choose from, and you could simply choose a trustworthy one for a reasonable fee. Not everyone should have to host but we need a certain number to.
Tee hee!
Great info for those of us in Brexitland/Europa
It takes me a long time BUT:
- Moving to FLOSS
- Moving to sustainable and interoperable hardware
- Supporting next generation of hacktavists
I self host, knowing the risks and challenges. Those same risks and challenges are why I don't push it on others, but I will fully support anyone who does want to take it on.
lots of middle ground people can do first before trying self-hosting.
use whatsapp --> use signal --> run a matrix server
ms office --> libreoffice/only office --> self-host collabra :D
lol
hard to get them off that sweet crack cocaine of "convenient" big tech stuff though.
Great piece. I am a year into self-hosting and did it entirely for the learning experience.
I quickly realised that I wasn't confident enough in my security to use Nextcloud or Vaultwarden and am aware that I have probably made my home network vulnerable to a concerted attack.
So it has ended up very much a small scale hobby, not hosting anything confidential or essential to day-to-day functioning (where I pay for Proton, Signal, Bitwarden etc.).
I have self-hosted stuff in the past, and occasionally think about doing so again for some stuff. But less and less of my computering is recreational and I choose to pay companies to worry about everything for me rather than spend my spare time doing more of what I do when I'm working.
Which is, of course, a kind of privilege in itself.
None of that is a criticism of course, just a reality that I've resigned myself to and remind myself of every time I think "oh, I could set that up myself".
I have built and operated my own hardware firewall, and it worked wonderfully, until my ISP changed my modem to a non-bridgeable one
I have built (numerous times) a software server on my laptop to host a local blog. I wrote tutorials to myself so I can replicate it on the next occasion—which always necessitated a refreshed tutorial
I want to have a hardware firewall between me and the internet. I want a non-google email. I kind of want my own Fediverse instance
I don't have the time, am low on money, and have few spoons
I will find that firewall, but it will likely be a secondhand commercial rig, and my networking knowledge/skills aren't there
I want to build a DMZ for IoT things, run my own cameras, keep my smart thermostats from calling home every fifteen minutes
But I can't find the energy to figure out how to freaking use the thermostats I have
I'm with you 100%—I could self-host... most things
But I can't, actually
and you gotta hope the tools you’re using is actually secure
in a better timeline, i'd like the idea of neighborhood/community hosting. like, the local library doing it, with help of tech savvy volunteers. that would be wholesome
@NocturnalNessa We, in Russia had such thing before (and looks like will have in the future 
), when the ISP prices were high and unaffordable for usual people, but there were a lot of coaxial cables and 10BASE2 (IIRC) network equipment on the market for a reasonable price. So, people just connected apartments and buildings with these cables and had a local network in the block (or between some blocks) for filesharing, gaming in local network, etc… Something like this: medium.com/@pv.safronov/moscow…
), when the ISP prices were high and unaffordable for usual people, but there were a lot of coaxial cables and 10BASE2 (IIRC) network equipment on the market for a reasonable price. So, people just connected apartments and buildings with these cables and had a local network in the block (or between some blocks) for filesharing, gaming in local network, etc… Something like this: medium.com/@pv.safronov/moscow…
@v
"It appears that “self-hosting” might mean different things to different people."
Definitely. We are able to do #selfhosting because we are local ISP for 25 years. We run our own datacenter from at least 8 years, and I was planning how to do the waste heat recycling for 10 years. We have access to second-hand hardware, because it would be hard to afford only the new hardware.
So #selfhosting is still collective activity of some kind, you need small business, or cooperative, or something, because you have expenses.
I suppose aggregating VPS-es for multiple instances on same physical hardware may have some advantages eg. with containers sharing the same ZFS filesystem (you can turn on eg. de-duplication of files)
So the #selfhosting usually anway about seeking some kind of cooperation. I would not be able to "selfhost" without cooperating with experienced admin...
This entry was edited (Tuesday, February 24, 2026, 1:31 PM)
@xChaos Oh wow — this is *super* interesting to me! How many people are involved in maintaining your local ISP? And it is cooperatively owned?
Do you recommend yunohost?
@unqualifiedtechbros I have no experience of it, I'm afraid.
That is an excellent article.
I do think that you can self-host on someone else's computer, shrinking all the financial costs to a couple of small ones (hosting fee and domain name; just those are cheaper than having any kind of home internet connection).
I run a root server in the cloud, on an OpenStack instance. The entire machine is defined by me. I regard it as self-hosting.
I think this summarizes why I'm suffering from more or less severe burnout symptoms on a daily basis.
When I'm done developing depressingly poor quality software projects designed to "generate" revenue for a greedy tech company for the day, I am expected to keep on top of the never-ending bot attacks against my self-managed services.
Every time I install a new kernel version or significant package updates, my anxiety peaks while I wait for machines and services - some of them being the backbone of my digital existence - to come back up.
I'd rather just pay for someone else to do it, but experience has shown that companies can't be trusted with anything anymore as increasing profits has become the driving force behind everything with quality and reliability being downgraded to a mere afterthought.
I hate that system and I hate being part of it but I have to pay my ever growing greed and inflation driven bills so what can I do?
@Neil Brown And (I'm yet to read the full article, but) the upcoming raise in hardware prices will make the entry price on self-hosting even more complicated to reach
Neil is absolutely spot on and we need to talk about the politics of this. Telling people to "just self-host" to escape surveillance capitalism is the exact same tactic as telling people to "calculate your carbon footprint" to stop climate change.
It individualizes a systemic, corporate failure.
Privacy and digital safety shouldn't be luxury goods reserved for cis-white-male engineers with unmetered fiber connections, rack-mount servers along with free time and financial stability to pull it off. When our response to the collapse of digital rights is "run your own infrastructure," we are engaging in digital redlining. The solution to predatory data brokering isn't forcing single mothers working two jobs to learn Kubernetes; the solution is ruthlessly regulating the data brokers out of existence.
@pheonix otoh, the history of eg email might be very different if companies had made it easy to self-host. They chose not to.
100% agreed, and this is coming from someone who has essentially set up a private internet for himself by now.
Feel like at the end it's yet another ongoing commitment, and people might not have the time, energy, and/or spoons to tend to it - regardless of their technical competence 
(Though big respect to those who do!)
This entry was edited (Tuesday, February 24, 2026, 1:18 PM)
Thanks for sharing. This is useful info :)
Sort of off-topic (and I agree that not everyone has the skillset/desire to self-host) but I followed your guide to upgrade my GlitchSoc instance and it was great. Thank you for posting it.
This entry was edited (Tuesday, February 24, 2026, 1:51 PM)
@fname I am so pleased - thank you!
When I was young there were organisations offering "shell servers" where you could run your software and host your website in a subdirectory of your home directory. That all was done for minimal amounts of money.
Maybe we need something like that again in the age of ad-phones.
Maybe we need something like that again in the age of ad-phones.
very good reminder for these days... i'm seeing a huge influx of people ditching windows for linux lately (for obvious reasons)
But instead of just taking the time to learn linux a lof of people are diving straight into self-hosting... without learning the fundamentals of system administration (i say this as a very amateur sysadmin). They just run 'docker compose up -d' and boom, you're a self hoster...
But when things inevitably go wrong they turn around and blame the software or even the OS.
I can agree on this, self-hosting is a different kind of rabbit hole.. requires near constant upkeep, and you spend most of that time looking at logs, analyzing, optimizing and improving what you already have..
You follow security checks and best practices to secure your self-hosted infra..
It's not for the faint of heart, can get daunting at times, but the result is satisfaction and knowing you took that first step forward from straying away from bigTech bros.. you feel free somehow..
Stepping out on a limb myself here. Here’s my response on why I think self-hosting *IS* an answer. But to understand this we need to, A. First understand the question being asked, and B. Not be overly literal or non-inclusive in terms of how we define “self-hosting”
shellsharks.com/notes/2026/02/…
Key concept: We don’t have a great vocabulary for describing the middle ground between pure “I own the hardware” self-hosting to “I use big tech platform A”. Let’s agree there’s self-hosting spectrum.
so I don’t self host, but that doesn’t absolve me of choices in what/where tech I buy, and I’ll tell people why I’ve made those choices - somebody has to pay (including volunteer/hobby time) somewhere. Is it £3/month to FB (current UK ‘don’t consent’ price) or the same to this mastodon instance? DropBox/OneDrive/Google, or hetzner for Nextcloud (and actually this one worked out cheaper for us)… Gmail or Runbox… GitHub pages or mythic beasts… the value is worth it.
This entry was edited (Tuesday, February 24, 2026, 2:35 PM)
"Self-hosting requires a heck of a lot of privilege."
This can be true, but it's easier (and probably cheaper) than ever before.
And I'm doing it mostly to just say no to corporations getting my data.
It's the online version of "just fork it", totally belies the amount of work that implies and some folk don't have that dog in them.
yeah and it even do not fix the core issue, of even if you self host, it will not allow grandma to join and such condemn it to never be mainstream.
linux user made this error back and the day and now that it's easy to use "strangely" the usage skyrocket.
couldn't agree more. Self-hosting is a hobby, and not something that most people will want to do, or should have to.
Self-hosting is ok, but it's a very individualistic perspective. Associative hosting is a very good thing for a lot of people. And of course the classic of having someone host something for a group of family or friends
self hosted EU soverign torment nexus
I do self-host a few services too, but there are several projects that I would love to self-host, but their setup is just not worth the effort. Like Overleaf, that documents their upgrade path as basically "yeah, drop the old install and start from scratch – or use Docker¹". Others like Gitlab CE are just PITA, introducing more issues in every release instead of fixing existing ones. There're enough situations, where self-hosting just doesn't fit …
I agree. I tried to self host email and IM services, but found that it's not efficient (in terms of costs and efforts)! I do self host lots of other things though!
I think we shouldn't see self-hosting as the goal, but as an act of resistance in an hostile world.
And I agree that everyone who can afford to selfhost at home are the lucky ones.
2026-02-24 22:54:25
2026-02-24 22:54:25
2026-02-24 17:24:06
9808202
Peter Vágner reshared this.
in reply to Amir
in reply to Dennis long
I’m proud to present the second add-on in my four-part NVDA add-on series.
Links Management is an opt-in NVDA add-on that improves Browse Mode navigation when Screen Layout is turned off. When enabled, each link behaves as a separate line for Up and Down arrow navigation. Trailing punctuation is intelligently attached to the link line to prevent blank or punctuation-only lines. Inspired by JAWS, this add-on focuses on clarity, predictability, and navigation control.
A dedicated settings panel is available at NVDA menu → Preferences → Settings → Links management.
github.com/amirsol81/links-man…
@NVAccess
Links Management is an opt-in NVDA add-on that improves Browse Mode navigation when Screen Layout is turned off. When enabled, each link behaves as a separate line for Up and Down arrow navigation. Trailing punctuation is intelligently attached to the link line to prevent blank or punctuation-only lines. Inspired by JAWS, this add-on focuses on clarity, predictability, and navigation control.
A dedicated settings panel is available at NVDA menu → Preferences → Settings → Links management.
github.com/amirsol81/links-man…
@NVAccess
GitHub - amirsol81/links-management: Links Management for NVDA is an opt-in add-on that improves Browse Mode navigation when Screen Layout is off. It forces each link to behave as a separate line for Up/ Down arrow movement, while intelligently attaching
Links Management for NVDA is an opt-in add-on that improves Browse Mode navigation when Screen Layout is off. It forces each link to behave as a separate line for Up/ Down arrow movement, while int...GitHub
reshared this
Peter Vágner reshared this.
This is a home run. NVDA couldn't do this years ago why? Nice Job!!!!!!
2026-02-24 22:52:38
2026-02-24 22:52:38
2026-02-24 13:07:15
9808685
Peter Vágner reshared this.
in reply to Frederik Braun �
in reply to Ricard Torres 👨💻
in reply to Frederik Braun �
in reply to Frederik Braun �
in reply to JA
New blog post 😊
If you replace all the innerHTML with setHTML, you will be free from XSS and other injection attacks. Goodbye innerHTML, Hello setHTML
hacks.mozilla.org/2026/02/good…
(Kudos to our folks for specifying, building and shipping!)
Goodbye innerHTML, Hello setHTML: Stronger XSS Protection in Firefox 148 – Mozilla Hacks - the Web developer blog
Cross-site scripting (XSS) remains one of the most prevalent vulnerabilities on the web. The new standardized Sanitizer API provides a straightforward way for web developers to sanitize untrusted HTML before inserting it into the DOM.Tom Schuster (Mozilla Hacks – the Web developer blog)
This entry was edited (Tuesday, February 24, 2026, 5:06 PM)
reshared this
Peter Vágner reshared this.
Can't wait for setHTML to have baseline support.
@dev Another browser is going to follow really really soon :)
Congrats! News are only highlighting the AI switch button, not the TT support 🙈
2026-02-25 14:27:38
2026-02-24 22:24:04
2026-02-24 19:58:35
9808943
Peter Vágner reshared this.
in reply to Jonathan
in reply to Peter Vágner
in reply to Peter Vágner
in reply to Jonathan
in reply to Peter Vágner
So thhat post I wanted to make about my experience with Graphene OS a month ago? Yeah it exists now.
jonathan859.mataroa.blog/blog/…
Please give me feedback because I'm really not convinced in its quality/usefulness.
Peter Vágner likes this.
reshared this
Peter Vágner and Tamas G reshared this.
@Jonathan Nice article. I had very similar experience a few months ago. My only difference is that I'm running this on Pixel 9A.
In order to give some more tips on what to explore next I'd say add @Izzy repo to your f-droid install from apt.izzysoft.de/fdroid/
Then some nice and accessible #opensource apps are:
@AntennaPod
#auroraStore as a google play alternative,
#catima - a privacy respecting app for lojalty cards,
@DAVx⁵ 🔄 - contacts / calendar sync from your own server if you wish to slowly stop using google for that,
#fairEmail - nice very accessible email client,
#makeacopy - for scanning paper documents,
#newpipe - for playing youtube / bandcamp / soundcloud content,
#openKeichain - for signing / encrypting your emails with fairEmail,
#rsaf - for connecting to various cloud storage services including your own smb / ftp / webdav servers,
sms import export,
Speak that! if the talkback notification presentation is not enough,
walkers guide - nice navigation app
Ytdlnis - an yt-dlp frontent.
Perhaps there are more but I have browsed my apps list and recommended what I think is essential.
@pvagner @AntennaPod @davx5app we also have some nice client apps for F-Droid repos, which include download stats and show you which apps are (re)built reproducibly by our builders, see the screenshots at izzyondroid.org/ (and head over to izzyondroid.org/quickstart/ for details on them) And if you want to check our download stats with your browser first, head over here: dlstats.izzyondroid.org/ 
Peter Vágner likes this.
@pvagner Yup, I knew a few of them but not all. A great list for sure. Please correct me if this has already been done, but I was thinking about making an english awesome android accessibility list about good and accessible android apps, similar how @radiorobbe has done it in german, great resource btw. No idea if that already exists, but it seems much more logical to me than having to dig through random accessible android articles which, imo aren't very logical sorted, and neither is the site very responsive at this point.
@Jonathan There is blind android users list at groups.io/g/blindandroidusers
I'm not subscribed to it as it's a continuation of so called eyes free list hosted at google groups a few years Ago.
My guess is that the community managed by @Steffen is much more organized.
I'm not subscribed to it as it's a continuation of so called eyes free list hosted at google groups a few years Ago.
My guess is that the community managed by @Steffen is much more organized.
@pvagner Thanks, but that's far too much of an honour. 🙂 I only manage a small German-language Telegram group with around 20 members, and the app list on GitHub and an associated website at droidblind.de. There's also a German mailing list hosted at BLINDzeln project, but I'm just an ordinary member there.
Peter Vágner likes this.
2026-02-24 08:43:07
2026-02-24 08:43:07
2026-02-24 08:43:07
9804802
Hey #showdown fans,
By any chance if you are coming to Showdown Slovak Open this year, please let me know so I can say hello to you.
2026-02-24 08:26:43
2026-02-24 08:17:21
2026-02-23 22:23:44
9804274
Peter Vágner reshared this.
in reply to Noah Tobias Carver 👨🏼🦯🇺🇦
in reply to Noah Tobias Carver 👨🏼🦯🇺🇦
in reply to Noah Tobias Carver 👨🏼🦯🇺🇦
in reply to Noah Tobias Carver 👨🏼🦯🇺🇦
in reply to Simon Jaeger
Heyah. What #VideoEditing software are #Blind #ScreenReader users using in 2026? Would appreciate any input. Thanks. #BlindMasto #BlindMastodon #BlindFedi
reshared this
modulux, André Polykanine and Peter Vágner reshared this.
Reaper and finalcut pro. These are 2 options I've heard of in which you obviously know of already. Have a great day, Noah. You're awesome.
This entry was edited (Monday, February 23, 2026, 10:26 PM)
I use Reaper with BVDA. Also I created videos with PowerPoint saving presentations in video formats, but it is a slower process. Microsoft ClipCham is working well on Windows 11, but the interface is more complex.
@Noah Tobias Carver 👨🏼🦯🇺🇦 On linux back in the days I was using an app called #avidemux. It worked fairly well.
I was also trying out #kdenlive however the version I was using has been crashing on my system so I've stopped playing with it.
Nowadays I am no longer dealing with videos and for simple cutting both audio and video I'm just using #losslesscut.
I was also trying out #kdenlive however the version I was using has been crashing on my system so I've stopped playing with it.
Nowadays I am no longer dealing with videos and for simple cutting both audio and video I'm just using #losslesscut.
@kaveinthran Reaper. It's the only thing that seems relatively foolproof.
There's another program called Lossless Cut, but because its aim is primarily to losslessly edit files, it can be hard to make seamless edits to videos without making the video drift out of sync--which is not something we can easily verify.
Reaper imports video files, and as long as your edits are removing parts of that video track, the video will match the audio. You can completely replace the audio track by turning the video track's volume down to nothing. (I don't know if muting also works; it might. Not something I could be bothered testing.)
There's another program called Lossless Cut, but because its aim is primarily to losslessly edit files, it can be hard to make seamless edits to videos without making the video drift out of sync--which is not something we can easily verify.
Reaper imports video files, and as long as your edits are removing parts of that video track, the video will match the audio. You can completely replace the audio track by turning the video track's volume down to nothing. (I don't know if muting also works; it might. Not something I could be bothered testing.)
@simon @kaveinthran Muting will cause the item to not be displayed FYI. Better to hit Control+F2 on the video item and set Audio to disabled. There's also a button in there to apply that to all video items in one shot if needed. Hope it helps!
2026-02-24 07:16:56
2026-02-24 07:16:56
2026-02-23 21:49:57
9804277
Peter Vágner reshared this.
#TIL London Underground Map was inspired by an electrical circuit.
#London, Design Museum.
in reply to Miroslav Suchý
The media in this post is not displayed to visitors. To view it, please go to the original post.
#TIL London Underground Map was inspired by an electrical circuit.
#London, Design Museum.
Peter Vágner reshared this.
I am reading your post and suddenly hearing a Map Men jingle in my mind
cc @jayforeman
youtube.com/playlist?list=PLfx…
MAP MEN
Buy MAP MEN MERCH from http://www.mapmenmerch.com Follow us on twitter: http://www.twitter.com/jayforeman http://www.twitter.com/markcooperjonesJay and Mark (YouTube)
2026-03-01 14:14:10
2026-02-23 19:58:53
2026-02-23 11:34:58
9800424
Peter Vágner reshared this.
in reply to Pratik Patel
in reply to Liam Erven
in reply to Pratik Patel
in reply to Liam Erven
in reply to Pratik Patel
in reply to Liam Erven
in reply to Pratik Patel
in reply to Yadiel Sotomayor
in reply to Pratik Patel
in reply to Yadiel Sotomayor
in reply to Pratik Patel
in reply to Andre Louis
in reply to Pratik Patel
in reply to Andre Louis
in reply to Pratik Patel
in reply to Andre Louis
in reply to Pratik Patel
in reply to Andre Louis
in reply to Andre Louis
in reply to Pratik Patel
in reply to Andre Louis
in reply to Pratik Patel
in reply to Captain Jack Sparrow
in reply to Pratik Patel
in reply to Pratik Patel
in reply to Pratik Patel
in reply to Quin
in reply to Tuukka Ojala
in reply to Tuukka Ojala
in reply to Tuukka Ojala
in reply to Pratik Patel
in reply to Pratik Patel
in reply to Pratik Patel
in reply to Pratik Patel
in reply to Pratik Patel
in reply to Pratik Patel
in reply to Bri🥰
in reply to Pratik Patel
in reply to Arfy
in reply to Pratik Patel
in reply to Pratik Patel
in reply to Pratik Patel
in reply to Pratik Patel
in reply to Pratik Patel
in reply to Pratik Patel
in reply to Pratik Patel
in reply to Joseph Lee
in reply to Pratik Patel
in reply to x0
in reply to Pratik Patel
in reply to Kaveinthran
in reply to Pratik Patel
in reply to Pratik Patel
in reply to x0
in reply to Pratik Patel
in reply to x0
in reply to x0
in reply to Kaveinthran
in reply to Pratik Patel
in reply to x0
in reply to Pratik Patel
in reply to The Ghostly Broken Panda
in reply to Pratik Patel
in reply to Cary ✨
in reply to Pratik Patel
in reply to Pratik Patel
in reply to Brandon
in reply to Cary ✨
in reply to Pratik Patel
in reply to Brandon
in reply to Pratik Patel
in reply to Brandon
in reply to Pratik Patel
in reply to Brandon
in reply to Pratik Patel
in reply to Brandon
in reply to Pratik Patel
in reply to Bill Dengler
in reply to Pratik Patel
I’m excited to announce the first public release of Terminal Access for NVDA: an add-on focused on improving the NVDA experience when working in terminal/console applications. Inspired by TDSR and Speakup, there's quite a bit of functionality here. I welcome contributions. Issues and PRs welcome. Lots of testing needed.
Open issues on the Github repo at: github.com/PratikP1/Terminal-A…
Get it here: github.com/PratikP1/Terminal-A…
GitHub - PratikP1/Terminal-Access-for-NVDA: This NVDA add-on enables functionality similar to TDSR in NVDA.
This NVDA add-on enables functionality similar to TDSR in NVDA. - PratikP1/Terminal-Access-for-NVDAGitHub
reshared this
Andre Louis, Pitermach, Khronos, Tamas G, 🇨🇦Samuel Proulx🇨🇦, modulux, André Polykanine, Oskar, LWorks, Mudb0y, Peter Vágner, johann, Derek Lane, Patryk Mojsiewicz and Tissman reshared this.
I will ge tit right now. lol
@lerven Lol. Lots of morning energy. Eh?
I work in terminal a lot. This would be helpful.
@lerven It's made for power users like you. Feedback is welcome. I already know of couple of issues. It's one of those things where I had to release before I went crazy.
Wil this do anything to ease the issue NVDA has with stalling/hanging when terminal becomes large?
@lerven Yes. It should. I've got a bit of a cacheing system going on. You can also put it in quiet mode while the screen is updating.
@lerven Does this improve access to WSL through the terminal?
@ysotomayor @lerven Yes. It should improve support for WSL and WSL2. I suspect WSl2 is better supported.
@lerven Oh! Perfect! An excuse to play with WSL instead of doing actual work! Thank you!
@ysotomayor @lerven Play with the different profiles to see if they work better in WSL. I have a WSL specific profile. It should launch automatically.
One issue immediately is that all of those keys are being stolen by the addon, so my guess is it's not using appmodules to detect them or something.
Please bare in mind I am an ignorant buffoon and know nothing about coding of any sort, everything I know about NVDA coding has been done by LLM, but this feels like that might be the issue.
For example NVDA+F3 doesn't usually do anything apart from in browse mode. When Terminal Access is running, it stole the hotkey when I was in a browser, and made it behave oddly. It also flipped the state of caps lock, my desired hotkey.
Since it's not a terminal window, it should probably not be triggering there.
Again, take everything I say with 0 understanding of what I@m on about because I'm a musician and know nothing.
@FreakyFwoof Ah. Crap. I thought I'd changed the NVDA+f3 combo. I'll double check that.
All the other keys show up as terminal specific as well, so you might want to consider that too if possible.
@FreakyFwoof The other keys should be terminal specific and not global. Somehow, the search keys got coopted by global.
Nope, they all show up anywhere on the system, at least when I checked.
@FreakyFwoof Is this on Windows 10 or 11?
10. Easy way to know is going into NVDA+1 and pressing a key when *not*focused on a terminal window.
@FreakyFwoof This is definitely a problem. I'll get right on it. Thanks for letting me know.
@FreakyFwoof I fixe it. Apparently there was a regression. I'd managed to fix it at one point and then somehow reverted it. Download the latest. Let me know if you have more trouble.
github.com/PratikP1/Terminal-A…
Release Release 1.0.45 · PratikP1/Terminal-Access-for-NVDA
Changed Raised minimum NVDA requirement to 2025.1 (Python 3.11 runtime); compatibility with earlier versions removedGitHub
How are you coding this out of interest?
@FreakyFwoof It's a combo of manual coding and Claude/Codex. I'm not confident enough in my Python writing skills that I can avoid LLMs
I'm sure there are some people on here who know wtf you're talking about.
I'm not one of them.
@Captain_Jack_Sparrow Luckily for me, there are plenty of people.
@Tamasg I'm curious, VIM is mentioned in the list of modified profiles but it seems to just quiet the status line. Is using VIM properly actually feasible with the addon as it currently stands? There's a number of issues that may make it tricky, most notably that a lot of navigation doesn't necessarily use the arrow keys, so I'm curious
@Tamasg I'll give it a look later. I'm no vim expert myself but can probably at least find a cheatsheet to figure out the basics.
One thing I do know is that it has the insert and command mode, which I believe is conveyed through the status bar. That bit, you MAY not want to silence as it's a bit of a thing to figure out otherwise :) Maybe that was already taken care of, will check and report back later
One thing I do know is that it has the insert and command mode, which I believe is conveyed through the status bar. That bit, you MAY not want to silence as it's a bit of a thing to figure out otherwise :) Maybe that was already taken care of, will check and report back later
Can you please explain why this seems to duplicate a lot of NVDA's settings in the add-on itself? This is something we very badly need in general, I think, NVDA's terminal support is so shit it can't even use ed properly, but NVDA already has 4 levels of punctuation, indentation reporting, character code announcements, word/character nav, etc. What does this do different?
My thoughts exactly. There are some interesting features (on paper at least - haven't tested this properly yet) but there is certainly a lot of overlap.
@tuukkao @TheQuinbox I considered all of that. When I started thinking about the features I wanted to implement, the one that made me pause was apps/terminal profiles. I knew that I could apply NVDA's profiles to different windows. But, I wasn't sure that profiles could go down to terminal app levels. For that, I needed to duplicate these settings. If it proves to be not necessary, I'll remove the extraneous stuff. A lot will depend on feedback.
@tuukkao @TheQuinbox If the community really tells me that they don't want something, I'll be happy to remove it.
@tuukkao @TheQuinbox I've tried a couple of tricks to avoid some of the NVDA's performance issues. Whether they work long-term needs to be seen. Where I hope this thing should shine is things like bookmarks, windowing, highlight tracking, etc.
I've bumped up the version number for Terminal Access to v1.0.45 with a fix for global keys issue. If you haven't updated it or got to the v1.0.43 release, you can update it here. I've also submited the add-on to the NVDA store. Hopefully, new updates will show up there once approved.
Terminal Access for NVDA 1.0.46 released with better cacheing support and a new configurable output buffer with nvda+alt+n. It automatically speaks newly appended terminal output as it arrives. Coalesces rapid output within a configurable window (default 200 ms). Speaks a concise summary: e.g. "47 new lines". Window monitoring is more efficient and reliable. Settings now give you four new options to control the output buffer.
I'm working on updating Terminal Access to support less complicated set of keyboard commands. Since I'm limiting the commands to terminal windows, things should be less confusing.
This entry was edited (Tuesday, February 24, 2026, 8:27 PM)
Terminal Access for NVDA 1.0.47 released.
• Simplified keyboard shortcuts: All terminal-scoped commands now use simpler 2-modifier combinations instead of 3-modifier combinations. User docs will provide details.
• Dropped Alt modifier entirely (key unchanged) for most commands
• Dropped Alt, key changed (to avoid NVDA global conflicts)
• 3-key → 2-key (dropped Alt from Shift combinations)
• Fixed announcement of "blank" after pressing enter
• Cleanup code
Terminal Access 1.0.51 released with even more simplified shortcuts such as for adding and reading bookmarks, more terminals supported including additional TUI profiles such as for apps like Claude Cli, code improvements for better performance, attempts at crushing persistent bugs like reading "blank" after pressing enter, and more. Read the documentation.
I'll make updates based on feedback since I believe I'm done with features for now.
reshared this
Paweł Masarczyk and Marek Macko reshared this.
@Bri @ZBennoui I should have copied and pasted instead of typing it out.
github.com/PratikP1/Terminal-A…
Releases · PratikP1/Terminal-Access-for-NVDA
This NVDA add-on enables functionality similar to TDSR in NVDA. - PratikP1/Terminal-Access-for-NVDAGitHub
meep, that link is missing your username :)
Thanks for this. If I may ask, is the repeating of letters expected? When I hit, for example, c, it says "c c space". This happens both in cmd and powershell.
The link doesn't work.
Terminal Access 1.0.51 released with even more simplified shortcuts such as for adding and reading bookmarks, more terminals supported including additional TUI profiles such as for apps like Claude Cli, code improvements for better performance, attempts at crushing persistent bugs like reading "blank" after pressing enter, and more. Read the documentation.
I'll make updates based on feedback since I believe I'm done with features for now.
Terminal Access will now show up in the official #NVDA add-on store. For now, it shows up in the beta channel since 2026.1 hasn't been officially released. Updates will show up there as well.
Terminal Access for NVDA 1.0.53 released with several changes.
Better Braille Support including cursor tracking showing the whole line instead of character at a time, add-on specific scripts actually showing content on the display and not just speaking it, and more.
The duplicate key echo issue reported by a few of you should be fixed.
As usual, please let me know if things aren't working as they should or if you want more features.
The code is very self-explanatory (thank you). One suggestion is to use Python 3.11 assumptions all the way e.g. no need to import typing.Dict and you can use enums for some constants. Let me konw if you would like more detailed reviews and feedback.
@joslee Thank you. I'd love to have a conversation with you. The reason why you see a mix of styles is because I'd originally planned to support older versions of NVDA. Because of that, the code needs to be cleaned up. I wanted to focus on the first release and responding to immediate feedback. I'll work on the clean up in the next few days. I'm also iterating since I'm not as confident in my Python skills as other languages. All feedback is welcome, good, bad, or ugly.
Wait I'm confused. Looking at this readme, are a bunch of these features redundant? The review cursor already can move by line, word and character and output the character at the cursor, with char info I can even output Unicode name. Why do you define separate gestures for that which are likely to clash with everything else? What makes this different from NVDA"s normal terminal support? Better selection and automatic profiling? How doe sit know what app you're running?
@kaveinthran @x0 Not yet. Claude code is on my list. I need to find the optimum profile for it.
Yeah, bookmarks, the cursor reading, better selection sound useful, but I kinda need it to suport my existing gestures for reading, these conflict with my add-ons and I simply don't have enough space left to assign them.
And what's with the new keystrokes for tabs and such? Doesn't wt already provide those? Basically I'm just confused as to how much of the README was written by AI and accidentally absorbed features into it that the add-on doesn't even need to get involved for.
@kaveinthran @x0 @joslee As far as I can tell, NVDA's profiles can be programmatically launched by Window/app names. Unless I'm mistaken, NVDA can't detect apps that are running inside terminal windows. This add-on tries to detect those apps. Whether or not it succeeds remains to be seen.
@kaveinthran @joslee Hmmm. That'd be interesting. If it actually parses out the command you run like the shell does it could maybe, but you have to be aware of pipes not just basic commands, and anything that relies on, say, $EDITOR is likely to break unless it can identify the application from some static output it emits.
@x0 @kaveinthran @joslee It's a work in progress. I'm still working on some of the logic in a local branch. Things like Git probably can't be detected unless I somehow try to use the command history feature.
would this work for TTCom?
@bluenova I'm not familiar with it. But, if you want to give it a try, I'd encourage it. If you want me to add it to the supported terminals list, please file a feature request or give me some details about it.
@munchkinbear so I'm curious, since NVDA already reads output in the terminal window, will the add-on read the terminal output twice?
@cary5871 @munchkinbear No. It will not double speak since the add-on uses NVDA itself to output speech.
@munchkinbear how do I fix it from saying the word space after every character I type?
@cary5871 Now it says blank when I use ssh profile, but it's a great idea, so hopefully that will get smoothed out. I don't have any other issues with it. Here's some debug showing what happens. === NVDA Debug Log Capture ===
Start: 2026-02-23 10:54:04
End: 2026-02-23 10:54:07
[nvda.log]
IO - speech.speech.speak (10:54:04.853) - MainThread (28368):
Speaking ['Debug logging enabled; capture started']
IO - inputCore.InputManager.executeGesture (10:54:05.429) - winInputHook (26196):
Input: kb(desktop):enter
IO - speech.speech.speak (10:54:05.431) - MainThread (28368):
Speaking ['blank']
IO - speech.speech.speak (10:54:05.513) - MainThread (28368):
Speaking ['root@serrebiradio:~# ']
IO - inputCore.InputManager.executeGesture (10:54:06.501) - winInputHook (26196):
Input: kb(desktop):enter
IO - speech.speech.speak (10:54:06.503) - MainThread (28368):
Speaking ['blank']
IO - speech.speech.speak (10:54:06.607) - MainThread (28368):
Speaking ['root@serrebiradio:~# ']
IO - inputCore.InputManager.executeGesture (10:54:07.069) - winInputHook (26196):
Input: kb(desktop):shift+NVDA+d
This entry was edited (Monday, February 23, 2026, 6:54 PM)
@serrebi this thing is really weird, I honestly don't think I'll be using it much, every time I type something like cd c:\users\username\dropbox, after every letter I type, it says space, even though I'm not writing spaces.
@cary5871 Well I will say give it a chance. I've played with writing an NVDA addon for this purpose, and it takes some skills and time to get this sorta thing right. Give it time. I gave up on mine.
@serrebi I'm using windows 10. I was doing this in the traditional cmd terminal. after every single letter I type, it says the word space. I can record an audio version of this for you if you want to hear it in action.
@cary5871 Whenever there's a new line it just says blank before the command output. I'm using windows console host as my default terminal app. Hope this helps. If you need anymore from me don't hesitate to ask.
@serrebi Are you hearing anything but Blank with Win Console Host after typing a command? What is your default profile? Press nvda+alt+f10. If you can go into NVDA's settings/Terminal Access/default profile and then change it to WSL, let me know what happens.
['Active profile: None (using global settings). Default profile: None']
@serrebi Sorry. Not sure how I missed your log snippet. If I'm understanding you correctly, you're using SSH in Win Console Host. What you expect to hear is the prompt and not "blank." Does this happen with every command? Or does it happen only when you press enter?
It happens with every new command after I press enter. Like if I type ls, it will say blank first, then say the ls command output.
@serrebi What if you're not in SSH? I'm trying to determine if the issue affects Conhost or SSH inside it.
@lerven Hi, I built a lot of NVDA’s modern terminal and console support. I’m curious, do you use NVDA yourself and have you switched to it recently? Your add-on implements a lot of things (flat review, formatted text, indentation) that NVDA already supports, though admittedly the NVDA-idiomatic way isn’t the most intuitive to figure out. CC @NVAccess
This entry was edited (Monday, February 23, 2026, 8:29 PM)
@bill I've done it this way because of profiles. I realize that there's cross-over. I hope that the profiles I've built look at what runs inside the terminal and apply settings. I don't think NVDA's profiles work that way. If there's a better way, I'm happy to remove the extra settings. @lerven @NVAccess
@lerven @NVAccess Yeah… the lag (especially in heavy text update scenarios) is a really hard problem to solve. It’s also at least partially a XAML bug github.com/microsoft/terminal/…
WT UI Automation memory leak · Issue #19710 · microsoft/terminal
Windows Terminal version 1.23.13503.0 (but issue was reported to be happening on earlier versions) Windows build number 10.0.26543.0 (but issue was reported to be happening on earlier versions) Oth...carlos-zamora (GitHub)
2026-02-23 19:07:22
2026-02-23 19:07:22
2026-02-23 17:59:55
9802786
Peter Vágner reshared this.
#movuary 20 Party on the moon
The media in this post is not displayed to visitors. To view it, please go to the original post.
#movuary 20 Party on the moon
Peter Vágner reshared this.
2026-02-23 18:43:23
2026-02-23 18:43:23
2026-02-23 18:05:24
9802881
Peter Vágner reshared this.
#movuary 21 The Frantic Robot
The media in this post is not displayed to visitors. To view it, please go to the original post.
#movuary 21 The Frantic Robot
This entry was edited (Monday, February 23, 2026, 6:08 PM)
Peter Vágner reshared this.
2026-02-27 09:11:35
2026-02-23 18:41:40
2026-02-23 18:13:02
9802916
Peter Vágner reshared this.
#movuary 22 Surfing on Neptune
The media in this post is not displayed to visitors. To view it, please go to the original post.
#movuary 22 Surfing on Neptune
reshared this
Peter Vágner, Andre Louis, Borris and Jonathan reshared this.
2026-02-27 08:23:23
2026-02-23 18:33:53
2026-02-23 18:18:31
9802942
Peter Vágner reshared this.
#movuary 23 A Run at Midnight
The media in this post is not displayed to visitors. To view it, please go to the original post.
#movuary 23 A Run at Midnight
reshared this
Peter Vágner, Andre Louis and Borris reshared this.
2026-02-22 19:56:46
2026-02-22 17:41:57
2026-02-22 17:02:16
9795001
Peter Vágner reshared this.
in reply to Jiří Eischmann
Today, a friend asked me how to record a podcast (from a closed platform), so I found a simple app called Reco that does exactly what he needs.
I really like the ecosystem of simple, easy-to-use apps that is growing around #GNOME and #Flathub.
In case you ever need it: 👇
Peter Vágner reshared this.
verify that you have removed the language segment from the path x3
2026-02-22 19:40:04
2026-02-22 17:25:55
2026-02-22 11:17:28
9793686
Peter Vágner reshared this.
in reply to Marco Zehe
@Tutanota I just recently tried the Tuta app on the web, Mac, and iPhone again, and am very pleased with the accessibility of the app nowadays. When I first tried it, several years back, it was quite not usable for me as a blind user. I'm happy to say that this has changed dramatically for the better. There are still some issues of course, but this is nowhere near as unusable as it was when I last tested it. Cudos to the team!
@Tuta
like this
Peter Vágner and FediVerseExplorer like this.
Peter Vágner reshared this.
Thanks for your feedback, happy to hear that!
2026-02-22 17:23:19
2026-02-22 17:23:08
2026-02-22 17:23:08
9795349
Dear !Friendica Support
I have just patched my friendica instance to exclude replies when returning home timeline to mastodon API clients and after looking at the time line in #tuba and @Tusky I am wondering why the default is to include replies?
Are there some references to older discussions on this topic?
It's just simple value change and mastodon, pleroma, gotosocial are excluding replies. Even status net / GNU social was doing it at the time.
Friendica Support reshared this.
2026-02-25 05:59:22
2026-02-22 17:04:32
2026-02-22 16:01:48
9794635
Peter Vágner reshared this.
in reply to Jamie Teh
in reply to Jonathan
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Scott
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Kara Goldfinch
in reply to Pratik Patel
in reply to Jamie Teh
in reply to Pratik Patel
in reply to Jamie Teh
in reply to Andy
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Day Garwood
in reply to Day Garwood
in reply to Day Garwood
in reply to Day Garwood
in reply to Ray Dog
in reply to Day Garwood
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Ray Dog
in reply to Jamie Teh
in reply to Ray Dog
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Andre Louis
in reply to Andy
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Jage
in reply to Jamie Teh
in reply to Jage
in reply to Jage
in reply to Jamie Teh
in reply to Jage
in reply to Jamie Teh
in reply to Jage
in reply to Pratik Patel
in reply to Jage
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Scott
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Brandon Tyson
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Brandon Tyson
in reply to Jamie Teh
in reply to Jage
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Brandon Tyson
in reply to Jamie Teh
in reply to Jamie Teh
in reply to x0
in reply to Jamie Teh
in reply to x0
in reply to Jamie Teh
in reply to x0
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Zach Bennoui
in reply to Jamie Teh
in reply to Zach Bennoui
in reply to Martin
in reply to Jamie Teh
in reply to Jamie Teh
in reply to Jamie Teh
I've just spent far too long writing the beginnings of a ridiculous music game for web browsers. You use either a MIDI keyboard or the computer keyboard (a s d f g h j k). The game plays a bar of random notes and you have to play it back, correct notes and correct timing. As you level up, the patterns get more rhythmically complicated. If you make 5 mistakes, game over; you have to start again. Still probably a lot more to do to make this more interesting, but I'm curious as to whether anyone finds this remotely worthwhile.
One challenge right now is how to calibrate if you don't have absolute pitch. I feel like I might need to add some sort of screen where it plays you a few notes until you find the right ones or something. For now, note that it's in A minor, A3 to A4.
files.jantrid.net/simus/
One challenge right now is how to calibrate if you don't have absolute pitch. I feel like I might need to add some sort of screen where it plays you a few notes until you find the right ones or something. For now, note that it's in A minor, A3 to A4.
files.jantrid.net/simus/
reshared this
johann, Peter Vágner, Andre Louis, Borris, victor tsaran, modulux and Zach Bennoui reshared this.
Sounds quite fun honestly. But yeah a way to test how which key sounds before would be nice.
@jonathan859 Note that you can press the keys in the first two bars before it starts playing notes to get a sense of what notes are where.
That sounds (oops, sorry about the pun) like something I'd enjoy playing! Maybe a "beginner" mode could spot the beginning note. Or the other way around, an expert mode that doesn't.
@technocounselor Oh, I'm down for this! I have perfect pitch and I would love it!
Got a bug for ya. If a note is still being held when the game over sound starts, often times here it'll stay playing that note forever and ever. Happens more with the computer keyboard I think, but I can't be certain of that yet.
The melody could be nudged a few dB louder by default. Sometimes hard to pick out the subtleties at the moment in a way that's more annoying than difficult.
I feel like that would keep some people busy for ages! LOL
Can you believe the A minor thing is killing me? I know which notes they are, I can probably hum then perfectly. But I keep thinking in C major for some dumb reason.
Goog idea, I love it. Currently, when I fail, last note plays endlessly.
I love this even though I keep failing at it miserably.
@KaraLG84 It's funny how terrible at this I am. I don't see this improving. I suppose it's just like typing Yes. I'm one of those people without perfect pitch.
Yes. Exactly. That's how I've learned to tune my guitars and how I used to learn piano music. @KaraLG84
@ppatel @KaraLG84 This is a wildly different concept to what you're doing but it popped into my head just now. With some modifications, what you have here could be a fantastic tool to learn pieces of music on the keyboard by ear. Take a thing of sheet music, split it into bars, allow user to pick one or two hands, keep playing the bar until the user plays it correctly, then move on. I don't think anything like this exists currently
How cool is this! You vibe code any of it?
@remixman OK now what would be kind of cool is if you used some kind of old-style machine learning a la the pi composer or whatever it was to output melodic fragments that are a bit less randomly generated and a bit more musically correlated? Trained on, MIDI melodies probably. Loads of free MIDI out there. And the higher level you get perhaps the influence factor goes down and heads towards randomness for difficulty levels.
@Andy
This might sound dumb but... how do you play with NVDA? If I'm in browse mode I'm obviously getting quick-letter navigation issues, if I try focus mode I get nothing - is NVDA still eating my keystrokes?
When I say I get nothing, I mean keys aren't playing notes and I'm still told to try again after five turns.
Love the idea though. Definitely a game I would play if I could work out what I'm doing wrong. It's not just this one I struggle with either - just can't seem to get NVDA to play properly with web games for some bizarre reason.
@daygar Yeah, same here. Looks like it would be fun if it would get my key presses.
@daygar I think you have to get keyboard focus to the right part of the window. What worked for me was as follows:
In browse mode, I saw a line with the word application. I pressed enter on that, then wwent into focus mode, and the game would then see my key presses.
@musicalman Yup, that worked. Thanks.
@musicalman @daygar That's weird. The game auto focuses that node when it starts. I don't understand why that would be overridden or ignored on your systems.
@musicalman @daygar What browser? And what's the value of that NVDA setting for automatic focus mode for focus changes? Using firefox here with that disabled and it still engages by itself, applications have always been the exception to that checkbox.
@technocounselor I'd summarily lose that game, though I need that game. My ear training sucks.
Oh boy I love this. Level 3 is a killer for me though, I'm great at melodies but am weaker with dense/syncopated rhythms. For me level 1 is easy, level 2 is moderate, level 3 is very hard (or impossible right now). Maybe I'll get better with practice.
My only suggestions are that the levels feel a tiny bit too long? Might just be me getting impatient though. Also, I think there should be a way to earn lives/mistakes. I need all I can get for level 3 hahaha
@musicalman Yah, I was wondering about the lengths of the levels. Part of the problem is that I'm running out of ideas to make more levels, haha. I probably need to add more loops so we can do faster tempos and stuff. Agree on adding lives. Any suggestions as to how to earn them? Maybe every 10 correct responses gives you 1 life back or something?
Yeah that was the first thing I thought of with getting lives back, and I think it would work well. I also like the idea of adding different loops at different tempos.
Big thing I'm worried about is just how to curve the difficulty progression of the melodies. I actually experimented with a similar idea a few months back. I tried adding more notes closer together to increase difficulty, but I didn't like it. I don't know if i'll be coming back to that project any time soon, but if I did, I'd likely make a probability system where certain steps of the sequence are much more likely to have a note dropped on them. As just one example, the first note of each beat has a higher chance than the rest. Or something like that, I'm still nowhere near done mulling it over. Anyway that could be a powerful dimension for difficulty adjustment.
@musicalman Yeah, I tried a pattern with all eighth notes and I couldn't do it. The pattern just wasn't melodic enough for my brain to process it. Wondering whether I need to manually enter patterns at various levels of difficulty and have it randomly choose them, but I'd need hundreds of them.
Yeah that would suck if that's the only way you could do it. In any case, an audio game that allows me to use my midi keyboard is certainly a new one. I know of other games that can make use of midi, but none that integrate it into gameplay. I hope you can continue to work on it!
I like this a lot. Great things.
@FreakyFwoof @remixman Haha, that's just a quick loop I threw together with REAPER, Komplete Kontrol and Shreddage. I need to make a few more at some point to keep it interesting..
@FreakyFwoof Hey I'm game. Sounds like fun.
That's cute. I noticed I lost a round and a note kept playing for some reason, had to Ctrl+F5 to stop it. Are there levels that you have to play more than one note at a time?
I'm here for interestingly nasty ideas. You could always implement a cord CAPTCHA.
bestmidi.com/captcha/cords.htm
bestmidi.com/captcha/cords.htm
Seriously, it's a cool idea though and the music gives it some life. Can you control the diffficulty, i.e. make level 1 on expert start faster
Ah didn't know if you wanted to just apply some tempo compression, but that probably makes the loops sound off.
Ya even if it's just 2, that being said, I guess I should beat the game as it is before asking for expert mode lol.
Just pushed a new version of Simus. Fixed the bug where the note gets stuck at game over. Loop music is a bit quieter so you can hear the melody more easily. You now get a life back (up to a maximum of 5) every 10 correct responses.
What's Simus?
Yes, thanks a lot. I found your earlier post. Should have read my Mastodon TL in reverse!:)
New in Simus: a second loop (hear it on level 2), more levels and less points required to level up. Also, you can press escape to "game over" early so you can see your score before you quit. files.jantrid.net/simus/
Please don't add a score board until I've figured out extracting absolute pitch from an unworthy host and redeploying it into my big stupid noggin.
@Scott Heh, I still don't think you should need absolute pitch to do this. As noted in the instructions, use the first few bars of the song to calibrate your ear to relative pitch by playing keys while you wait. That said, it does require very snappy relative pitch that way.
Oh sure, I can do it. That's not the issue. The issue is I'd like to do it well. :P
man this thing is fun. I think I need a midi keyboard though. Using qwerty feels very strange for some reason. Maybe I just have to get used to it.
This could very well just be me, but I feel like sometimes I'm playing the notes correctly and on the correct rhythm and it still occasionally seems to be counting it as wrong sometimes. Again this could very well just be me not hearing something right, but this has happened a number of times now where I'm wondering if there's something going on under certain conditions? I am playing using my computer keyboard. This is a great game and I am really having a great time with it! Thanks for putting this together.
@BTyson Yeah, I think I've seen that once or twice, particularly if I game over and restart. But I'm not entirely sure if it was just me either, haha. I was seeing bugs like that before I released it, but I thought I'd squashed all of those bugs. Can't rule out one lurking somewhere though. Getting the timing right for this in the code is tricky.
@BTyson Played it quite a bit just now, getting to level 3 both times with score 54 or so, and didn't see any problems. 🤷♂️
@BTyson Level 4 sorry. Oh also, note that if you play notes while it's playing you the melody, those will currently count as errors.
I wonder if I happened to start playing as the last note was finishing up instead of fully waiting for it? I'm not sure. I felt like I saw this even on level 1. It was also late at night so maybe I was tired too.
@BTyson That should be fine. I only meant if you pressed other incorrect notes long before you were supposed to start playing. There's a bit of timing tolerance, so starting a tiny bit before should be fine.
@BTyson How many levels are there? So far I have not got past 3 so I guess I can't ask for faster modes lol. Are you tracking people's scores? In addition to tracking correct notes, 1 point per loop it seems like currently, perhaps some score bonus, or more points based on timing, or points for streaks, etc. It's a cool timewaster for sure.
@BTyson @Jage I think I did 6 levels or something? Can't quite remember right now. I've only reached level 4 myself. Yeah, I guess the score could be multiplied by the level or something, though ultimately it just results in a bigger number. :) I've more been wondering whether something should gain you extra lives, since you really need them later in the game. I'm not tracking any stats at this stage. That would require server side code, which I can do, but there's always a chance of abuse: some troll overfilling the database with data, etc. I don't have the cycles to do all of the required security work to protect against that.
@BTyson Ya fair enough, no reason to overengineer things. Oh boy are 5/6 16th notes? Maybe it does an end of round bonus based on some things.
I also have occasionally hit a thing where sometimes, it sounds like one of the notes will get stuck while it's playing the sequence to me, and that lead me to think that the note was supposed to be played twice, but it counted it as an error. So I'm wondering if it's very occasionally missing a note when it should be playing another. This also seemed to happen on level 1.
I think I just saw the bug happen again. When it happened, I heard both the sound of getting it correct, plus the wrong sound at the same time. Then, after that, every other sequence I tried to play back correctly was counted as wrong. This seems to happen for me on level 1, at completely random times.
@BTyson Urgh. I'm sorry. I can't reproduce it, which is going to make it rather hard to debug. It sounds like there's a timing issue where somehow two melodies get scheduled at once - only one plays but both get checked - but I don't know why that would happen.
@BTyson I'll keep an eye out for it. Maybe some inspiration will strike at some point and I'll be like, oh, oops, there's some questionable code here :)
OMG it's like how I played Simon Trickster! Was there ever other hardware that did this? I loved doing this kind of thing, when I was in Jazz Band at my high school for a year me and one of the percussionists would do this. He would play phrases at me on the vibes and I'd play them back on the piano, and vice versa.
Getting used to that input mode on a keyboard is certainly a challenge though! Might connect my M32 just to do that.
Feature request; Can we get a pause button? ANd is there a quit button? Perhaps p and q, or x, or escape or something. If I pause and resume mid-playback, it should reset to the top of the bar. To avoid cheating you could put a grace period of a sort on pausing so you can't use it every turn or something like that.
I mention that so that score can be printed for sharing. You were on level x with a score of x, etc.
Wow, I'm really awful at this game, but I kind of like it anyway. Thanks!
Hell yeah I'm gonna win this game lets go!
@ZBennoui I'll flag that there's no winning at this point, plus there are only 3 levels, with the third going on forever. Not where I intend to leave it, just need to come up with more ideas and more time, heh.
Lol yeah was kinda joking. Really cool game though, and if you ever need more music to use as loops for different levels I'm happy to help out.
@ZBennoui But but but, I can't get it working. I'm on PC. Is there something I should try? I even tried it without a screen reader running.
@ZBennoui I wwas on Crhome. Byt yeah, I will try a different browser with no extensions to see if that works.
I love this! It’s a great riff and there’s some simple joy obtained copying the patterns along that riff. I can’t wait to see how you add to it, if you decide to do that.
2026-02-22 17:01:23
2026-02-22 17:01:23
2026-02-22 16:04:24
9794650
Peter Vágner reshared this.
Playing with my Move and I'm missing something obvious. I've watched Andre's intro video but there's a bit of a gap between what he's doing and what I am.
I'm doing a kind of boring looping thing just to get familiar. I have a simple bass loop I want to play throughout the entire track. There's also a drony synth riff I want to loop, but it starts 4 bars in. And I want to lay all these down in maybe something like a 32-bar arrangement so the bass, synth lead, and drum loop are all playing by bar 12 or so when a vocal lead is added--basically a gradual stack, and then the drums/synth/bass are unstacked at the end so the song starts/ends with the solo bass, if that makes sense.
How do I do this? My initial thought was note mode and just lay everything out on one timeline, but it isn't clear to me how to start track 2 bass looping at 1, track 4 synth looping at bar 5, etc. I tried to start playing my synth loop at bar 5 by using the arrows to start at 5 but that just seemed to loop *everything, as if what I did was start at bar 1 even though the thing clearly said "bar 5" and that's where I hit record.
Maybe what I should be doing is structuring these as clips, and I've sort of done that. I now have my bass and synth parts mapped to the second and fourth leftmost pads, but it's not clear to me how to actually structure those into a song so the first clip plays for 4 bars and keeps looping, then the synth clip starts on cue. I thought maybe I'm supposed to record in session mode and trigger clips, but that doesn't seem possible.
Reading the manual, it's a great overview of each slice of functionality but nothing seems to tie everything together. What am I missing? Andre builds everything as a single loop, and that's a great intro but I'm aiming for something that feels just a bit more complex and for me there's an understanding gap.
reshared this
Jonathan and Peter Vágner reshared this.
2026-02-22 16:59:19
2026-02-22 16:59:19
2026-02-22 02:51:24
9794752
Peter Vágner reshared this.
Huge thanks to @maximiliano and @TheEvilSkeleton for reviewing & merging this very long-awaited UX improvement in GNOME Calendar's infinitely scrolling month view: the previous/next buttons (and corresponding keyboard shortcuts) now properly clamp to the beginning of months when switching months! 
See the "before" vs "after" demonstration videos in the merge request: gitlab.gnome.org/GNOME/gnome-c…
#GNOMECalendar #GNOME #UX #productivity #calendaring #planning #OpenSource #FLOSS
Jump to the first day when going to the next/previous month (!369) · Merge requests · GNOME / gnome-calendar · GitLab
See commit messages for details. FixesGitLab
reshared this
Peter Vágner reshared this.
2026-02-22 15:25:38
2026-02-22 15:25:38
2026-02-07 12:20:21
9674427
Peter Vágner reshared this.
Let's try again with the LWT 1996 ident because I fucked it up last time. #NVDAComposer
NVDA_COMPOSER_CLIP v2
# tempo: 120
n67 t120 v127
n66 t120 v127
n64 t120 v127
n62 t120 v127
n67 t120 v127
n66 t120 v127
n64 t120 v127
n62 t120 v127
n71 t120 v127
n69 t120 v127
n67 t120 v127
n66 t120 v127
n76 t120 v127
n74 t120 v127
n72 t120 v127
n71 t120 v127
n72 t120 v127
n71 t120 v127
n69 t120 v127
n67 t120 v127
n72 t120 v127
n71 t120 v127
n69 t120 v127
n67 t120 v127
n74 t1440 v127
n86 t480 v127
n83 t480 v127
n81 t240 v127
n83 t960 v127
n55 t3840 v127
The media in this post is not displayed to visitors. To view it, please go to the original post.
Let's try again with the LWT 1996 ident because I fucked it up last time. #NVDAComposer
NVDA_COMPOSER_CLIP v2
# tempo: 120
n67 t120 v127
n66 t120 v127
n64 t120 v127
n62 t120 v127
n67 t120 v127
n66 t120 v127
n64 t120 v127
n62 t120 v127
n71 t120 v127
n69 t120 v127
n67 t120 v127
n66 t120 v127
n76 t120 v127
n74 t120 v127
n72 t120 v127
n71 t120 v127
n72 t120 v127
n71 t120 v127
n69 t120 v127
n67 t120 v127
n72 t120 v127
n71 t120 v127
n69 t120 v127
n67 t120 v127
n74 t1440 v127
n86 t480 v127
n83 t480 v127
n81 t240 v127
n83 t960 v127
n55 t3840 v127
reshared this
Peter Vágner reshared this.
2026-02-22 15:24:18
2026-02-22 15:24:18
2026-02-07 13:03:36
9674446
Peter Vágner reshared this.
Inspired by the chord sounds that seem like I automated them but I didn't. Just playing harder or softer opens up the filters. Quite fun really. A short, simple idea using sounds from Roland Cloud Zenology.
The media in this post is not displayed to visitors. To view it, please go to the original post.
Inspired by the chord sounds that seem like I automated them but I didn't. Just playing harder or softer opens up the filters. Quite fun really. A short, simple idea using sounds from Roland Cloud Zenology.
Peter Vágner reshared this.
2026-02-22 15:20:44
2026-02-22 15:20:44
2026-02-08 07:54:49
9684027
Peter Vágner reshared this.
#movuary 08 The Cheap Producer. My #mac has been kind of a pain in the ass when it comes to using the screen reader, but my phone pulled through for me. Enjoy.
The media in this post is not displayed to visitors. To view it, please go to the original post.
#movuary 08 The Cheap Producer. My #mac has been kind of a pain in the ass when it comes to using the screen reader, but my phone pulled through for me. Enjoy.
reshared this
Peter Vágner reshared this.
2026-02-22 16:04:30
2026-02-22 13:40:02
2026-02-08 17:35:59
9685182
Peter Vágner reshared this.
Apparently, there was an obscure Windows port of the Macintalk speech system called WinTalker. I was able to successfully build it and run it.
in reply to Anthony C. Bartman
in reply to aaron
in reply to Alex Chapman
in reply to aaron
in reply to Anthony C. Bartman
in reply to Anthony C. Bartman
in reply to Peter Vágner
The media in this post is not displayed to visitors. To view it, please go to the original post.
Apparently, there was an obscure Windows port of the Macintalk speech system called WinTalker. I was able to successfully build it and run it.
Peter Vágner reshared this.
@alexchapman Cool. I wonder if this could possibly be an NVDA addon, ha.
@fireborn Oh boy that'd be cool, or even ported to SAPI5, looly.
@alexchapman @fireborn Yeah. As long as Apple doesn't catch wind of it. They still have those voices on their new macOS systems.
This entry was edited (Sunday, February 8, 2026, 6:01 PM)
@fireborn @alexchapman Yeah probably. Only time will tell when that'll be. The code I have is pretty barebones at the moment.
Oh wow, that's cool!
@ROMMIX There is also a czech speech synthesizer called wintalker voice. There is nothing special just the name is somewhat familiar to me.
@pvagner yeah. Pretty sure it was a placeholder name when Mark Cecys used that name.
Bri🥰
in reply to Brandon • • •Brandon
in reply to Bri🥰 • • •Brandon
in reply to Brandon • • •Brandon
in reply to Brandon • • •Martin
in reply to Brandon • • •Brandon
in reply to Martin • • •Noah Tobias Carver 👨🏼🦯🇺🇦
in reply to Brandon • • •Brandon
in reply to Noah Tobias Carver 👨🏼🦯🇺🇦 • • •GitHub - thgcode/orca-remote: An Orca screen reader plugin that makes it communicate with NVDA Remote
GitHubNoel Romey
in reply to Brandon • • •Noah Tobias Carver 👨🏼🦯🇺🇦
in reply to Noel Romey • • •Brandon
in reply to Noah Tobias Carver 👨🏼🦯🇺🇦 • • •Dickson Tan
in reply to Brandon • • •Brandon
in reply to Dickson Tan • • •