First, available technical features don't impact community aspects. It's something to be aware of, and influenced my own decision to join toot.cat, though the fact that the admin is someone I've known online for over a decade played a larger role.

Secondly, the issues around moderation also matter, but:

- Blocks are reasonably rare, but generally implemented with good reason. Well-administered instances which play well with others tend very much not to get blocked.

- Increasingly, moderators are sharing blocklists through Fediblock (<https://fediverse.wiki/wiki/Fediblock>) or similar mechanisms. So if a block exists on one well-administered instance, it's likely to exist on numerous others.

- Blocked instances typically have a small number of members, though there are a few exceptions. (Large Japanese instances on which Loli features predominantly is the principle exception of which I'm aware, or alt-right sites such as Gab and Truth Social.)

- Instance moderation is on a continuum. "Limited" simply means there is limited federation, but that content won't appear on the Federated timeline, and that follow requests require an explicit approval. (I see this largely for profiles on infosec.exchange.) That's a relatively low bump.

- Toot.cat does moderate instances extensively, with 476 moderated instances listed at <https://toot.cat/about>. 138 of those are "Limited" (see above), 338 are suspended, meaning no federation occurs. The top 20 moderation rationales are:

  50  Reason not available
  18  Freeze peach instance
  17  Transphobic admin
  11  Admin using a wide variety of slurs
  10  Admin using racist slurs
  10  Covid denialism
   8  Edgelord admin
   8  Racist admin
   6  Lamp instance(s)
   5  Admin shitposting to #fediblock
   5  Anti-moderation admin
   5  Cryptocurrency-oriented instance
   5  Harassment, slurs, fascism, etc.
   5  Transphobic harassment from admin
   4  Adming spamming #fediblock
   4  Anti-LGBTQ+ admin
   4  Conspiracy theory
   4  Covid denialism and other disinfo
   3  Admin using ableist slurs
   3  Anti-moderation instance

"Freeze peach" is often code for limited or no moderation, and is associated with pronounced and unaddressed abuse. "Lamp" is a specific abusive user, see: <https://mew.toot.cat/mw/Usr/lamp>.

I haven't (yet) checked to see how many profiles exist or are active (w/in the past month) on the blocked instances, but I strongly suspect that most are quite small, or no longer active. Just for grins, I'm looking at a random selection of 10 instances blocked by toot.cat ('sort -R | head -10' off the list of blocked instances):

  neckbbeard.xyz: Server not found
  hermitmountain.top:  Timeout on Web access
  dobbs.town:  Timeout on Web access
  ortodogs.info:  Unable to connect
  touha.me:  Single-user
  nnia.space:  282 active users (blocked as pedophile instance)
  gamers.exposed:  Appears inactive?  Video site?
  neenster.org:  Unknown (users not reported).  Principally Nina Paly's personal instance as I understand.
  cybercriminal.eu:  Timeout on Web access
  mastinator.com:  Read-only implementation, presumably no users.

So that's 2 actives out of ten listings, with the largest having 282 members (small by Fediverse standards).

> First, available technical features don't impact community aspects.

That's a naive take, but I think it's worth moving on to something much more important...

> Blocks are reasonably rare, but generally implemented with good reason. Well-administered instances which play well with others tend very much not to get blocked...

I notice that you keep using weasel words like "reasonably" and "generally" and "tend" and "typically" with casually dismissive motte-and-bailey appeals to "well-administered" and "good reason" and "small number of members". You do not appear to be engaging in good faith.

Please address the fact from your own statements that your participation globally depends on whether another instance administrator likes your instance's administrator and explain why you believe that's not an issue relating to separation of communities and server administration policies.

> The top 20 moderation rationales are:

> 50 Reason not available

> 17 [Someone who isn't you did something that someone else didn't like]

> 11 [Someone who isn't you did something that someone else didn't like]

> 10 [Someone who isn't you did something that someone else didn't like]

> 8 [Someone who isn't you did something that someone else didn't like]

> [etc.]

I hope you understand the absurdity here of denying the significance of binding identity provision to distributed administrative policies.

The starting point of this discussion was spankalee's statement that "Making your Mastodon account provider also be a community is a fatal flaw of Mastodon". I pointed out that this was largely not the case.

The Fediverse presently has something north of 14 million registered accounts: <https://mastodon.social/@mastodonusercount/11110302515627643...>

I'd responded with what factors are distinguished between specific instances, noting specifically which of those affect broader community interactions and which don't. That's based on my own participation on Mastodon since 2016-7, across a number of different instances.

For the rest, I'm describing both in narrative and using actual instance and sampled data what the impacts of moderation are. You don't seem to like either my use of data, or how I narrate it, but I can in fact assure you that both correspond strongly to my experience over the past seven years. In particular. I'll note that the moderation rationales I'm reporting come directly from toot.cat, and that though I'm friends with the admin I'm not a moderator there myself. You're the one who's editorialised those rationales.

The impact of blocks is best measured by how many active Fediverse profiles are affected by them. That's something which instance block counts addresses poorly. My sampled check helps validate my own gut impression, again, that most blocks are against small, and often inactive instances. You're ... completely ignoring this fact in your umbrage.

I'd also make the argument that harassing people for their backgrounds, beliefs, and lifestyles which are in no way harmful to others is highly disruptive. You may choose to deny this, but that's not well supported by extensive experience and research.

A key function and attraction of the platform is that individual instances can make their own individual federation options. And yet, even with extensive moderation the total impacts seem fairly minor. I'll concede that current mechanisms make this less fully transparent than it might be, but the information is available, and in general issues around instance-level moderation is an ongoing area of Mastodon and Fediverse development. But if you're so opposed to such a fundamental aspect of the Fediverse, you'll probably not find the overall concept particularly attractive.

I'm interested in question of the magnitude of blocking in terms of affected profiles, by the way, and am looking for better ways to get a broader read of how widely blocks are used on different instances, and how many active Fediverse members are restricted by such blocks. Right now, that involves my manually looking up individual instances in a GUI browser session, though I believe there may be APIs which can help with that.

Just for grins, I'm looking at infosec.exchange, itself moderated ("Limited") by toot.cat (on the basis i.e. hosts law enforcement and intel agency accounts). Jerry Bell doesn't strike me as especially socially sensitive.

Infosec.exchange moderates 145 instances, 137 are Suspended, 8 are Limited.

Reasons (complete list):

  114  Reason not available
   3  Hate speech
   3  Unmoderated racism
   2  Part of free/hate speech network
   2  Violent extremism
   1  Admin uses hate speech
   1  Almost exclusively an instance for posting porn.
   1  An instance dedicated to bots that autopost animated/anime porn. Pretty massive draw on resources at best. Going to get someone fired for what’s in the timeline at worst.
   1  At owner instance request
   1  Blocked to resolve confusing bleed over due to Glaceon blocking infosec.exchange
   1  Bypasses blocks and makes all toots public without consent
   1  Generally awful free speech instance supporting racial and homophobic hate
   1  Highly offensive and violent hate speech instance
   1  Highly offensive content
   1  Hosts far right content, offensive racial memes, hate speech, and clearly attempting to out MAGA truthsocial.com
   1  Mutual silence in place
   1  Racism promoted by admin
   1  Should be obvious
   1  Significant source of hate speech, racial slurs, and targeted harassment.
   1  Supports/tolerates hate speech.
   1  The instance appears to be up to no good. Block in place due to shenanigans from a user on that instance, will reassess.
   1  Twitter scraper creating account mirrors
   1  Under review
   1  Unmoderated homophobia
   1  Unmoderated racism and hate speech
   1  Unmoderated spam and other nonsense

Again, this on an instance which is not particularly following the Left's social Zeitgeist.

One of the "Limited" moderations is on toot.cat, as a mutual silencing.

I've not compared the T.C. and I.E. lists for shared blocks, though I see a few.

One of the largest Mastodon instances is mastodon.social (operated by Mastodon's author, Eugen Rochko). Moderated: 235, Suspended: 220, Limited: 15, rationales:

  68  Third-party bots
  64  Hate speech
  43  Inappropriate content
  23  Harassment
  18  Spam
   4  Misinformation
   4  Transphobia
   3  MAPs
   2  Reason not available
   1  Bot that announces blocks
   1  Botnet
   1  Conspiracy theories
   1  Mirroring other Mastodon accounts without permission
   1  Misinformation and conspiracy theories
   1  Preparing for announced shutdown

For those three instances, unique / shared blocks are:

582 1 (blocked by only one instance) 77 2 (blocked by two) 40 3 (blocked by all three)

Finally, I also recommend moderator dang's HN guidance: <https://news.ycombinator.com/item?id=37366825>

One other point that I've been trying to ignore, but have decided I can't:

I notice that you keep using weasel words like "reasonably" and "generally" and "tend" and "typically"...

Yeah, well, if I'd used absolute language, the obvious attack would be "Hah! I've found one example among the ~25k Fediverse servers which fails to follow the moderation practices you describe!" Which is to say, of course, that with roughly 25,000 individually-administered servers, of course there is variation in moderation practices. Again, this is a key feature of decentralised networks.

But again: in most cases the factors which call for, erm, moderation in moderation mean that excessively broad-brush actions tend not to be undertaken by larger, or even modestly-sized, instances, and that the total affected populations are relatively small. That poorly-administered instances face backlash (or departure) by their own affected members is another element to this.

That said, there are general tools for moderation, which are discussed in Mastodon's technical documentation, see: <https://docs.joinmastodon.org/admin/moderation/>

Those tools are there for a reason, their design is inherent to Mastodon's function (other Fediverse apps and servers have similar capabilities), and as with all regulation, tools which are effective when manipulated with skill and wisdom are of necessity dangerous when used carelessly or maliciously. That applies to every tool starting with the first scratch-stick and hammer-rock.

There are also recommended guidelines and practices, though much of that remains somewhat informal. Mastodon's creator, Eugen Rochko, has written on the topic however: <https://blog.joinmastodon.org/2018/07/cage-the-mastodon/>. Though broad, and focusing on technical details, Rochko details why features such as global default search ( People and brands search for their own name to self-insert into conversations they were not invited to") and quote-reply (Heed, my followers, how I dunk on this fool!). That is, there is some direction for how well-administered instances are expected to operate. The server rules of Rochko's own instance can also serve as a rough outline:

1. Sexually explicit or violent media must be marked as sensitive when posting

2. No racism, sexism, homophobia, transphobia, xenophobia, or casteism

3. No incitement of violence or promotion of violent ideologies

4. No harassment, dogpiling or doxxing of other users

5. Do not share intentionally false or misleading information

<https://mastodon.social/about>

Toot.cat posts a code of conduct, generally restricting: Personal attacks, Trolling or insulting/derogatory comments, Public or private oppressive language or actions, Other unethical conduct, red flags (detailed in a link). Again I'll note that the admin someone I've known for a long time, and she's run, moderated, or helped design community criteria for a wide range of diverse communities over decades, with infinitely greater patience and tolerance than I'd show myself.

<https://mew.toot.cat/mw/Toot.cat/CoC>

There are also freestanding guides to moderation (this is, again, a decentralised federated universe), such as "A Guide to Moderating in the Fediverse: The Basics" <https://tenforward.blog/a-guide-to-moderating-in-the-fediver...>. And, I'm sure, numerous others.

As to the predictable, and extant, objection that those rules are vague and subjective, I'll note that this is similarly the case on the forum we're discussing this matter on, Hacker News. The guidelines are kept intentionally brief, and much of the actual guiding lore is to be found in the many, many, many moderation comments by dang (> 47k comments), sctb (~6k), and pg (~10k) over the years. (Totals are all comments, some concern matters other than moderation.) This is by design. As dang writes:

Here's a brief refresher: Curious conversation is good. Substantive comments are good. Thoughtfully sharing personal experience is good. Flamebait is bad. Personal swipes are bad. Ideological boilerplate is bad.

Those are excellent ideals for virtually any healthy community in my view.

<https://news.ycombinator.com/item?id=26613795>

More by dang on HN's brief guidelines: <https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...>

And no, I don't always agree with dang. That ... would be obsequious, pandering, and utterly unimaginative. But I very frequently do, and even where I disagree, can understand why he's moderating as he is.

See, e.g., <https://news.ycombinator.com/item?id=37566315> <https://news.ycombinator.com/item?id=37258356> <https://news.ycombinator.com/item?id=29493683>.

And finally: I really don't expect that BugsJustFindMe will be persuaded. I feel rather like the narrator of Tim Minchin's exquisite beat poem "Storm", and we'd all as well be back nine minutes in time. But I'm hoping that other readers might find reason and fair claim here, and reach an understanding and appreciation. As much as my faith in persuasion through rational argument has taken a profound hit over the past decade or so, that hope remains live for me.