Got a new manager at work. Caught myself special-interest-info-dumping about Douglas Engelbart again during our first one-on-one meeting.
Managed to get infinite scroll and search (almost) working on this site generated from a few of my Mastodon exports. Kinda neat for a fully static site. (Relative vs absolute paths are a pain, though)
This also reminds me that I should grab another export from around here, since my last backup was from January. But, I also want to play with fetching outbox endpoint pages to catch up incrementally
Everyone quotes "information wants to be free" but they never seem to include "it also wants to be expensive"
We're in the "wants to be expensive" phase right now
@corbden @Are0h @thegibson FWIW, with the streaming API on Mastodon instances, you can build a personal algorithm against the toot firehose.
That said, it's not exactly simple. Would be neat if fediverse servers and/or clients supported sharable plugins for personal feed algorithms
A thing that maybe folks don't realize about ActivityPub & the fediverse:
ActivityPub is for publishing your activities. It's mainly designed to disseminate and spread things around. It's not really designed to keep things secret or compartmentalized.
There are copies of your stuff in lots of places. Your server purposefully sends your stuff to other servers. Even "direct messages" are just sort of a friendly agreement between servers and admins, not a private channel.
If you don't want what you say to end up in front of many eyes, don't use this network to say it.
I'd think that these things are obvious, but then I see folks expressing surprise that things they said turn up in far-flung databases.
Your name goes through at least, like, 17 computers every day
@ana Oh, that looks perty!
Another dumb project idea:
A thing that connects to a Mastodon instance's streaming API and runs pluggable algorithms implemented in WASM against the firehose to generate personal custom feeds.
Maybe WASM could help keep the algos sandboxed well enough to be relatively safely shared around?
@george What do you mean if we hold our club meeting in the middle of a downtown sidewalk that random people can wander through?
Maybe elon musk should have tried running a mastodon instance first before buying twitter
@reconbot Yeah, I'm thinking more about algos applied over an extended period where a server / long-running app might work better.
Like, give me the top links shared in the past day. Or give me just the folks I follow who tooted least in the last week that I'd otherwise miss
Hmm, thinking I might try switching over to tinysearch for my little Mastodon export site generator.
Pagefind works really well for finding entire pages (duh) but not-so-well for indexing individual elements within a page.
Since I'm generating per-day pages full of multiple posts, that makes Pagefind slightly less handy
So, I think GitHub Pages isn't super happy with me pushing a 650MB website at it repeatedly. But, I think I found the right baseUrl option to get Pagefind to offer up the right links
Lots of fiddling that could be done with the HTML templates, but I think I got a website of toots here
twitter
So I peeked in on Twitter for the first time in awhile.
Uhhh, no thank you.
(can't think of what i did to signal interest in this person other than maybe gawk back in 2018 when she handcuffed herself to the front door of Twitter HQ)
Part of me respects the optimism represented by this "Send to Phone" button offered by a YouTube ad playing on my TV.
Like, some group of people at some point in a conference room reviewed some UI design mock-up with this button sketched out, and they apparently decided "yeah, people will enjoy willingly sending an advertisement from their television to their smartphone, this is a worthwhile feature to build."
