Re:The Follower Factory
by DonHopkins on Saturday January 27, 2018 @02:30PM
(#16247674)
Thank you for those links, and your great work. If Twitter refuses to do the kind of investigation and analysis that you performed because they're making too much money from bots, then the free press and open source community needs to take up the slack!
Hey Rich! The graphics are on point for this article. I really liked the unfading faces at the start. Also, the times should put a tip jar on these kinds of articles, I'm not an avid reader to subscribe, but would want to appreciate any good articles I find.
On the graphics, it looks like the density changes for the rest of the dates whenever the identified bot families show up (creating the differently colored columns). When the identified bot families stop, the density changes back. Is that an artifact of how you're displaying it, or is it an indication that the followers before/after are also bots, but they've been mixed with bots created at other dates, where the bot families identified are just the low quality (non-mixed) bot families?
[dead]
Re:The Follower Factory
by DonHopkins on Saturday January 27, 2018 @03:02PM
(#16247825)
You gotta at least give Rich credit for contributing critical bug fixes back to open source tools he uses like Svelte, even when he's under a deadline!
https://twitter.com/Rich_Harris/status/956961714578317312
https://github.com/sveltejs/svelte/pull/1137
It seems like Svelte is worth taking a good look at!
I've been shopping around for a "magical disappearing UI framework" to re-implement my old jQuery pie menu component. I was considering Polymer because it was minimal and followed the latest web standards for components, shadow dom, events, etc. How does Svelte compare and interoperate with Polymer?
http://www.donhopkins.com/mediawiki/index.php/JQuery_Pie_Men...
We'd like to eventually open source some of the stuff we built to do the analysis as well, though it depends on time and priorities.