Misskey talking about leaving the Fediverse
-
Not sure what this is about. ActivityPub is not a complex protocol. At the end of the day, it is just a REST API. Servers send activities to inboxes, which then gets processed.
Do they want to centralize Misskey? They should look at Mastodon and realize they aren’t that big. FediDB literally shows ~11.3k MAU.
If it was originally written to sync blogs or wikis, I can understand not wanting to continue running or developing it.
I think they are just saying “I was writing misskey for a different purpose, and it (nor I?) can handle the speed of ActivityPub and its development.”
I assume there’s a Google translation issue that caused the post to be confusing.
-
If it was originally written to sync blogs or wikis, I can understand not wanting to continue running or developing it.
I think they are just saying “I was writing misskey for a different purpose, and it (nor I?) can handle the speed of ActivityPub and its development.”
I assume there’s a Google translation issue that caused the post to be confusing.
It wasn’t. Even the very first appearance of ActivityPub in W3C mentioned it being a social networking protocol:
The ActivityPub protocol is a social networking protocol based upon the ActivityStreams 2.0 data format. It is based upon experience gained from implementing and working with the OStatus and Pump.io protocols.
OStatus and Pump.io are both social networking protocols as well.
I wouldn’t say that “social networking” is just syncing blogs and wikis.
The vocabulary is also not that small and the flexibility is also there. It is clearly not meant for just blogs and wikis.
-
It wasn’t. Even the very first appearance of ActivityPub in W3C mentioned it being a social networking protocol:
The ActivityPub protocol is a social networking protocol based upon the ActivityStreams 2.0 data format. It is based upon experience gained from implementing and working with the OStatus and Pump.io protocols.
OStatus and Pump.io are both social networking protocols as well.
I wouldn’t say that “social networking” is just syncing blogs and wikis.
The vocabulary is also not that small and the flexibility is also there. It is clearly not meant for just blogs and wikis.
This doesn’t really apply to why misskey was originally created though, which is what this post implies.
What misskey was used for and why it was created can be two things.
-
Google Translate:
First of all, the ActivityPub system is not suited to something like Misskey, which flows at lightning speed.
It was originally designed to connect blogs, small-scale SNSs, and wikis.
How can it handle the TL hell where tens of thousands of requests fly in per second?
It’s based on the idea that it would be nice if various small services could send each other updates, so it’s quite costly.
So what is Miss Key? Is it like Tumblr? What’s the culture like?
-
I think the main alternative to Misskey is its fork, Sharkey? I’m not personally familiar with them though
I think misskey.io is the flagship instance for Misskey so they would not switch since it’s their own software? other instances might
Uh sorry, I was a bit unclear. I meant their alternative to leveraging ActivityPub for what they do.
-
And do they have some alternative? I mean without anything to replace it with, that probably means switching off the networking and either split up or all agree on one large service?!
It sounds a bit like they drank too much ATproto coolaid, which is designed for “firehose” like feeds, but does so at the expense of horizontal scalability, making Bluesky a near natural monopoly on the protocol.
-
It sounds a bit like they drank too much ATproto coolaid, which is designed for “firehose” like feeds, but does so at the expense of horizontal scalability, making Bluesky a near natural monopoly on the protocol.
May the best team win.
I’m not sure if trading in these things is very helpful, it probably doesn’t get us all the way. But I can empathize. I feel some nostalgia for the times when I ran mostly text based communication on a potato. And that’s not how any of the more modern tech works. But there’s an entire complicated story behind it.
-
Misskey has 3,535 monthly active users according to https://misskey.fediverse.observer/dailystats
edit: actually fediverse observer seems to be unable to get active user counts from Misskey instances, many show 0 active users but with thousands of registered users https://misskey.fediverse.observer/list
Might be a privacy feature, I checked out the NodeInfo of one of the 0 active users instances (which I checked and did have active users) and it had “null” in the relevant fields. Only the total users were given.
So the data just doesn’t seem to be there for those instances. Considering it works for some and not for others, I assume it’s opt-out for some reason.
That said, FediDB gives 11.3K MAU for Misskey.
-
May the best team win.
I’m not sure if trading in these things is very helpful, it probably doesn’t get us all the way. But I can empathize. I feel some nostalgia for the times when I ran mostly text based communication on a potato. And that’s not how any of the more modern tech works. But there’s an entire complicated story behind it.
How can the best team win when ATproto is structurally giving Bluesky an advantage over any possible other team? The best solution is not to play such a rigged game.
ATproto and ActivityPub are not in a competition, despite many people having this wrong impression. ATproto ultimatly tries to solve a different problem than ActivityPub.
ATproto tries to make the influencer type social media (many follow a few) more robust by giving those influencers better data ownership and outsourcing moderation to community operated filters. At least in theory, as most of it didn’t work out with Bluesky dominating everything and few others interested in becoming part of it.
ActivityPub ultimatly tries to solve the social network problem where network effects lead to large centralized walled gardens like Facebook or Reddit locking in users. The overall design is thus modeled on that problem and it isn’t particularly suited to cater to social media influencers with large amounts of followers.
-
How can the best team win when ATproto is structurally giving Bluesky an advantage over any possible other team? The best solution is not to play such a rigged game.
ATproto and ActivityPub are not in a competition, despite many people having this wrong impression. ATproto ultimatly tries to solve a different problem than ActivityPub.
ATproto tries to make the influencer type social media (many follow a few) more robust by giving those influencers better data ownership and outsourcing moderation to community operated filters. At least in theory, as most of it didn’t work out with Bluesky dominating everything and few others interested in becoming part of it.
ActivityPub ultimatly tries to solve the social network problem where network effects lead to large centralized walled gardens like Facebook or Reddit locking in users. The overall design is thus modeled on that problem and it isn’t particularly suited to cater to social media influencers with large amounts of followers.
ActivityPub ultimatly tries to solve the social network problem where network effects lead to large centralized walled gardens like Facebook locking in users. The overall design is thus modeled on that problem and it isn’t particularly suited to cater to social media influencers with large amounts of followers.
I never thought about it this way. Probably that’s why Mastodon was always struggling with discovery, while on the other hand federation on the Threadiverse works quite well
-
And do they have some alternative? I mean without anything to replace it with, that probably means switching off the networking and either split up or all agree on one large service?!
They’re working on their own federation system which they’ll only switch over to when it’s ready. They expect things to be fine for the next few years still.
-
ActivityPub ultimatly tries to solve the social network problem where network effects lead to large centralized walled gardens like Facebook locking in users. The overall design is thus modeled on that problem and it isn’t particularly suited to cater to social media influencers with large amounts of followers.
I never thought about it this way. Probably that’s why Mastodon was always struggling with discovery, while on the other hand federation on the Threadiverse works quite well
Mastdon’s discoverability issues are as political as they are technical. It fundamentally requires some aggregating and process of people’s data, something Mastodon users are infamously opposed to. Not for entirely illegitimate reasons, but the extent some go to oppose it is troubling.
-
Mastdon’s discoverability issues are as political as they are technical. It fundamentally requires some aggregating and process of people’s data, something Mastodon users are infamously opposed to. Not for entirely illegitimate reasons, but the extent some go to oppose it is troubling.
Ah, that makes sense. I assume than that discoverability is better on Sharkey and other key forks?
-
Ah, that makes sense. I assume than that discoverability is better on Sharkey and other key forks?
Maybe? You’d have to ask someone who’s used them.
I’ve actually never used anything beyond my basic mastodon.social account.
I’m hoping to finally be able to move of it if FEP-1580 goes anywhere: https://neuromatch.social/@jonny/115376513498494866
-
Maybe? You’d have to ask someone who’s used them.
I’ve actually never used anything beyond my basic mastodon.social account.
I’m hoping to finally be able to move of it if FEP-1580 goes anywhere: https://neuromatch.social/@jonny/115376513498494866
Let’s hope!
-
They’re working on their own federation system which they’ll only switch over to when it’s ready. They expect things to be fine for the next few years still.
Is there any background information on this? Wikipedia mentions this but they only link to the very same post.
-
Is there any background information on this? Wikipedia mentions this but they only link to the very same post.
I saw a separate post linked in the Reddit thread on this, where it was said.
https://misskey.io/notes/af5udbqosfca05lo
Via machine translation:
I won’t be leaving the federation immediately, but I will be migrating to the low-overhead federation system MisskeyHQ is developing. Servers incompatible with that system may eventually lose connectivity.
The full transition will likely take several years, so it should be fine until then.
Source for the full quote of link + translation (Reddit alert)
Don’t know more information than this. But the wording there suggests to me that what they’re switching to isn’t fundamentally incompatible with the fediverse. They seem to be treating loss of compatibility as an allowed risk, not a guarantee. So they’re probably still using ActivityPub as a foundation for what they’re building.
-
How can the best team win when ATproto is structurally giving Bluesky an advantage over any possible other team? The best solution is not to play such a rigged game.
ATproto and ActivityPub are not in a competition, despite many people having this wrong impression. ATproto ultimatly tries to solve a different problem than ActivityPub.
ATproto tries to make the influencer type social media (many follow a few) more robust by giving those influencers better data ownership and outsourcing moderation to community operated filters. At least in theory, as most of it didn’t work out with Bluesky dominating everything and few others interested in becoming part of it.
ActivityPub ultimatly tries to solve the social network problem where network effects lead to large centralized walled gardens like Facebook or Reddit locking in users. The overall design is thus modeled on that problem and it isn’t particularly suited to cater to social media influencers with large amounts of followers.
Sure. I was coming from a different perspective. We have all these competing protocols, I’ve used Matrix, XMPP, eMail, several dialects of AP and there’s a bunch more out there, and a plethora of different and/or overlapping use-cases. My definition of a “win” would be to connect people. In a meaningful way and to contribute to their lives. Make it easy to participate and all these things. I don’t really care that much about the exact protocol, that’s just a tool to achieve some goal.
I think it’s way more important what we build with these things. My ideal platform would do away with the unhealthy social media dynamics we’ve inherited. It’d be full of people discussing hobby projects, share their travel stories and what’s important to them. Have answers to weird computer problems… Politics would be more than a shouting match. It’d empower people to be constructive while talking to each other. I think the protocol is important to lay out the groundworks for that, but ultimately it is not in itself the standard by which we measure our success.
-
yea, the Threadiverse probably gets a relatively high number of requests since the upvotes/downvotes are also requests
but everything seems to be working fine over here
I don’t know if I would say it’s working fine. If you don’t clear your cache regularly the main page takes like a full minute to load.
Hopefully it’s fixable but this is a level of performance most internet users would not tolerate.
-
I don’t know if I would say it’s working fine. If you don’t clear your cache regularly the main page takes like a full minute to load.
Hopefully it’s fixable but this is a level of performance most internet users would not tolerate.
if a client side action (clearing cache) affects the issue, then it’s not related to the ActivityPub protocol
