everything having its own complete app is so incredibly Western and capitalist so you can see why China's embrace of the superapp concept is so compatible with their culture/society.
that said, I don't know how we can break people free from the "app fatigue" problem we've created. People just have zero motivation to try anything that will require they spend 30 seconds learning a new app layout.
But they'll spend 6 hours in 30 second increments watching Tiktoks
whole thing sucks and people are not learning from their mistakes because it takes years before their app finally becomes so unusable, intrusive, or finally shutdown. They won't change until they're forced to change, so these companies just boil-the-frog over time.
someone should create a bot that harvests all kinds of metadata about the chat members, logs all the conversations, etc for analysis and starts cross referencing it with any other data that can be scraped off the internet.
Invite the bot to the chat under an unassuming name.
Wait a while, and then have it start publishing reports about people exposing their lives and habits and stuff. Then when they are shocked about what they're seeing you can say "this is what Facebook/Meta is learning about you but you choose to keep giving them this information instead of switching to a more private communications app"
pre-Elon takeover I think Twitter could have evolved into that if they really wanted to.
You can't do it if half the world hates you though.
Amazon at one point could have pulled this off. I'm not sure anyone else in the West could do it at this point. We'd need a universally trusted brand and it's not Big MegaCorp, it's not Social Media Behemoth, and it's definitely not Fintech Banking Startup.
@feld I think Google probably had the most chance of success, but they can't design a good UX to save their life. If they could, they've have locked down search (and the web in general), messaging, social media, phones (both the physical devices, voip/SMS, and the OS), shopping, etc. But instead they created weird, confusing (and competing) products that with horrible UIs that they then canceled.
Even Android's UI was pretty bad (imho) for the first decade+.
If you have to choose between America being dominated by Amazon or Walmart who do you choose?
Everyone picks Walmart. Every time. And Walmart is winning the long game anyway. They have a far superior logistics and supply chain, plus brick and mortar presence.
This doesn't dismiss the very righteous hatred for how they've destroyed small businesses in communities.
But nobody's going after Walmart (or Costco) in the same way they attack other corporations for their behavior. It's just not palatable. They like shopping there too much.
@feld @xdydx Curious why you're lumping in Costco. They've been one of the best corporate citizens in resisting Trump, so if we have to have corporate citizens at all, I'll choose them.
Instead of a superapp, we should have open protocols. We don't need half a dozen different apps for email, why do we need that many for chat or social media?
With email we expect a small subset of features. With social media and instant messaging there are so many ways to do things that an open protocol isn't sufficient because you just end up with massive fragmentation. This is why proprietary apps thrive. They can tightly control the user experience and ensure equity across supported platforms.
XMPP tried this and failed. The XEPs and design-by-committee did not work. Inevitably this led to iMessage and WhatsApp being proprietary XMPP implementations.
The same thing is happening on the Fediverse. Sure, ActivityPub is an open protocol, but you can't view the new Pixelfed Stories on Mastodon or Pleroma; many implementations have emoji reactions but Mastodon does not.
How does an open protocol solve this issue on its own?
see also: Mastodon limitations on polls and post length
@feld @mcv I'm not sure that I agree that a) embrace & extend is a feature, and that b) open standards can't define a minimum featureset or result in fragmentation. Xmpp being a good example here - fragmentation is huge not due to committee, but by the very idea of a tiny core with a million extensions.
Every jabber server has optional extensions. Every client has a mix of supported extensions. A boon for proprietary users, who use it as a base for their closed systems (googlechat, zoom...)
I'm not saying it's a feature, only that if you have the capability to control the clients on all platforms and the server you can guarantee people have a good experience and never run into weird compatibility issues. You also get the same app/branding on each platform too, so there's no confusion.
Somehow people can figure out that "Outook", "Mail.app", "Gmail" etc are all "email" and compatible with each other, but once you try to get them to use some other open source service where the clients on each device they own are completely different in design and features they just can't figure it out.
"Why can't Flupple on my Mac make a video call to Dogwag on my friend's Android? It says something about incompatible encryption? Why are we bothering with this, why can't we just use Whatsapp" is essentially what you end up with
I will say though, this is why I think DeltaChat will persevere and has a high chance of wide adoption vs other open source secure messengers.
They've made a cross-platform Rust core that implements ALL the core functionality and you just have to put a UI on top of it. So they did, and they have 99% feature compatibility across Mac, Windows, Linux, Android, iPhone, Ubuntu Touch, cli/terminal.... there's even a feature phone (flip phone!) implementation, but it needs to be refreshed as it was mostly just an experiment...
@feld @mcv I kinda thought we were there with matrix, but then I ran a server for a few years and wow, what a dog. Have heard about but haven't tried Deltachat yet!
Matrix has so many weird issues people don't even realize, like first party clients that have existed for years not implementing all the crypto.
And yes, the server is a dog. It's basically a git repo or a blockchain. The data will grow and grow forever. Deletes don't actually exist; they just get hidden from users with tombstones. And this has a whole nest of other problems from a security perspective: if your Matrix server has open registrations someone can make an account and join a room from another server which imports ALL the data.
Like illegal data. Images and videos you never want to be associated with. But now your server has a permanent copy of them in its database. Forever, or until you nuke the entire server and make a new one.
Good luck trying to explain that one to the Feds doing a no-knock raid on you at 3am.
Matrix IMHO is suitable only for situations where you *need* the data to be permanently stored as an audit trail // immutable, like corporate and government internal communications.
It doesn't make sense for IoT like they tried to advertise. And it is not a good fit for regular people's needs.
Different cloud file sharing services fight for control over files I provided nextcloud services for a few ad-hoc groups that needed to collaborate for a limited time on a project. These groups wer...
Uh, I am sorry to ask, but, is that burrito and quesadilla you meant? I am sorry, it's a genuine question, I just could not understand. If so, damn, I am jealous but also, I can't believe food is sooo expensive! 😱
Dear #Mastodon, please, please, if I ask really nicely could you kindly refrain from erupting into yet another fire storm of controversy at least for one afternoon?
Yep, I agree with this, let's give it a hundred thousand thumbs up and send it into space! Woo! Now, there are healthy arguments, but they are discussions or debates, the unhealthy ones are just unnecessary drama and conflicts that quite frankly, no one needs. I'm boosting this, if it's okay.
So far I’m having a great day. Got matching tattoos with my best friend this morning and we went out to a nice lunch at the 99. Relaxing for a bit before open mic
Alba's hot take for today is: I don't necessarily disagree with the judge who ruled Meta's pirating of materials for training to be fair use. the problem here is that intellectual property, like any form of private property, is a flawed premise from the start
I have to say I'm really alarmed by what seems to be a much IP-friendlier turn. I recognise the unfairness of Schwarz getting hounded to death for doing something others get away with and do for profit, but the solution isn't to make the enclosure of knowledge even stronger.
@modulux For what I've seen, it's less of a matter of people upholding IP laws, and more like people being angry that those IP laws that were supposed to be neutral and apply to everyone, only work if Disney wants to fine you to oblivion for pirating one of its movies.
I think the general opinion that big companies abuse their copyrights has not changed, it's just that many people is pointing out the hypocrisy there is in all that situation.
@nasd1066 @modulux idk, I don't follow law that often but I don't think laws only work for big companies. I don't think we're there (yet). people say that because the rulings are unfair, but they aren't because judges/juries are biased towards companies but because (among other things) laws don't take into account the wealth of involved parties and the fact that there's a difference between doing something for profit and doing it for leisure or even to survive
@modulux Okay, maybe Disney suing someone to oblivion is an exaggeration (in the EU at least), but people do get fined for seeding torrents, sometimes several thousand euros.
Meanwhile, if a minor artist were to sue an AI company for having scrapped their data, they'd need to pay a lawyer, in some places the costs of the trial too, there would be readjournments, etc. So it's very difficult to enforce those laws when you're not a big company.
@nasd1066 @modulux the charges were for seeding, not for downloading. pirating things is legal; enabling others to do it isn't. so scraping, even if you consider it piracy, would be legal.
here the judge was applying the law "neutrally" when he declared that, since Meta had taken precautions to avoid seeding, it wasn't doing illegal piracy (at least AFAIK, not super familiar with US law)
@nasd1066 @modulux and yeah the point of the justice system being hard to access for the not-ultra-wealthy is perfectly valid, but that's unrelated to whether judges are applying the laws equally to everyone or not
I think this is generally true. But I do think it bends at the edges. By this I mean, if the application of law would destroy an economic sector, I think courts would be extremely reluctant and creative about whether and how to apply it.
On our way back from an absolutely fantastic weekend at a cottage. Spent tons of time in the lake and quality time with wonderful friends. Also fantastic food! Two days at work this week then I’m heading to Winnipeg for a long weekend.
I do like def. That said, my criteria is whatever's easiest to write, pronounce, and takes up less space. So fun is probably it for me, closely followed by fn.
I remain a fan of consuming my news via RSS feeds. There are #accessibility and time-saving benefits to getting all your news in one accessible app, and there are no naughty algorithms trying to filter your view of the world. For many years, I've used Lire as my RSS reader on my iPhone, in conjunction with a feed aggregator service called The Old Reader. Sometime last year, The Old Reader hit a rough patch, but it seemed to have come right, until a couple of months ago. At that point, I began experiencing serious issues accessing The Old Reader from Lire. Most of the time, Lire would time out. I worked out that if I tried to access the service repeatedly, it usually succeeded on the third or fourth attempt, but that was time-consuming and frustrating. One other person contacted me to let me know he was experiencing the same problem. I contacted The Old Reader, and they seemed uninterested in pursuing the issue, apparently because they weren't receiving many reports of it. I contacted the developer of Lire, who couldn't have been more helpful. He asked for my credentials for The Old Reader so he could test with the app, and he duplicated my findings right away. He was able to identify the issue with The Old Reader API that was causing the time-out, and gave this information to The Old Reader, who didn't even acknowledge his message, let alone commit to resolving the issue. So over the long weekend, I resolved to find a better service. After doing some deep research with ChatGPT, I settled on Bazqux Reader. The website has a few rough edges from an accessibility point of view, but I was able to export my feeds from The Old Reader, import them into Bazqux, and log in via Lire. The one thing that is very different is the speed. Retrieving article from Bazqux is way faster than The Old Reader ever was, something that my deep research told me I should expect. So, sadly I can't recommend The Old Reader anymore, but so far so good with Bazqux. Lightning fast and simple setup. There is a cost for this service, I think about $39 per year, and Lire does support using iCloud to store your feeds. I find that for the large number of feeds I have, a service like Bazqux adds value, but if your requirements are more modest, using iCloud may be sufficient. However you choose to use it, RSS is worth using.
Oh Jonathan, I should have missed your suffering with RSS if you wrote about it before. I'm so sorry because I have a great recommendation for you. It's called Miniflux (Flux being the French for stream, or in this context, feed, as in RSS feed). It's at miniflux.app, it costs something like 15 euros a year, it's open-source, 100% accessible (I'd say 99% because I remember finding some tinny tiny quirks). And it's awesome, I've been using it for almost a year now. Also, if you have a TweeseCake account (and you do if you use the app), they provide their instance of Miniflux for free, if I'm not mistaken. go try it, it's fully accessible and thrice cheaper, if not self-hosted.
I'm sorry, but you might be wrong. From their official page: "Support for Feedly, Feedbin, Feed Wrangler, FeedHQ, The Old Reader, Inoreader, Newsblur, BazQux Reader, FreshRSS, Miniflux, Nextcloud News, and Tiny Tiny RSS is currently included." So you might want to reconsider.
@menelion Yeah I have it up and running now. It's slower than Bazqux. I have my RSS feeds sorted into folders such as blindness, tech news, sports etc. Do you know if there's a way to make that folder structure show up in the Miniflux web UI?
Those are called categories here. And you might view feeds of only one category. The only weak point is that you cannot have a tree of categories and subcategories.
@menelion Thanks, I couldn't find anything eitehr. Bazqux is refreshing every five to 30 minutes, which on a day like the Apple events is really handy.
Jave upi troed uh what was its name? News, something with news, it has RSS. I know it's pretty accessible, and yes, RSS can be worth it. I still haven't made it a habit to use my reader. I am just using miniflux on the web because I am lazy. Thanks for sharing, though. Appreciated it.
I use icloud for my rss, but that's mainly because I haven't found good rss reader on windows. I know there is Luna RSS, but it doesn't provide any support for any sinking service.
Yo computer touchers: there's a history of Black invention and ingenuity you may or may not know about, *and* a history of Black inventors' ingenuity being denied at a systemic level
Estoy metido en una idea un poco rara. Hacer tutoriales de mates accesibles a personas ciegas. Los principios:
Notación en ASCII lo más puro posible. Explicaciones desde el principio, y desde los principios, magia mínima, todo lo que se pueda probar se prueba. Una guía de notación en cada página donde venga toda la notación que se usa. Accesible. Libre. Gratis. Para siempre.
De momento solo tengo configurada la plataforma para hacerlo realidad, pero tengo curiosidad por saber si a alguien más le puede interesar, como usuario, contribuyente, o ambas cosas.
Hola, cuando te refieres a tutorial, incluyes la enseñanza de las mates? O solo como escribirlas. Yo tengo una guía en docx para escribir mates usando unicode math que se lee bastante bien tanto por nvda como por jaws y creo que no es tan engorroso como lateX o ascii. En mi experiencia ascii puede ser de uso básico, para apuntes y así, pero en la práctica si el docente es cerradito, no va a funcionar mucho a nivel universitario sobre todo. No he probado cambiar la guía de formato, aunque sí pensé en crear un epub a ver si la lectura se mantenía igual de bien que en docx. Es un proyecto que puede ser útil, sí. Saludos.
Me refiero a enseñar la materia en sí, no como se escribe. Aunque inevitablemente para poder enseñar la materia hay que elegir una notación. Pero la utilizo solo como vehículo.
Bueno, eso me parece estupendo, la verdad. La única pega que le veo al ascii a nivel avanzado es con la notación para ciertas funciones. Justo por lo que acabas de decir, escribir los símbolos memorísticamente es un coñazo, lo sé de experiencia porque me tocó en el técnico. Pásame un correo por privado y te la envío sin problema. el unicode math tiene su sintaxis similar a lateX, pero un poquito más simplificada, entonces no te toca memorizar en sí la codificación, sino como se escribe la sintaxis. ej: en unicode math en el editor de ecuaciones de word, una potencia se escribiría así X^2 y al pasar a estilo profesional, los lectores leen algo como: matemáticas, equis elevado al cuadrado o equis al cuadrado, creo que varía entre nvda y jaws. No he probado con voice over. De antemano sí te digo, hasta ahora en PDF no va. Ni con uno ni con otro. Lo otro es que puedes navegar la ecuación y eso es algo que a muchos estudiantes ciegos les importa bastante.
Sí, lo de navegar la ecuación es importante. Yo me arreglo con ascii aunque para ecuaciones complicadas la cosa puede ser difícil. La idea por ejemplo de navegar por términos en MathML es buena, aunque la implementación no me guste nada.
Fíjate, mathML yo lo veo un poco inútil porque los docentes ni por asomo pasan por ahí, entonces el estudiante invierte tiempo y esfuerzo y no le va a servir de mucho. Quizá para crear apuntes que solo sean leídos podría ser, pero para la práctica de estudiar, resolver, presentar evaluaciones nada de na. Dame un chance que miro bien el email y te lo paso.
No, para utilizar así yo MathML tampoco lo veo, pero cada vez lo veo en más sitios (wikipedia, páginas web que compilan fragmentos de latex a MathML...). Pero no me parece un formato razonable para trabajar en él como tal.
Para trabajar sería con ASCII, pero para exponer una fórmula o explicar una demostración, mejor MathML ya que no sabes si la persona que lo lee está con voz, braille o zoom, así cubres todo
No sé, yo o he tenido muy mala suerte o no me convence nada la forma en que, por ejemplo, NVDA trata MathML. Dicho eso, si consigo una forma de sacar salida de MathML que no me rompa mucho la cabeza, no me importaría ponerlo también.
Unser Onlineshop erstrahlt in neuem Glanz – klar, modern und benutzerfreundlich. Mit viel Arbeit und Liebe zum Detail haben wir unseren Onlineshop neu gestaltet. Schaut vorbei und lasst euch inspirieren!🙂
The hidden burden of undiagnosed ADHD is constantly needing to invent excuses in an attempt to explain or justify behavior that you don’t even understand yourself.
Europick.eu is live! 🐘✨ Easily find European alternatives and share them with your friends on Mastodon or other social media. Created by PixelUnion. Check it out! europick.eu #Europick #EuropeanAlternative
Here we go again. Whenever policy makers insist that there’s some “nerd harder” solution to tricky societal problems, actual experts have to spend a ridiculous amount of time explaining basic reali…
So, I later learned that posts don't transfer, which is unfortunate. We had some banger posts on there.
Anyway, I just want to make some observations upon setting up our homepage:
Federated timeline is WAY more busy, wow. I was hoping it was, but this is a lot.
Images break a lot faster. Only a few minutes in and I'm seeing a cascade of gradient images, which... I don't know how to fix that? I mean, they eventually load, but I imagine it's because the federated timeline is filling up that fast.After further mucking about, it seems like some images just... stop loading and stay stuck partially loaded, which, yeah, that's unfortunate.
The usuals we're used to seeing in the federated timeline are now gone.
I'm sure we'll get used to it, but this is definitely a change, but hopefully a good one.
Hi and welcome! As for the images, that is very strange and probably shouldn't happen. I'll have a look at the logs to see if I can find anything. Most of us here can't see them, certainly none of the mod/admin team can, so it's entirely possible something went wrong somewhere and nobody ever noticed. That also means they've probably been broken for years! Oh no! Sorry!
I mean, that makes sense because the alt text loads just fine, so it wouldn't be a problem for most people on this instance. Yeah, for us, some images on our home dock take either several minutes to load, and then others will partially load where a top chunk of the image has loaded, but the rest is just blank. How large this loaded chunk can vary between a sliver and about 1/4 of the actual image, at least from what I've seen.
What's odd is that some newer images load just fine while older ones below are either broken or are still loading. Not sure what the pattern there is.
Wait... half loading images? How... how does that even make sense? Oh gods how do I even test that... OK this will be interesting... do they stay broken if you refresh?
@talon Yeah, it's really weird. For example, on the federated timeline, this person's post contains one of those half-loaded images: wandering.shop/@silvermoon82/1…
But if I go on their page, the image loads just fine, but even that's not 100% consistent.
I tried clicking a specific account with these half-loaded images and it seems like they're still not loaded, but it's not consistent. One image that was partially loaded then became blurred when I refreshed the page while another half-loaded image stayed half-loaded. It's really weird.
They don't tell anyone this but Australians/Kiwi's come with a built in 500ms typing/action buffer in their brain that they use for typing prediction for terminals on systems far away from them (most of the world)
I didnt need to be so harshely called out, thankyou very much.
On this note, I have a friend who constantly bitches about having such high latency for certain things. A terrifying collossal latency of... 70ms. I give her a lot of shit for this. A *lot* of shit.
@MissAemilia I find myself in Melbourne at the moment, and a very harsh reminder of what the typing latency to the UK is like here. Though the websites are not as bad as I thought it would be!
Да, така и предполагах. Добре, че не са ни давали да я четем, всъщност ако са, най-вероятно съм пропуснала, ама надали. Сега ако питам кога точно се развива действието на книгата, се надявам да не звуча изключително тъпа. Аз наистина нищо не знам за писателя, и за книгата или пък не съм се постарала да запомня каквото и да е било, след изречението "Книгата е тежка".
Възможно е, всички приемаме нещата по различен начин. Но за да я разбера, трябва да прочета. А аз не съм голям фен на тежките книги. Аз си обичам фантастиката и LITRPG книгите. Рядко чета нещо тежко. На никого след осем часа пред компютър не му се четат супер тежките неща. Като цяло на никого след работа не му се четат супер тегавите работи. :) Ще се пробвам да я прочета, обаче. Може би в края на този месец, ако всичко с моите планове е наред. :)
Update4: the part hasn't shipped. Not ETA. Yet I got asked to evaluate the support handling. I hate to break it for you @frameworkcomputer, but until I get the part, the support ticket is still open. That mean, I hope you read the feedback comment. Someone why can take action. Each and every day with an ETA just make things worse.
> Be GNOME foundation > Huge potential, clout, possibilities to take Linux to the next level with proper planning and community organising > kick out board member for trying to actually improve things and ruin his reputation on the way > huge shitshow > executive director fails to secure enough funding to even pay her own salary and leaves > eventually find someone new to be ED > also kick him out after 4 months for trying to actually change things for the better
i think something needs to be done about the 20+ year GNOME veterans on the board who aren't willing to adopt change
i'm not a foundation member and don't plan to be, but i do care about the future of GNOME and despite the best efforts of many wonderful people it's not looking good rn
The simple solution to crashing birth rates is to tax billionaires and make basic living affordable. It's not that difficult, but the rich are more worried about their own bank balance than they are about birth rates. As usual they want to blame everyone but themselves for the problems they create.
I am sharing this link but only because it makes me so goddamn angry:
https://www.telegraph.co.uk/business/2025/08/28/birth-rates-plummeting-space-might-be-best-hope-cure/
No, the fuck it is not.
Playing Hootananny's Brixton one night, I had my keyboard on my back and that stage is extremely high off the ground. A couple more steps forward and I'd have fallen off and probably broken some part of me. Next thing I know the percussionist grabs me from behind and says, 'Mate, you blind or summink? Few more steps forward we wouldn't have a keys player at all. Watch where you're going!' Perfect delivery, everyone bust out laughing. That's the way a response should be done.
I recently talked with a guy who builds software for small & medium sized German insurance firms and he says almost all of their clients are working on removing US services from their tech stack due to perceiving them as highly risky. This includes cloud services, office tools, AI tools, etc. So far he said the switch is a mixed bag with some successes and some blunders, but the push is very clear.
Highly anecdotal stuff, but maybe this cut is deeper than I expected.
are there any good, trustworthy resources for learning how to perform safe and effective hypnosis? i'd like to get into it, but i currently feel like i'm kinda stumbling around in the dark
There's a handout folder from the local kinky hypnosis group. (Link follows in DM).
Mostly in German, but three of them are available in English. Potentially also more could be translated into English, but idk how helpful these are without the workshop part actually.
feld
in reply to Andres • • •everything having its own complete app is so incredibly Western and capitalist so you can see why China's embrace of the superapp concept is so compatible with their culture/society.
that said, I don't know how we can break people free from the "app fatigue" problem we've created. People just have zero motivation to try anything that will require they spend 30 seconds learning a new app layout.
But they'll spend 6 hours in 30 second increments watching Tiktoks
whole thing sucks and people are not learning from their mistakes because it takes years before their app finally becomes so unusable, intrusive, or finally shutdown. They won't change until they're forced to change, so these companies just boil-the-frog over time.
i hate it
feld
in reply to feld • • •someone should create a bot that harvests all kinds of metadata about the chat members, logs all the conversations, etc for analysis and starts cross referencing it with any other data that can be scraped off the internet.
Invite the bot to the chat under an unassuming name.
Wait a while, and then have it start publishing reports about people exposing their lives and habits and stuff. Then when they are shocked about what they're seeing you can say "this is what Facebook/Meta is learning about you but you choose to keep giving them this information instead of switching to a more private communications app"
Perhaps THAT would wake them up.
(probably not though)
Andres
in reply to feld • • •feld
in reply to Andres • • •pre-Elon takeover I think Twitter could have evolved into that if they really wanted to.
You can't do it if half the world hates you though.
Amazon at one point could have pulled this off. I'm not sure anyone else in the West could do it at this point. We'd need a universally trusted brand and it's not Big MegaCorp, it's not Social Media Behemoth, and it's definitely not Fintech Banking Startup.
Who is left that is not universally hated?
Walmart and Costco?
Andres
in reply to feld • • •@feld I think Google probably had the most chance of success, but they can't design a good UX to save their life. If they could, they've have locked down search (and the web in general), messaging, social media, phones (both the physical devices, voip/SMS, and the OS), shopping, etc. But instead they created weird, confusing (and competing) products that with horrible UIs that they then canceled.
Even Android's UI was pretty bad (imho) for the first decade+.
feld
in reply to Andres • • •you're right, but I discount Google because they stopped adhering to the "don't be evil" mantra so long ago...
Around 2012 they could have done this. It should have been a key part of their Android launch actually...
Death by Lambda
in reply to feld • • •Uh.. I have news for you about #Wal-Mart
feld
in reply to Death by Lambda • • •nobody's boycotting Walmart.
If you have to choose between America being dominated by Amazon or Walmart who do you choose?
Everyone picks Walmart. Every time. And Walmart is winning the long game anyway. They have a far superior logistics and supply chain, plus brick and mortar presence.
This doesn't dismiss the very righteous hatred for how they've destroyed small businesses in communities.
But nobody's going after Walmart (or Costco) in the same way they attack other corporations for their behavior. It's just not palatable. They like shopping there too much.
Death by Lambda
in reply to feld • • •Everyone picking Wal-Mart over Amazon every time does not preclude Wal-Mart being universally hated.
feld
in reply to Death by Lambda • • •the polls don't agree. One from March 2025 put Walmart at the top and it was universally selected across all age groups
Men - 61.6%
Women - 64.6%
Gen Z - 63.2%
Millennials - 62.8%
Gen X - 64%
Baby Boomers - 63.3%
commercial.yougov.com/rs/464-V…
edit: if Walmart could run for President, it would win in a landslide. Simple as that.
M.S. Bellows, Jr.
in reply to feld • • •feld
in reply to M.S. Bellows, Jr. • • •Martijn Vos
in reply to feld • • •@feld @Andres
Instead of a superapp, we should have open protocols. We don't need half a dozen different apps for email, why do we need that many for chat or social media?
feld
in reply to Martijn Vos • • •With email we expect a small subset of features. With social media and instant messaging there are so many ways to do things that an open protocol isn't sufficient because you just end up with massive fragmentation. This is why proprietary apps thrive. They can tightly control the user experience and ensure equity across supported platforms.
XMPP tried this and failed. The XEPs and design-by-committee did not work. Inevitably this led to iMessage and WhatsApp being proprietary XMPP implementations.
The same thing is happening on the Fediverse. Sure, ActivityPub is an open protocol, but you can't view the new Pixelfed Stories on Mastodon or Pleroma; many implementations have emoji reactions but Mastodon does not.
How does an open protocol solve this issue on its own?
see also: Mastodon limitations on polls and post length
Andres
in reply to feld • • •@feld @mcv I'm not sure that I agree that a) embrace & extend is a feature, and that b) open standards can't define a minimum featureset or result in fragmentation. Xmpp being a good example here - fragmentation is huge not due to committee, but by the very idea of a tiny core with a million extensions.
Every jabber server has optional extensions. Every client has a mix of supported extensions. A boon for proprietary users, who use it as a base for their closed systems (googlechat, zoom...)
feld
in reply to Andres • • •I'm not saying it's a feature, only that if you have the capability to control the clients on all platforms and the server you can guarantee people have a good experience and never run into weird compatibility issues. You also get the same app/branding on each platform too, so there's no confusion.
Somehow people can figure out that "Outook", "Mail.app", "Gmail" etc are all "email" and compatible with each other, but once you try to get them to use some other open source service where the clients on each device they own are completely different in design and features they just can't figure it out.
"Why can't Flupple on my Mac make a video call to Dogwag on my friend's Android? It says something about incompatible encryption? Why are we bothering with this, why can't we just use Whatsapp" is essentially what you end up with
feld
in reply to feld • • •I will say though, this is why I think DeltaChat will persevere and has a high chance of wide adoption vs other open source secure messengers.
They've made a cross-platform Rust core that implements ALL the core functionality and you just have to put a UI on top of it. So they did, and they have 99% feature compatibility across Mac, Windows, Linux, Android, iPhone, Ubuntu Touch, cli/terminal.... there's even a feature phone (flip phone!) implementation, but it needs to be refreshed as it was mostly just an experiment...
Andres
in reply to feld • • •feld
in reply to Andres • • •Matrix has so many weird issues people don't even realize, like first party clients that have existed for years not implementing all the crypto.
And yes, the server is a dog. It's basically a git repo or a blockchain. The data will grow and grow forever. Deletes don't actually exist; they just get hidden from users with tombstones. And this has a whole nest of other problems from a security perspective: if your Matrix server has open registrations someone can make an account and join a room from another server which imports ALL the data.
Like illegal data. Images and videos you never want to be associated with. But now your server has a permanent copy of them in its database. Forever, or until you nuke the entire server and make a new one.
Good luck trying to explain that one to the Feds doing a no-knock raid on you at 3am.
Matrix IMHO is suitable only for situations where you *need* the data to be permanently stored as an audit trail // immutable, like corporate and government internal communications.
It doesn't make sense for IoT like they tried to advertise. And it is not a good fit for regular people's needs.