13.76480
i’m not terribly excited about mastodon. though it’s obviously the best existing network of its type, federation creates a worst of both worlds scenario.
note my thoughts here are about centralization and distribution of control of a system, not of the implementation of a system. mostly because any sufficiently large system is going to be technically distributed.
the primary benefit of centralization is the ability for the actor who controls the system to change and improve it relatively quickly and easily. this is a benefit because systems generally do not spring into existance fully formed, they grow and change and accumulate features and undergo restructuring. the drawback of course is that the other actors don’t have control, and if the system changes in a way they don’t like, they can’t do much about it. from a security perspective, this enables both the timely fixing of security bugs all across the system and bakes in a enormous vulnerability. that actor, or anyone who can influnce that actor, can introduce vulnerabilities.
in particular, almost all centralized systems reveal user data to the controllers (or anyone who gains access to the computers used by the controllers) of the system.
on the flipside, the primary benefit of the distribution of control over a system to the users of that system is adaptability. the end users can choose different implementations (or change their implementations) to suit their needs, they can operate multiple independent systems, and they can have some confidence that any private data remains private. of course, this comes at the cost of a great coordination problem. to interoperate with others’ implementations, my implementations needs to be compatible with them. even if half of the users of the system makes the same update, the network as whole may be left mostly operating under the previous standard.
in federated systems, control is distributed to an extent, but not to the end users. the end users can choose between instances and move between instances, with various amounts of difficulty, but unless they start their own instance they only experience a limited amount of control.
so the above is mostly talking about the centralization of control of the technical aspects of the system. but with technical control comes social control.
Twitter is vastly different than Mastodon in that there’s one instance with one moderation policy, so all a user can do over there is complain about that policy if they disagree with it.
Here people actually have some agency: they can choose a different instance with a different policy if they find that the current one is not what they want.
In fairness, I think it is a new thing for most people. It’s sometimes said to be like email, in that people can have different email providers, but we don’t talk about how an email provider should make their moderation policy visible to potential users because we don’t expect moderation there. With public communications, we do need moderation so there’s an important extra piece there that is pretty opaque to new users.
ironicly, because the types.pl instance of mastodon limits the mastodon.social instance due to a lack of moderation
(note that this section of the page is only available to types.pl accounts),
as a types.pl user, i couldn’t see the reply until @ahfrom boosted it.
and that right there is why i’m not terribly excited about mastodon: if i want to have a reasonable experience interacting with various groups, i either need to be on multiple mastodon instances as not to get caught in instance-level politics, or start my own instance and hope instance-level politics don’t come for me.
With public communications, we do need moderation […]
no. what we need is for users to be able to effectively control what content they share, see, and platform.
when i post something here, it is both a public communication and not subject to moderation. i control what i’m sharing. that probably is at least solved, and easy to do so.
when you look at my website, you may not want to read all of it. you can’t know which parts you want to read and which parts you don’t without actually going and doing so. of course, your user agent could be doing this for you, noting that you might not want to read this page, perhaps even (removably) censoring within the page for you.
there are some crude tools for users for this sort of thing. the sophisticated ones are designed to fight spam and aid moderation. in both cases the tools are domain-specific, tied to a single company’s ecosystem, and usually don’t allow a user to say something as simple as “always accept communications from this other user”.
hence tags, fulfilling the dual purpose of linking to other content about the same topics, and informing the user about the content they may or may not want to read. there are obvious shortcomings - most notably that it shifts the responsibility to the creator to anticipate what people may or may not want to see!
now… i don’t have comments on this page, but if i did, you could leave a comment. indeed, you could read a comment that other people - who did want to read the musing - would not want to read. even if readers had the tools to automatically remove such comments from thier view of my musings, i might not want to enable other people finding those comments, which suddenly puts me in the moderator’s chair.
and of course, twitter and mastodon allow replies, and seeing what the replies are to a post, analogous to comments on a blog. yet, i don’t think i’ve seen a moderation technique of allowing each poster to unlist replies to their posts, much less an automated tool that tailors itself to the user’s preferences to take some of that burden.