The one thing I do think the AT Protocol is better at than #ActivityPub is data portability. Being built-in at the protocol layer so that you don’t have to do the migration shit (which doesn’t bring your posts) is better. It just is. I think ActivityPub has more potential overall for the social web beyond just Twitter clones, but that part of #ARProtocol is better.
I think data portability is incredibly important and the fact that Mastodon hasn’t prioritized that, to me, is a failing. And I say Mastodon, not ActivityPub, b/c AP does have some options there. But Mastodon wasn’t designed that way, no matter what people want to pretend. It just wasn’t. My followers/following list is only part of my data. My content is equally important.
@film_girl this and many other things about the project (like UX issues that haven’t been fixed in years) is telling me that it won’t be able to move beyond a niche
@thomasfuchs I sadly agree. I’ve long-argued that I’m long ActivityPub, I’m not long Mastodon. You can’t do these things by committee and I think Gargron has done a good job trying to steer, but has also let too many other people bikeshed.
@film_girl I’m a bit sad about it because if the project had taken the influx and Twitter woes as an opportunity to improve on UX it could have gone so much further already.
I’m hoping for an ActivityPub <> AT Protocol bridge so you can follow and chat with people in both worlds.
@thomasfuchs that would be ideal (unified timeline). There is already support at Bridgy Fed for AT Protocol stuff but I don’t know what work is ongoing to make things better.
@thomasfuchs@film_girl I imagine there's also going to be an element of fatigue and drift in folks' social graph as people continue (or opt not to) movie. I'm not sure I'm in the minority but I'm not sure if I'd prefer to simply log off over joining another service (I don't have the highest opinion of Jack either).
@thomasfuchs@cory that’s totally fair. My feeling has always been that I’m very pro decentralized protocols but I’m convinced that a sane centralized instance with decent UX and rules will be the one that gets most people to use said protocols.
@film_girl@thomasfuchs oh certainly — I imagine Mastodon may not scale beyond the first-adopter set but that may still be a success. (I also think driving folks to mastodon.social to reduce friction is a smart, pragmatic move.)
@erickrm not even close. I’m talking about people able to take my old posts and messages and media and being able to import them into a new instance. An archive is great. But that’s not data portability. That’s an archive.
@codinghorror you can request your data but you can’t do anything with it like bringing it to your own instance. It did take them a long time to even get the download thing tho, tho that is there!
Have been told so often that it's easy to move servers and that followers come with me...which is fine.... but not the content?! My timelines are like a sort of live note taking service for work. I don't want to risk losing that just because my instance host shuts down, for example!
@film_girl I wonder if it could be possible (either in the protocol or in Mastodon) to support a server-to-server transfer directly of user content. Obviously it doesn't support that right now, but I wonder how hard it would be to implement. If it would be, I wonder if it should be a change in the protocol or if it just needs to be a change to the implementation.
@film_girl “You can just move instances!” solves way fewer problems than Mastodon apologists would suggest it does, and is only a half-truth in the first place.
@cadence022 yeah. It was really disheartening to read through years of issues related to this ro see a ton of bikeshedding/excuses/justifications for why this was never a priority. And that’s fine. I know about dev cycles and constraints. But what makes me pause is that it is very clear this was never considered when designing Mastodon. Data portability was very much designed through the lens of the server admin/instance not the user.
@film_girl exactly, and it shouldn't be that way. great, admin can pick up and run off with everyone's data, is this not what we were trying to avoid with decentralization? we should be able to download archives whenever required or wanted of our content and even be able to port them over to the new account with accurate time stamping.
@Cutterpillow which I totally get — but that should be up to the server instance/user, not the project, imho, if you claim you care about data portability. Offering an archive that you can do nothing with isn’t data portability.
@film_girl i keep trying to explain to people that ATProtocol's approach is way more thoughtful then people realize, but I keep getting yelled at about it.
@film_girl like lots of people yelled at them for not just going with activitypub, but they literally considered it and said "this has a bunch of limitations, we can build it right from the ground up."
@mmasnick exactly! When I read the spec again today, I was struck again by how thoughtful it is and how better-designed a lot of it is. And that’s not a slight at AP, but it was clear a lot of things with AP weren’t designed for/considered/prioritized at all over the last 7 years.
@film_girl yup. i'm obviously a fan of AP. But it has limitations. Of course, lots of people try to start from scratch on things like this and it fails, but I'm really encouraged by the approach with ATP.
@mmasnick fully agree. And I’ll continue being unpopular and say that using VC money you’re able to get (or from a rich benefactor, whatever) to pay people to work full-time on designing this stuff is a good thing. It’s easy to reflexively hate anything that has money attached to it but I think there is value in that at the beginning of a standard/protocol. Take the money and use it well! If the app fails, that doesn’t mean the protocol will!
@film_girl@mmasnick I’m curious about how if at all an app could implement parts of both protocols. After all Mastodon (and other federated platforms) use more than just Activity Pub to federate. I need to look into it further.
@andrewfeeney@mmasnick yeah, I’m curious about that too. https://fed.brid.gy has AT Protocol support so I think it theoretically is possible but I don’t know if that pushes up against anything specific to Mastodon
@film_girl@andrewfeeney@mmasnick it’s very possible, and may well mirror something like how email clients speak IMAP and SMTP and POP3 and nobody even cares or notices.
@film_girl@mmasnick I'm curious how, technology aside - you can reconcile this with the fact Jack is still at the helm, and money still comes from Twitter? Last I read (and please correct if I'm wrong) the purse strings are very much tied there.
If that still is the case and he pulls all support - it still feels very messy. You all tend to follow this close than I so like I said - happy to stand corrected if that isn't valid info anymore. 🙂
@cdlhamma@mmasnick I think it started as a Jack/Parag thing but has evolved. He’s not the CEO, tho he may have hired her. I don’t know if he’s even active but like for me, it’s the same as any other company. You get in while the good times are rolling, you leave when they end. At least this is a protocol. I’m not saying it’s going to work or win, I’m saying I like the design of the protocol.
@film_girl@mmasnick yeah I get that! The protocol does look promising 🙂 its all the other baggage I have a hard time with, the non technical stuff of how its funded, how they'll make money etc. Time will tell I suppose.
@cdlhamma@mmasnick totally. Over the weekend, Jack was apparently bitching about Bluesky on Nostr (the other decentralized app but this one is crypto) so I guess we’ll see, as you said :). As for sustainability, I guess it depends on how much they raised and their runway but like, stuff is open. If it works, it’ll work beyond just Bluesky. If it fails, it’ll be just like 99% of the social apps out there.
@film_girl@cdlhamma jack is on the board, but disagrees with the direction Bluesky is going in, and is devoted to Nostr instead, which he thinks is a better approach (both are interesting to me)
@misc@mmasnick@cdlhamma he doesn’t like that Bluesky is a Twitter clone. Which seems weird. B/c of course you start with the Twitter clone then you can get weirder.
@film_girl@misc@cdlhamma yeah, my understanding is he thinks Bluesky is too much like Twitter. He also thinks they should have launched before moderation tools were ready.
Some of this is driven by how much he likes "simplicity" of nostr, in which they just defined the most basic concepts and then leave the details up to everyone else.
@ryanbooker@film_girl@cdlhamma I don't disagree. Nostr is not at all user friendly, and I started using it when there were less than 1000 people there. But it has some interesting things going on, and I know some people are developing more consumer friendly onboarding.
@ryanbooker@mmasnick@film_girl@cdlhamma coming from people who've dealt with the onboarding into mastodon*, that says a lot. but yeah, as soon as you see JSON on the front page you are in for a rough time.
@film_girl@mmasnick This is an interesting stance. I can’t think of a single time that a VC funded protocol has survived as an open one, but I can think of a bunch of times that openness has been quashed for profit.
Though perhaps I’m just not remembering something obvious that is convincing you this is a good path forward.
Their theories are interesting and I have an account to check it out, but I doubt I’ll stick with it.
@amd@mmasnick I’m not comparing this to JavaScript, to be clear, but literally every aspect of JavaScript from its initial creation to Ajax to React, to Angular to Node to npm to TypeScript have all been funded by VCs or already public companies. jQuery is the only library I can think of that never had a big VC daddy. But every part of ECMAScript is from public or VC-funded companies.
@skotchygut@mmasnick right now, there aren’t examples for setting up your own alternative node that I can find (Bluesky is still in invite-only mode) but it’s all there in the docs and in the way the protocol works. Obviously, this is still early. But yes, the idea is very much that. https://atproto.com/docs
It's good for everyone in the decentralized social space to learn off of each other. Not just be reflexively in tribal camps.
And #ActivityPub folks need to move faster on their own nomadic identity systems, rather than diss others. But most I see that are serious on the AP side are seeing that as better way: learn and move faster.
@lapcatsoftware it does! I’ve held off migrating to my own instance for this very reason. And attempts to get a solution supported (a PR is basically there for the taking) is ignored and argued over by the masterotti.
Add comment