Donnodubus

@techlife@indieweb.social No it is not "the functional equivalent of not using an algorithm," because it is literally only achieved through the implementation of an algorithm. The notion that an algorithm should only sort according to time, and that chronology is the most "objective" or "neutral" criteria, is actually a totally subjective value judgment. People often equate "algorithm" with data-driven predictive models, but that's a separate concept. @hildabast@mastodon.online @mission_europa@universeodon.com

@hildabast@mastodon.online @mission_europa@universeodon.com I'm going to be pedantic here and point out that systematically sorting the posts, boosts, and replies of people you follow into chronological order IS an "algorithm."

@techlife@indieweb.social My point about algorithms was unrelated to QTs. I just think the canned talking points about "algorithms" fail to grasp what the word actually means, plus the fact that chronological sorting isn't really value neutral like people claim. I feel Mastodon would benefit from adding another timeline sorting method, but it's impossible to start that conversation if everyone is just going, "Meep, meep! Algorithm bad!" 😆

@techlife@indieweb.social Those don't need to be the criteria on which an algorithm operates, though. I'm not in favor of using data collection to predict people's interest in posts whatsoever. Here's a toot from a while back where I summarized my idea of what I think could be a useful alternative for sorting a home feed, just using metadata that I think is already implemented in Mastodon: https://mastodon.social/@donnodubus/109599868495621148 @hildabast@mastodon.online @mission_europa@universeodon.com

@techlife@indieweb.social The definition of an algorithm is a finite sequence of rigorous instructions. No more, no less. Assembling the posts, replies, and boosts of accounts you follow into a chronological sequence is exactly that. And it is not neutral. There are actually several judgments and assumptions that go into that algorithmic model. @hildabast@mastodon.online @mission_europa@universeodon.com

@techlife@indieweb.social To be honest, I don't actually know how Mastodon's code works under the hood. But since users are able to manipulate their timeline by filtering out boosts, for example, it seems to me that the timeline is not simply a naive stream of data. I think it must be processed and organized into a list according to metadata, which would necessarily make it algorithmic. People who say Mastodon has no algorithms are also forget the trending post and hasthag tabs... @hildabast@mastodon.online @mission_europa@universeodon.com