@spgreenhalgh, Are you checking for comments manually only?

I've seen a few doing this pattern before:<br /> - https://boffosocko.com/2021/03/25/hypothes-is-as-a-comment-system-receiving-mentions-and-notifications-for-your-website/<br /> - https://boffosocko.com/2020/05/26/55771462/

I use an RSS feed through IFTTT to trigger email notifications to help out and @judell's facet tool can be useful: https://jonudell.info/h/facet/?max=10000&exactTagSearch=true&expanded=true&addQuoteContext=true&wildcard_uri=https%3A%2F%2Fspencergreenhalgh.com%2F*

@pratik Mastodon caps at 500 as a default. The information density of the particular language/character set is certainly part of the calculus.

Here's a few to start (and see their related references): - https://www.semanticscholar.org/paper/How-Constraints-Affect-Content%3A-The-Case-of-Switch-Gligoric-Anderson/de77e2b6abae20a728d472744557d722499efef5 - https://www.nature.com/articles/s41599-019-0280-3

Another potential method (or an additional filter) for finding posts with photos, or more specifically posts whose main purpose is a photo or image is to use use the post type discovery algorithm. Given that Micro.blog is built on a variety of IndieWeb building blocks, most photos could/should have a class of u-photo on their img tags, so you could search for these instead or in addition to. I believe there are a set of parsers and tools out there that do this in a few languages already and someone in the IndieWeb Dev chat can direct you to them if they’re not linked to the page above.

https://collect.readwriterespond.com/antennapod/

I feel your pain here Aaron.

Perhaps it helps, perhaps not, but I've been using AntennaPod for a few years now. In particular I love it on Android because I can use the share functionality to share to a custom email address which posts to reading.am for an account that aggregates everything I'm listening to. Then I port the RSS feed of that back into my site. It's a stupid amount of manual work, but it mostly works.

Alternately you could share material you listen to to Huffduffer and pull data out that way as well. My problem here is that Huffduffer is more of a bookmark service than a "listened to this" sort of service, though you could always add a "listened" tag to the things you've heard in the past.

The tougher part of all this is that podcasts have "canonical" links for the podcast episodes (sometimes) and an entirely different link for the audio file which has no meta data attached to it (presuming you can even find the URL for the audio file to begin with.)

AntennaPod allows you to pick and choose what you want to share, so usually I default to the audio file to get that in to the workflow and finding/adding the data for the particular episode is a bit easier.

I will say that this is one of the ugliest and most labor intensive workflows I've got for social posts, so I'm usually only doing it and posting publicly for things that I really think are worth the time that make for interesting notes/observations that go along with the post.

I'm curious to see what others come up with for this workflow.

My general heirarchy is that bookmarks are things I want to come back to (and usually read) later, reads are things that I've read, like are things I've read and want to send appreciation for, and replies are things that usually are both read, liked, and needed even a bit more.

Here's more on how I've thought about it before: https://boffosocko.com/2018/03/10/thoughts-on-linkblogs-bookmarks-reads-likes-favorites-follows-and-related-links/