Mastodon Architecture favors Calomnies

Dear TAG,

    Having worked on decentralised social networks for 20 years, I started an 
account on Mastodon 9 months ago with the big move from Twitter.

This last 9 months experience has led me to think that there is a serious 
architectural problem with Mastodon that the TAG should be aware of, 
and that needs Web Science or other fields to investigate. 

I think that the Mastodon architecture helps spread slander in 
two ways:

A) Slanderous Toots spread across instances via tags [1]
B) Political Activists put pressure on part-time admins to censor Toots that 
    respond to those slanders [2] by redoubling the slanderous position

Most Admins are well-intentioned, I believe, and took on this role to help balance 
the problems that centralised social networks pose. 

Being part-time and more interested in particular subjects - mathstodon admins
in mathematics, w3c.social in w3c standards, etc… - they don’t have the 
resources to dive deep into any controversial issue. As a result, to avoid being 
defederated - the equivalent of ex-communication - they comply with these requests, 
but without fixing the problematic slanderous toots spreading via the tags.

As a result, we have a social machine that promotes only one specific view on
political topics, those favoured by extremist Activists, who also stop any debate 
on topics that risk questioning their position. Leading to more and more extreme
positions.

Compare this with blogs and RSS feeds. People there can write their 
own content on a server of their choosing and can choose to follow 
whomever they want without an intermediary admin to supervise what is seen or
said. 

The problem we had with the blogosphere was the inefficiency of conversations 
which, due to spam, which had to be filtered. I am working on a way to resolve a part 
of that problem with Solid access control extensions. 

But I don’t think that one solution only will do. For example,
for certain types of conversations, p2p duplication of signed content as 
in https://nostr.com/ will be a requirement for serious conversation 
(to avoid people changing or retracting their statements at a later date). 
Perhaps a Solid extension to IPFS would be needed for that.

It’s not an easy problem [3].

Henry


[1] They don’t reliably spread, as some instance filter some out and not others
   Here is a list of tags across instances of the RFKJr tag
   https://w3c.social/@bblfish/110768025055570367
[2] You can see this in action precisely in this thread here:
     https://w3c.social/@bblfish/110778950392224298
[3] https://teesperky.com/product/we-do-this-not-because-it-is-easy-but-because-we-thought-it-would-be-easy-shirt-unisex-tee/