-

Scott M. Stolz

Entrepreneur & Author

Authorship Studio is part of a decentralized social media network powered by Hubzilla.

Mon, 12 Feb 2024 22:43:04 -0600

boris@toolsforthought.social

Ryan Barrett posted about his forthcoming #ATProtocol <> #ActivityPub bridge and had a lot of people lose their minds

https://fed.brid.gy/r/https://snarfed.org/2024-02-12_52106

He wrote a great article back in January entitled “Moderate people, not code” that is good background

> “Whether ActivityPub or ATProto or webmention, the underlying technical protocol a community uses to interact online is a poor way to judge who they are and whether you might like them.”

https://snarfed.org/2024-01-21_moderate-people-not-code

show all 29 Comments

Mon, 12 Feb 2024 22:08:41 -0600

Boris Mann

boris@toolsforthought.social

I’m going to attempt to explain in general terms how I think about an #ActivityPub bridge.

I’m over on Bluesky. I see this interesting ActivityPub account coop-AT-cosocial-dot-ca. I’d like to follow it!

It has open follows, so the bridge subscribes to that account on my behalf. It’s kind of like a “bridge instance”.

Depending on how the bridge is implemented, maybe you see boris-bluesky-AT-bridge-instance

Mon, 12 Feb 2024 22:10:24 -0600

Boris Mann

boris@toolsforthought.social

This is similar to your feed reader (a foreign network protocol) subscribing to the RSS feed of your Mastodon account.

It’s open. The rules of the protocol are being followed. There is no opt-in for that RSS feed — other than turning on “request to follow” in your Mastodon account.

Mon, 12 Feb 2024 23:13:40 -0600

Mark Shane Hayden

msh@coales.co

@boris respectfully I strongly disagree with "moderate people not code".

It is *extremely* important to moderate both people AND code equally and simultaneously. Just as there are toxic people there is toxic software. When we encounter toxic software we must work towards fixing it, and if it cannot be fixed or the developers are unwilling to fix it then it should indeed be moderated like toxic people are...because toxic software attracts and creates toxic users.

I am not totally and permanently closed minded to BlueSky, but in its present state I have serious concerns. The proposed "Composable Moderation" is currently under-specified and there is less control over federation than in AP protocol. As such I will be watching activity from such bridges and absolutely will "moderate code" if need be.

Mon, 12 Feb 2024 23:17:04 -0600

small patatas

smallpatatas@mstdn.patatas.ca

@boris since you mention moderation: if the protocol you're being connected to is one that has either no moderation system, or one that is incompatible with the systems that are standard on the protocol you're using (where/to whom do you report a bluesky user from a fediverse account, exactly? How do you sign up for a bluesky moderation service from the fediverse?) then concern, or even anger, about this being opt-out seems entirely rational and justified, and I don't fault anyone for blocking anyone or anything they choose, quite frankly

Mon, 12 Feb 2024 23:18:43 -0600

small patatas

smallpatatas@mstdn.patatas.ca

@boris as for the RSS comparison, I was unaware that people could reply to your RSS feed?

Mon, 12 Feb 2024 23:20:34 -0600

Boris Mann

boris@toolsforthought.social

@msh it’s perhaps worded a bit clunkily.

Do you agree that implementations of a well-specified protocol should interop with each other?

eG Mastodon should be able to talk to Pixelfed

I know Ryan, I know his software isn’t ill intentioned. I can’t say the same for what someone other than Ryan would do running it.

I’m happy to say “yes and” to moderate people AND code… but the code is mostly going to be “an instance”, run by a person, because that’s how it works.

Mon, 12 Feb 2024 23:21:30 -0600

Boris Mann

boris@toolsforthought.social

@smallpatatas it’s an example of not respecting the “request to follow” that’s in Mastodon. RSS still available even if you have that setting set.

Mon, 12 Feb 2024 23:22:57 -0600

Boris Mann

boris@toolsforthought.social

@smallpatatas on the ActivityPub side of the bridge, all the AP tools are available.

Thinking of this as a “bridge instance” is what I found most helpful.

And so talking to Ryan as the operator of a new server (both code and instance) might be helpful.

Mon, 12 Feb 2024 23:28:04 -0600

small patatas

smallpatatas@mstdn.patatas.ca

@boris Ah, ok so Ryan will be taking on sole responsibility for moderation of a server with millions of people on it doubled from a server with zero internal moderation? Do you not see a glaring issue with this?

Mon, 12 Feb 2024 23:37:11 -0600

Mark Shane Hayden

msh@coales.co

@boris no I don't when I think about it.

Even if the protocol is perfectly well defined I would (and have) blocked AP instances that implement that protocol poorly (even if it was a technically correct implementation).

I think, on the whole, Mastodon and Pixelfed are relatively high quality and are responsibly maintained, so I encourage such interoperability. OTOH I have blocked "indexing" and "discoverability" applications in the past that, while perfectly implemented the AP protocol, abused user's consent. I will not interoperate with toxic applications, whether or not the toxicity was not intentional.

So basically I treat bad software like I treat badly moderated instances of good software...I block them.

This has always been my policy having run email servers for 25 years. You don't federate with open email relays or servers full of spammers just because they do SMTP perfectly.

Mon, 12 Feb 2024 23:40:45 -0600

Boris Mann

boris@toolsforthought.social

@msh right. You’re moderating the people _behind_ the code who are using it with ill intent.

eG a spammer using Pixelfed or a spammer using Postfix

Now if you find SpamMailer9000 software is only built for spam use cases — then yes, see if you can fingerprint that software and block it!

Mon, 12 Feb 2024 23:48:10 -0600

Boris Mann

boris@toolsforthought.social

@smallpatatas individuals on AT Protocol will choose to follow individuals on ActivityPub, and vice versa.

The Bluesky AT Protocol instance is well moderated.

Or at least, as well moderated as Mastodon-dot-Social.

Since this kind of bridged follow is going to have a higher technical bar, I don’t expect a lot of users right away.

Yes, Ryan is going to have to figure out how to take in reports and pass them along in both directions. And probably some mass follow detection / API limits.

Tue, 13 Feb 2024 00:02:20 -0600

Chris Evelyn

chris_evelyn@troet.cafe

@boris But if I don‘t want my content on bsky in the first place?

Tue, 13 Feb 2024 00:06:37 -0600

Scott M. Stolz

scott@authorship.studio

@Boris Mann

Yes, Ryan is going to have to figure out how to take in reports and pass them along in both directions. And probably some mass follow detection / API limits.

Does ActivityPub even have the ability to pass on reports to other ActivityPub-enabled servers? I think we would need to add that feature to ActivityPub before we could add it to the bridge.

Tue, 13 Feb 2024 00:09:25 -0600

Boris Mann

boris@toolsforthought.social

@chris_evelyn block the server, turn on request to follow, post followers only are some of the tools for controlling both who and what server can see your posts.

The bridge instance is a full ActivityPub server instance.

On the other side of it, eventually, might be me running my own ATProtocol Personal Data Server (PDS).

Tue, 13 Feb 2024 00:09:54 -0600

Mark Shane Hayden

msh@coales.co

@boris Well the developers' intent isn't the sole factor in determining toxicity. Sometimes things are just designed badly and result in unintentionally bad consequences.

There are good people who use bad software, and I will still defederate from instances of bad software run by good people, just like I defederate from instances of good software run by bad people. It is all about evaluating the consequences.

So when it comes to this particular bluesky bridge I will be evaluating it not based on there being good people on bluesky, nor on the professed character or intentions of the person who developed or runs the bridge. I will base my decisions on how the software works. If it is doing mass follow requests bases on an "opt out" policy, and if it is building up a massive PDS full of fedi posts to dump in the firehose that can *enable* bad actors to evade fedi blocks, then the bridge is blocked right away.

I will NOT wait for bad actors to take advantage of the toxic software before I block it.

Tue, 13 Feb 2024 00:11:57 -0600

Boris Mann

boris@toolsforthought.social

@scott Mastodon has this built in, so not a standard, you are correct.

I was thinking of my own experience as both admin and user with the Mastodon “send report to remote admins”

By “pass on” at this point I meant talk to the Bluesky team and ATProtocol client devs.

Tue, 13 Feb 2024 00:12:26 -0600

SomeAnoTooter

SomeAnoTooter@mastodon.online

@boris @msh I think one important think is if "who" and "should" maintain. First it should be those who want to have those connections and second no one explicitly should (have to) do it.
The three e tactics known from m$ could still be problematic. But I think floss is resilient enough.

Tue, 13 Feb 2024 00:14:34 -0600

Chris Evelyn

chris_evelyn@troet.cafe

@boris So the work is on me, right.
How about people who want to be on bsky get an account there?

Why not make the bridge opt-in?

Tue, 13 Feb 2024 00:15:41 -0600

Scott M. Stolz

scott@authorship.studio

I think the point that some people are missing is that there is a difference between blocking a protocol and blocking a server.

If a server is misbehaving, by all means, block it. If users are misbehaving, by all means, block them. That is what these moderation tools are for.

But a few rowdy people on a server is not every user on the server, and a few rowdy servers on a protocol is not every server.

And soon the AT protocol is going to have servers that are not operated by Bluesky, just like how ActivityPub has servers that are not run by Mastodon or even using Mastodon software.

Tue, 13 Feb 2024 00:16:38 -0600

Boris Mann

boris@toolsforthought.social

@msh Re: mass follow I was speaking of individuals, it’s a behaviour I have seen in many early networks, where accounts will follow everyone to get follows back.

In any case, yes, that’s how you should evaluate any ActivityPub instance and the software that runs it!

Tue, 13 Feb 2024 00:18:05 -0600 last edited: Tue, 13 Feb 2024 00:22:21 -0600

SomeAnoTooter

SomeAnoTooter@mastodon.online

@boris @smallpatatas might be a bit off topic, but as long as there are feasible tools for, I guess blocking is the right word, I as a normal user wouldn't mind too much about which protocols I'm interacting with.

Tue, 13 Feb 2024 00:22:03 -0600

Boris Mann

boris@toolsforthought.social

@chris_evelyn those are the default tools you have as part of the AP protocol.

It’s the same for managing any ill intentioned ActivityPub server instance.

I could setup an Akkoma ActivityPub server instance and have open sign up and host it in (country where you personally don’t want your data).

If your AP server is setup to federate automatically as is the normal case, either you or your server admin would need to block the server.

Tue, 13 Feb 2024 00:28:04 -0600

Boris Mann

boris@toolsforthought.social

@chris_evelyn Chris, it looks like you’ve got “Request Follow” turned on, so you’re all set.

Tue, 13 Feb 2024 01:57:39 -0600

Chris Evelyn

chris_evelyn@troet.cafe

@boris So if I have that set, my posts won‘t federate to bsky?

There are reasons why some people left the corporate social networks.

There‘s a reason why I (and others) have follow requests on and post mostly unlisted or followers only.

I like the idea of a big, open internet but I also like the idea of having a modicum of control over what I share and with who.

Tue, 13 Feb 2024 03:08:56 -0600

Ben Thompson

jbenjamint@mastodon.scot

@boris thanks for sharing, it's a good read.

The focus on moderation in the piece is right, but the politics will scupper this - I think until experience shows BS will not keep any form of monopoly control of their platform, it's going to be seen as another capitalist attempt to make serious money off of network effects. I imagine many folk will object to contributing to that. So i think this project would be better waiting for that evidence - it might not be a long wait.

Thanks