Live data from Hacker News

Bluesky announces data federation for self hosters

bsky.social

361–370 of 471 posts

Re: Bluesky announces data federation for self hosters

#361
post #5

Hey HN, the engineering team at Bluesky is especially excited to get to this point! We're happy to help answer questions and help anyone trying to run their own PDS host. Things should work pretty well for self-hosters right now, but we're standing by to help if there are any problems. Technical details and the installer are in the GitHub repo https://github.com/bluesky-social/pds And we're on Discord available to he…

Given the PDS server works on ports 80/443 and I'd like to use a domain (@nytimes.com in the documentation, but say @example.com), how does it interoperate with existing services that already operate on @example.com , for example a website, blog, cloud.

I'd imagine this use case is quite common for self hosters. If it can't operate alongside an existing, say, nginx on this port, are there recommended alternate practices?

I'm excited at separating identity from hosting, of which self hosting identity gets us closer.

Re: Bluesky announces data federation for self hosters

#362
post #64

Earlier quoted context omitted.

I think everything you said was fair, but you also mentioned Twitter being a conservative cesspool, and a lot these features like federation and composable moderation are designed to help prevent the whole "rich guy buys the company and turns it into something you don't like" scenario.

It's really unfortunate that the tech companies set the precedent in the first place by pushing hard political agendas into their policies and moderation biases. If it was truly neutral in the first place we would not be having this conversation. All the people complaining only now about Twitter doing this are part of the problem.

There is no "true neutral" when it comes to moderation. There are a million examples, but the most obvious are of the form "you can have group X or people who hate group X and are dedicated to driving them off the platform". Somebody's not going to have "free speech" in that case. And even if you go for what most "true neutral" advocates want, which is a lack of rules, you'll quickly find that quite a lot of people don't want to hang out at the place that's filled with Nazis or scam artists or spammers or whatever.

So in practice you have to make choices, or you'll end up running the new 4chan and being sad about your life. As happened to the guy who ran the old 4chan: https://www.rollingstone.com/culture/culture-features/4chans...

Re: Bluesky announces data federation for self hosters

#363

Earlier quoted context omitted.

Basic anti-spam. Just look on Mastodon (and the thankless job of many instance mods) or on how Twitter gets overrun with porn, crypto, crypto scam or troll bots...

Ah yes, incredibly low limits with no way to say "hey look we're quality, let us do more". That's "basic anti-spam". Sure. It's actually "you can host yourself and your friends but anyone who doesn't know someone with a server and technical knowledge must use our platform"

> Ah yes, incredibly low limits with no way to say "hey look we're quality, let us do more". That's "basic anti-spam". Sure.

Then please suggest how they can remedy this problem without giving spammers the same opportunity.

- Requiring upfront financial commitments would deter most low-yield spams, at the cost of raising requirements for running your own node - Domain-based filtering likely works for this, but this brings back in the topic of censorship (see Mastadon's domain blocking)

> It's actually "you can host yourself and your friends but anyone who doesn't know someone with a server and technical knowledge must use our platform"

You're complaining about the immediate first steps and how it doesn't immediately land on the moon. Simplifying the deployment process for everyone else still takes time. At minimum:

- Purchasing gateway for a custom domain name (requires partnerships with domain name sellers) - Server/Compute+DB rental gateway (because the average normal person doesn't have the hardware to run the service 24/7, nor do they want to know about Linux) - Togglable automatic updates to the federation service

Each simplification would require effort on the devs & partners to setup & maintain.

Re: Bluesky announces data federation for self hosters

#364
post #300

Earlier quoted context omitted.

No. There are two components: data storage, and "indexing". You can always, feasibly, own your data. Being your own indexer is less feasible, but you'd choose one like you'd choose a Mastodon instance or something. Portability is trivial because you have the data. "Usk" buys your indexer to put ads on it, you can just move. Moderation only controls what you see on your feed, using a different indexer just means loggi…

I'm confused, your first word is "no," but your answer seems to suggest "yes, usk would be able to have considerable influence" Which is to say, neither storage nor indexing -- but moderation -- the ability to have control over that would be the most important thing. "Usk" buys (or otherwise leverages money to be able to put his thumb on the control of ) bluesky's moderation. Data storage just gives you the ability t…

Posts on AT are content-addressed. The feeds, as they're called now, just return content links. If "Usk" buys a feed provider it's trivial to move to another one. If you're curious why running a feed is more expensive than simply storing data, it's because you need to ingest the actual data for these links before figuring out what to serve in a feed. This storing, indexing, and aggregating is expensive.

Re: Bluesky announces data federation for self hosters

#365
post #360

Earlier quoted context omitted.

I have only brought this up once before on HN and it was over 2 years ago. Not adopting a new project because it is missing something niche is an extremely common reason why people stick with tried and true, mature software. I do not see anything wrong with pointing out niche issues because to some people these issues are important. Because it's broken out of the box it is allowing people who aren't aware of this pro…

Curious. What is the use case here? I’ve spent tens of thousands of hours of my life on the Internet and a lot of that as a sysadm and I’ve not once heard of people accessing or linking to sites this way.

Personally, I just want to properly handle this edge case and it would bother me if my sites didn't handle it correctly. There are advantages for using FQDNs since they are not ambiguous there are extra optimizations that can be done. I don't want my sites to be problematic for people who want to use them so I make sure my sites properly handle them. Usually handling FQDNs is easy as it just works out if the box on most server software.

Re: Bluesky announces data federation for self hosters

#366
post #337

Earlier quoted context omitted.

>If Google announced that they were going to ingest all Mastodon content and post it in a new Google Groups kind of thing, they'd be pretty understandably upset about that, too. exactly like they did with usenet without any issue?

Well, at least they paid money for Deja. Slight difference no?

I'm completely confused under what moral framework the fact that Google paid to buy the Dejanews archive makes a any difference.

To make it clear, for people who don't know:

Google Groups was originally Dejanews, which was a web based archive and front end to Usenet. Google started searching Usenet, but didn't have historical archives so they bought Dejanews.

Obviously no one who posted on Usenet got paid under this transaction.

It's like if Google bought a Mastadon archive off someone now: this argument seems to indicate that would be better somehow than Google archiving Mastadon posts themselves.

I don't understand why at all?

Re: Bluesky announces data federation for self hosters

#368

Earlier quoted context omitted.

I'm surprised that the tool in question is Bridgy Fed. Bridgy Fed has existed for a long time and is a very useful tool. Its alternative, Bridgy, has also been used to bridge between closed social networks and the open IndieWeb. Why are Fediverse people only angry about it now? It's an open protocol. If you want privacy, don't publish something for the entire world to see. That's just basic common sense. At the very…

Well first not everyone on the fediverse is opposed to the bridge. I agree that public is public. But there are concerns about moderation being incompatible, it’s normal to voice them. As for the fediverse destroying the open web for everyone else, I think you’re hyperboling quite a bit, the fediverse has done mountains to make social media more open, probably more than everyone else.

Yeah you're right, I think I did overgeneralise there. I was meaning more of the culture of "Mastodon users"; Mastodon itself has done a lot to help the open Web too.

Though I think "voicing concerns" is a bit of an understatement. I feel really bad for the developer of Bridgy Fed, working on their passion project and just getting caught up in all this heat and harassment.

Re: Bluesky announces data federation for self hosters

#369

Earlier quoted context omitted.

It's really unfortunate that the tech companies set the precedent in the first place by pushing hard political agendas into their policies and moderation biases. If it was truly neutral in the first place we would not be having this conversation. All the people complaining only now about Twitter doing this are part of the problem.

There is no "true neutral" when it comes to moderation. There are a million examples, but the most obvious are of the form "you can have group X or people who hate group X and are dedicated to driving them off the platform". Somebody's not going to have "free speech" in that case. And even if you go for what most "true neutral" advocates want, which is a lack of rules, you'll quickly find that quite a lot of people d…

True neutral means moderating consistently, judging behavior without regard to identity.

The ideal, platonic version of this would be that moderators only see an "identity scrambled" version of each tweet/post when they make their moderation decision. Like a screen that blinds orchestra musicians when they audition, the human would see a statement like "I hate New Yorkers" and not know if the original message said "I hate New Yorkers" or "I hate Floridians." So they would have to make a decision based on the general principle of whether a statement of this form is allowable.

Anywhere you want to draw the line is fine with me, as long as you draw it consistently.

Re: Bluesky announces data federation for self hosters

#370
post #360

Earlier quoted context omitted.

Curious. What is the use case here? I’ve spent tens of thousands of hours of my life on the Internet and a lot of that as a sysadm and I’ve not once heard of people accessing or linking to sites this way.

Personally, I just want to properly handle this edge case and it would bother me if my sites didn't handle it correctly. There are advantages for using FQDNs since they are not ambiguous there are extra optimizations that can be done. I don't want my sites to be problematic for people who want to use them so I make sure my sites properly handle them. Usually handling FQDNs is easy as it just works out if the box on m…

Can you provide a URL to a page containing a link of this style? I’ll concede that it’s useful to someone if I see it! The term ‘FQDN’ does not always imply the dot at the end?
Post reply on HN