Private and Public Mastodon

Does Mastodon have something to the effect of robots.txt (or is robots.txt already robust enough to express an answer to the question "This site administrator is willing to have these toots indexed for search?")

At the end of the day, it's a federation protocol. The protocol's creator can have an opinion on whether things should be indexed for search, but it's up to users of the protocol, not the protocol's creator, how it's used.

(... and I have to comment on this bit from the article. "The problem would be a public search engine that Gamergaters and Kiwifarmers use to hunt down vulnerable targets." This heavily implies that Gamergaters and Kiwifarmers can't just build that for their own purposes, which would, of course, ignore robots.txt settings).

There is a story [0] from November doing the rounds on Twitter today (but which was flagged on HN) that does a good job of expanding on why there is no search function on Mastodon. Namely, it's* full of child porn, or "lolicon" which is drawn child porn, illegal in many Western nations but not in Japan. In the words of Mastodon Founder Eugen Rochko:

> Lack of full-text search on general content is intentional, due to negative social dynamics

I guess you could frame this as a privacy argument, but I don't buy that as the primary reason people were upset with "Fedisearch." They were upset because it exposed their communities of lolicon enthusiasts. Do they deserve privacy or is that more of an excuse to keep their content off the radar of law enforcement and out of reach of polite society?

I did find this 2017 GitHub issue [1] to be slightly less inflammatory and full of fruitful discussion about legal implications of caching strategies for media content on Mastodon instances.

[0] https://www.secjuice.com/mastodon-child-porn-pedophiles/

[1] https://github.com/mastodon/mastodon/issues/1847

* "It" being the wider fediverse, where the 2nd and 3rd most popular instances allow "lolicon" content (which is illegal in many jurisdictions). Of course not all of Mastodon is filled with this, but all it takes is an obsessive minority to poison the grapevine with a large amount of content. Also, you can expect generative AI to lead to even more of this stuff.

Facebook got sued a whole bunch. Between the FTC's various privacy-related consent decrees, Cambridge Analytica, and the GDPR, they functionally can't do the whole "push people to overshare" thing anymore.

I will say though, it's specifically the old Facebook mentality that made me hyped back when Google+ was announced. And then it crashed and burned - which was exactly the point in time where I stopped wanting to be on literally any new social network.

Use a private group chat, and it can be decentralized too using matrix!

The fediverse is a public protocol, and indexing public content is fair use(ML is not indexing, its very different), so any kind of licensing won't matter.

Discovery is already crap in mastodon and the fediverse in general, please don't stop people from improving it.

It may look like eternal September for fediverse natives, but that's what going mainstream is like, don't like it? Create something with some kind of intentional barrier like HN to stop the masses from joining.

I find the fear about getting doxed and having your words watched and used against you from fediverse natives very funny, because I guarantee you that most of them support cancel culture.

In general I find old fediverse communities too fragile, one example are the crazy content warnings(just mute words and topics you don't like, simple!) and now this.

Absolutely, I support E2E and signatures. It's funny that there is nothing new since GPG, it is decentralized identity with E2E encryption and verifiable signatures. But original poster was questioning the delivery mechanism. Given that there is no E2E, it already works as efficiently as it can with pub/sub.

I don't want to start discussing what Fediverse should be, but I have had this discussion in past, few points:

- ActivityPub has two signatures: HTTP signatures, which are ephemeral, and more persistent JSON-LD object signatures. They are contentious. Some servers disable JSON-LD object signatures, because they allow to proof cryptographically that someone posted something, thus deleted posts become "liability".

- E2E encryption will have similar ramifications, because not all servers agree that crypto be used.

Currently only way to do "privacy" is to deliver messages to just the people you want, and hope servers won't send them to third parties, like in email. That's what Mastodon does.