Visualizing Mastodon server moderation | InfoWorld

In Mapping the broader fediverse I confirmed how a brand new desk added to the Mastodon plugin—mastodon_domain_block—allows queries that discover which servers are moderating which different servers. For instance, listed here are servers on nerdculture.de’s record of offenders.

choose
  server,
  area,
  severity
from
  mastodon_domain_block
the place
  server="https://nerdculture.de"

+------------------------+--------------------------------+----------+
| server                 | area                         | severity |
+------------------------+--------------------------------+----------+
| https://nerdculture.de | roysbeer.place                 | silence  |
| https://nerdculture.de | *.10minutepleroma.com          | droop  |
| https://nerdculture.de | *.activitypub-troll.cf         | droop  |

...snip...

| https://nerdculture.de | shitposter.membership                | droop  |
| https://nerdculture.de | wolfgirl.bar                   | droop  |
| https://nerdculture.de | www2.gabbers.me                | droop  |
+------------------------+--------------------------------+----------+

I used the brand new question sample to construct a dashboard to report, for every server within the residence timeline:

• The record of blocking servers.
• The rely of blocked servers for every blocking server.
• The record of blocking servers for every blocked server, and the rely of these blocking servers.

This was a great begin, however I had a hunch that relationship graphs would reveal non-obvious connections amongst these servers. And certainly they do! Right here’s a view of the brand new graph-enriched dashboard.

IDG

The left pane exhibits the blocked servers shunned by a blocking server chosen from these within the residence timeline. The correct pane has the inverse view: the blocking servers that shun a specific blocked server. These had been the 2 classes I outlined for the primary iteration of those graphs.

class "blocking_server" {
  shade = "darkgreen"
  icon = "server"
}

class "blocked_server" {
  shade = "darkred"
  icon = "server"
}

Right here’s the code for one of many graphs.

graph {
   node {
    base = node.blocking_server
  }
   node {
    base = node.blocked_server
  }
   node {
    base = node.blocked_and_blocking_server
  }
   edge {
    args = [ self.input.blocking_server.value ]
    base = edge.match_blocked_server
  }
   edge {
    args = [ self.input.blocking_server.value ]
    base = edge.match_blocking_server
  }
}

Right here is the definition of node.blocking_server, which refers to class.blocking_server.

node "blocking_server" {
  class = class.blocking_server
  sql = <<EOQ
    with servers as (
      choose distinct
        blocking_server,
        blocked_server
      from
        blocking_servers(${native.restrict})
    )
    choose
      blocking_server as id,
      blocking_server as title
    from
      servers
    order by
      blocking_server
  EOQ
}

The FROM clause calls blocking_servers(), a set-returning perform outlined like so:

create or substitute perform public.blocking_servers(max int) returns desk (
  blocking_server textual content,
  blocked_server textual content
) as $$
with servers as (
  choose distinct
    server as area,
    'https://' || server as server_url
  from
    mastodon_toot
  the place
    timeline="residence"
  restrict max
),
blocking_and_blocked as (
  choose
    s.area as blocking_domain,
    d.area as blocked_domain
  from
    servers s
  be part of
    mastodon_domain_block d
  on
    s.server_url = d.server
)
choose
  blocking_domain,
  blocked_domain
from
  blocking_and_blocked
order by
  blocking_domain, blocked_domain
$$ language sql

I assumed these substances would suffice. However once I started poking round within the graphs made with these definitions, infosec.alternate behaved unusually. Generally it could seem as a blocking server, different instances as a blocked server. I used to be lacking a class!

class "blocked_and_blocking_server" {
  shade = "orange"
  icon = "server"
}

As you possibly can see within the graph, infosec.alternate not solely blocks 73 servers, it is usually blocked by two servers: faith.masto.host and weatherishappening.web. Why?

weatherishappening.web calls infosec.alternate a “HOSTING DEATH SPIRAL FASCIST ORGANIZATION” and blocks it with “Restricted” severity.

faith.masto.host blocks infosec.alternate on the “Suspended” degree, however doesn’t clarify why (“Motive not obtainable”).

Though these servers may, in flip, be blocked by others, no such blocks seem within the block lists of my instant server neighborhood.

> choose rely(*) from blocking_servers(100) the place blocked_server="weatherishappening.web"
+-------+
| rely |
+-------+
| 0     |
+-------+

> choose rely(*) from blocking_servers(100) the place blocked_server="faith.masto.host"
+-------+
| rely |
+-------+
| 0     |
+-------+

There’s one other change hitter in my neighborhood, although. c.im is blocked by me.dm and octodon.social.

IDG

Why?

me.dm (Medium) has “Suspended” c.im for “Hate speech”.

octodon.social has “Suspended” c.im for “Motive not obtainable”.

When the opinions and insurance policies of your server differ from these of mine, we see totally different realities via our respective lenses. May such fragmentation drive the fediverse-curious again into the arms of Large Social? I’m positive that may occur—certainly is going on—to some extent.

However I hope that a few of us, no less than, will study to thrive in various networks of on-line communities, conscious of the kaleidoscopic interaction of filters however not overwhelmed by it. That ability will serve us nicely IRL too. To amass it, we’ll want to visualise the operation of our filters. One good way to do this: SQL queries that drive relationship graphs.

This collection:

1. Autonomy, packet measurement, friction, fanout, and velocity
2. Mastodon, Steampipe, and RSS
3. Looking the fediverse
4. A Bloomberg terminal for Mastodon
5. Create your personal Mastodon UX
6. Lists and folks on Mastodon
7. How many individuals in my Mastodon feed additionally tweeted at this time?
8. Occasion-qualified Mastodon URLs
9. Mastodon relationship graphs
10. Working with Mastodon lists
11. Photos thought of dangerous (typically)
12. Mapping the broader fediverse
13. Protocols, APIs, and conventions
14. Information within the fediverse
15. Mapping individuals and tags in Mastodon
16. Visualizing Mastodon server moderation