Leaving twitter

I've spent most of my spare time posting farewells to the people I was following on twitter. I have now more following on mastodon than there, despite starting not far from 400 followed people. The place has become aggressive to creative people with the TOS update of November 15th.

I created myself a bluesky account too, because it's clear most of the gamedev and pixelart people will / have migrated there. And a ko-fi supporter account for those who haven't migrated and possibly won't.

I'm glad I had decided to reduce my implication on twitter. I'm glad I had captured most of the key interactions already in 2022 when twitter became unreadable for non-subscribers, although I still have many posts linking to items trapped in twitter.

Because the harshest discovery of the week was that twitter archive does not feature any of the replies, likes and everything you *received* from others. Only what *you* did. That means any conversation is lost and any thing I could have said in reaction to someone else's post is missing key context. With over 14000 tweets in the archive and no more public access to the platform's content, automating recovery is not an option.

So I'll do my best to recover what I can manually, with the help of blogger archive. I'll post recovered things under the fromTwitter. I'll have some more key replies cross-posted as blogger comments (apologies to comments authors if they feel offended that I took that freedom. You have the right to make me remove them if you want to, of course). When possible, I'll augment the link to twitter threads with an abbr of what was said in the post.

And Mastodon ? Well, it will be about the same: the exported data has an 'outbox.json' and a 'media_attachments' folder, but that's it (apart for likes, avatar and bookmarks). inReplyTo entries will list some other's status entries (actually an URL with https://instance.domain/users/userid/statuses/tootid), but that URL is nowhere to be found in the .json file. We *can* get some stuff out of it though, like calling /api/v1/statuses/tootid/context, which gives json arrays ancestors and descendants, the second having all the replies. They have the initiator type xhr in firefox network inspector. And the thing I replied to is in /api/v1/statuses/tootid. That should be scriptable enough.