Mastodon Web Space Search Tool
researchbuzz.me
researchbuzz.me
Since most people think of mastodon as one single service, why can’t you just search all known Instances by default and then allow users to reduce to specific instances as a sort of filter option?
Reason is because this early on , not looking like there is any real logic to which instance you are going to find one user or content and you’ll essentially have to guess where to point your search query at.
This was somewhat covered previously https://news.ycombinator.com/item?id=34221851 but is worth having a look at.
In short though, any attempts to do this tends to result in huge block lists, the hashtag #fediblock circulating around, and in some cases outright personal attacks on the people behind it.
Then a movement to defederate anyone who allows indexing.
Then hopefully a move to allow-list federation which I personally have suggested for a while now as the only sensible move if they really want some privacy.
(Checking for “web search tool”, #Google and #Fediblock from mastodon.social, it seems like people have not noticed yet.)
I would like to point out that HN has actually discussed the “no search social media” idea before, where it was called pseudo-privacy[1].
Whitelisting makes me worried for the future of decentralised social media, as higher barriers for creating new instances encourage consolidation instead of diversity and decentralisation. It’s true that it would make moderation easier, though. Indeed, one of the main barriers to spam right now is perhaps the cost of domains.
Speaking of which, this also makes me worried how the Fediverse will deal with its inevitable spam problem, without simply saying “the volunteer mods will take care of it”. Or perhaps that is a legit strategy, I’m not sure.
I have a feeling that Hashcash could be a solution, and where instances do not participate in it you could slow their connections in. Just an idea though.
The list is available on Wikipedia[1] but the original ISO document is ridiculously behind a $180 paywall[2].
I would also take this opportunity to point out that Lemmy already has full-text search.
From my experiments, the user’s own pages are opted-out, but their content is federated to other servers (by design), which is then rendered on other servers without the opt-out. Or at least, where their content is part of another piece of content (eg as a reply, or as a parent post).
1) Privacy by obscurity. The platform was too small for people to care, so content was sort of private, until it was put into the spotlight.
2) Social shaming as a way to maintain the above. However, others have pointed out that this does not work on bad actors, or corporations such as Google.
As a server admin, I attempt to resolve this by making it very clear in the privacy policy that all content is public. Even direct messages can potentially be public as it is not encrypted and federated to a number of untrusted servers.
Indeed, even the default Mastodon UI warns the user of such, telling them not to share sensitive information over direct messages, and when deleting their account, mentioning that their content may still (very likely) live on in other servers.