Thread about some differences and limitations that might surprise or trip up folks trying out #Mastodon. I'm no expert, just sharing info gleaned from reading up a bit and playing with it for a few days - hopefully can shorten learning curve for a few folk on the #TwitterExodus
In short, if you are hoping for a drop-in, like-for-like replacement that works just like Twitter, it's not that, despite the superficial similarity. Most important, the way posts propagate, and who may be able to find them, is totally different to what twitter users are used to.
Your instance (the server you joined) hosts toots from you and its other members. But you can follow people on other instances, and they can follow you, and you will see each others' posts. Hence the suggestion "it doesn't matter which instance you join". Well, only sort of...
The way that works is federation. When you follow someone on a different instance, your instance basically gets a copy of their posts. They will show up in your feed. Posts from people who are followed by *other* users of your instance will also be available to you.
But if someone on another instance isn't being followed by anyone on your instance, they almost might as well not exist for you. You won't ever see their content or be informed that they exist (although if you know their handle you can type it in and follow them).
As on twitter, you can discover people by looking at other people's follows/follower lists. However these are often very incomplete - not sure why, perhaps some instances do not export their members' follow lists and you only get to see the ones who happen to be on your instance?
And the federated propagation limitation applies to search, too. When you search for a hashtag it will only search the posts that originated on, or have been federated to, your instance.
There is no global, exhaustive, search, and there really can't be, because no entity ever holds all of the messages. Another limitation that may grate is searches only work on hashtags - not the whole text. So if you want your posts to be discoverable, use hashtags well.
But even when you use hashtags well, the person searching for that hashtag may not find your message, depending on the vagaries of federation - has your message reached their instance? Quite likely not!
None of the above is intended to disparage Mastodon. It's a different beast to Twitter by design, and has many fine virtues, probably the foremost of which is that no one person or entity owns or controls it, so it cannot be bought out or taken over.
Here is another surprise. If you follow someone on another instance, you will not get access to their previous posts, unless it so happens that they were already being federated to your instance. Your instance will start to collect their new activity but will never go back...
... and backfill their historical material. So the available history of activity of a remote user depends on when they were first followed by a user on your instance. Someone following them on another instance may see more, or less. People will find that very random and confusing
I posted an introduction toot, full of searchable hashtags (as recommended), but that post will only show up in searches made on instances that were already federating my content in at the time I posted it. As that post was one of the first things I did, that was likely few!
So if you see an introduction toot by someone you like, boosting it will help to get it to spread
I found this paper on #Mastodon interesting. You don't need to follow the technical details to get some useful nuggets about consequences - good and bad - of the federated architecture. Disclaimer: 2-year-old paper, things might have changed. umm-csci.github.io/senior-seminar…
• • •
Missing some Tweet in this thread? You can try to
force a refresh
It's that map again. A short thread.
I am seeing this map being posted by leave supporters in response to the news of Vote Leave dropping its appeal. The argument appears to be: any swing introduced by illegality of cheating could not been enough to affect the outcome. 1/10
When the referendum result was reported, I was dismayed by this map. We had a referendum, a simple count across the whole voting electorate. The map was in effect re-interpreting the numbers, after the fact, as if it was a first-past-the-post election. 2/10
Everyone knows why FPTP is disliked. Votes cast for the losing side(s) in any constituency are effectively ignored. In the case of the referendum, they were of course not ignored. But this map contrives to do that, and has fed a narrative of "overwhelming majority want out". 3/10