honestly for what i use social media for it could probably be handled with some shell scripts and a public atom feed
yes i know this is just OStatus
what i mean though is i don't actually like social media that much for having actual conversations, and for status updates you can get away with something much simpler
yes there is still the hypothetical problem of “status updates for a limited audience” but in practice i don't find that to be that pressing; just send an email or chat message if you need someone to know you're doing something but don't want the whole world to know
beyond that i think the exclusivity of the limited audience leads to a strengthening of the parasocial aspects like
yes you are learning things about a person that not everybody knows but that still doesn't mean you like, know them at all as a person, and it especially doesn't mean they know you
the major advantage mastodon gives me from a publishing perspective is just that it has an app i can use from my phone
but i wonder if i could automate some git stuff with Shortcuts to make pushing a small text file of a status to a repo just as simple
i did make this shortcut and i'm working on the post-receive hook and some minimal html files to display the thing but it will probably be done tomorrow
python HTML processing is bad so the only markup i will support is <example://link> or <example://link>="link text"
the url will be status.kibi.exposed
i'm now wondering if in addition to normal timestamp-based statuses it would be useful to also have tag+uuid7-based statuses for things like notes or snippets which are not so ephemeral; that might be future work
or tag+SHA1; shortcuts can generate those and it would guard against adding the same thing twice
this is all set up on my end i’m just waiting for my girlfriend to figure out the routing for the server; i gave her a somewhat complex list of demands
the post-receive hook is written in python and i hated every minute of it
but it will be low-maintenance and stable, and it has zero dependencies beyond the standard library
for something which needs to interact with the filesystem and output both xml and json that’s pretty good
working copy has shortcuts support; i bet i could real easily and just have a CI pipeline to turn those text files into structured data