Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Building on ATProto

lukekanies.com

Post reply on HN
Print view View original

91–99 of 99 posts

Re: Building on ATProto

#91

Post by inigyou »

embedding-shape wrote: post #85
inigyou wrote: post #73

Earlier quoted context omitted.

Matrix's biggest blocker is that it doesn't work very well . You try it, you get a worse experience than Discord, you switch back to Discord. I hope the Matrix project figures this out one day.

What do you hope they'll figure out by ambiguous feedback like this? What exactly is worse? Is it possible it's just different, and that translates to "worse" although maybe it's actually on purpose? What specifically is it that doesn't work well for you?

I mean... just use it for a while and see? It's death by a thousand paper cuts. There's no single big problem. It does get messages from A to B.

For instance when you click on a new room and the messages are stuck in a loading state forever. Or when you sign out and sign back in and need to reset decryption keys and can't read message you received before signing in

Re: Building on ATProto

#93

Post by verdverm »

inigyou wrote: post #74
verdverm wrote: post #12

Earlier quoted context omitted.

How do you prevent a player from manipulating game state outside of the app view? There is a subgroup in the Atmo working on games, have you found them yet?

Publish the moves, not the game state?

This would imply the game state is stored in a centralized app view, at which point the general atproto ethos that you can move your data if an app becomes bad is up to the app.

I believe the viable solution in a rich permission system is one where the user cannot directly modify a record (as defined by a permission schema for a specific app/type) and can only call methods.

Re: Building on ATProto

#94

Post by solarkraft »

embedding-shape wrote: post #85
inigyou wrote: post #73

Earlier quoted context omitted.

Matrix's biggest blocker is that it doesn't work very well . You try it, you get a worse experience than Discord, you switch back to Discord. I hope the Matrix project figures this out one day.

What do you hope they'll figure out by ambiguous feedback like this? What exactly is worse? Is it possible it's just different, and that translates to "worse" although maybe it's actually on purpose? What specifically is it that doesn't work well for you?

Acute awareness of and attention to the overarching problem. Almost all of the individual issues are likely already known, but too often hand-waved away as minor inconveniences. This is common in the FOSS world and IMO the #1 hurdle for wider adoption.

That said, it would not be fair to not acknowledge that the Matrix 2.0 efforts have been real and massive improvements. Refocusing like this is a massive organizational effort and takes a non-trivial amount of resources. I’d like to take the opportunity to really commend that.

(That said, I have paused my close following of the project due to my frustration with the state of things and rate of progress, but I totally understand some of the reasons for it. It is only becoming more and more important to have an open standard for instant messaging, but I wouldn’t be surprised if at some point something new appeared and immediately captured all attention)

Re: Building on ATProto

#95

Post by verdverm »

blackqueeriroh wrote: post #65
verdverm wrote: post #29

Earlier quoted context omitted.

the cultiness is why I left crypto and atproto, it plagues many great ideas by creating in and out groups around opinions on the right way to do things

Really strange, considering there’s more debate, disagreement, and lack of conformity in atproto than I’ve ever seen in crypto. I think you might mean, “people who disagree with me intensely because they believe strongly in the way they view things.” Atproto is very, very far from a cult. It is a bunch of poasters, though, and nobody should be surprised when a social protocol has a bunch of, well, really social peopl…

I hope you will agree that Paul's personal attacks towards me are unacceptable and unbecoming for the CTO of Bluesky. This is not the first time, I hope it is the last.

(see his peer comment flagged by the community here)

Re: Building on ATProto

#96

Post by verdverm »

inigyou wrote: post #78
blackqueeriroh wrote: post #63

Earlier quoted context omitted.

You guys are really confused about what atproto has to do with Bluesky and how much of the protocol is developed by or managed through Bluesky. It really seems you all think the only company or community doing anything on atproto is Bluesky, when that couldn’t be further from the truth. Have you even paid attention to what’s happening lately?

Well the whole protocol relies on a "PLC server". There can only be one and guess who owns it.

I think a more potent example is who controls the source for the spec and official implementations. This is a notable lack of community governance around this part. In this sense, Bluesky can and does change things at will. I have experienced breaking changes from their devs changing how things work and merging their own PRs in a matter of minutes, while not using versions. (indigo)

When it comes to private/permissioned data, Bluesky people have decided the shape it will take while not participating in the conversation until they shared their already decided upon design. They are letting the community give feedback, but the community never was able to participate in the design phase.

Re: Building on ATProto

#97

Post by inigyou »

verdverm wrote: post #93
inigyou wrote: post #74

Earlier quoted context omitted.

Publish the moves, not the game state?

This would imply the game state is stored in a centralized app view, at which point the general atproto ethos that you can move your data if an app becomes bad is up to the app. I believe the viable solution in a rich permission system is one where the user cannot directly modify a record (as defined by a permission schema for a specific app/type) and can only call methods.

You can easily reproduce a chess board from the moves if you can implement chess.

Re: Building on ATProto

#98

Post by inigyou »

inigyou wrote: post #91
embedding-shape wrote: post #85

Earlier quoted context omitted.

What do you hope they'll figure out by ambiguous feedback like this? What exactly is worse? Is it possible it's just different, and that translates to "worse" although maybe it's actually on purpose? What specifically is it that doesn't work well for you?

I mean... just use it for a while and see? It's death by a thousand paper cuts. There's no single big problem. It does get messages from A to B. For instance when you click on a new room and the messages are stuck in a loading state forever. Or when you sign out and sign back in and need to reset decryption keys and can't read message you received before signing in

Or when someone gives you a link to a Space and you click on it and get a loading spinner for a while and then an obscure error. What is a Space anyway? Why can't I organise my own channels into groups but only the admin of a Space can do that? It isn't intuitive. Obviously it makes sense to have groups of channels like Discord guilds, in some way, but it's currently confusing.

When I know the address of a channel I can never find the right textbox to put this in. And when it is the right one, it doesn't give feedback that it's the right one. IIRC it's the same box that searches channels on the current home server, but if you put an address there it shows 0 results, but if you click on continue then it joins the channel anyway. But there's another box somewhere in the app that also does this search thing but can't join a channel address. I think. And neither of them are in obvious places. In Discord, at the end of your list of guilds, is a big + button, which opens a text box asking for the address of a guild.

I really think you should just sit down with some inexperienced or first-time users and watch them do some basic tasks while thinking aloud. This used to be the bread and butter of the UX industry, which has atrophied.

Re: Building on ATProto

#99

Post by cyphar »

inigyou wrote: post #97
verdverm wrote: post #93

Earlier quoted context omitted.

This would imply the game state is stored in a centralized app view, at which point the general atproto ethos that you can move your data if an app becomes bad is up to the app. I believe the viable solution in a rich permission system is one where the user cannot directly modify a record (as defined by a permission schema for a specific app/type) and can only call methods.

You can easily reproduce a chess board from the moves if you can implement chess.

Most games are not perfect information games. How would you implement poker?
Post reply on HN