jamietanna[m]aaronpk: looking at the sign in with apple example, I see the comment `Save your private key from Apple in a file` does that mean they provide both the public and the private to a dev, rather than going through certificate signing process?
[Ian_Forrester] and [kevinmarks786] joined the channel
[tantek]my point is if the user *chooses* to publish an email address (e.g. already does so, plenty of folks do), *then* it can be used for IndieAuth via Apple sign-in
[grantcodes]Alrighty, pretty close to having my microsub middleware idea working. I got it to pull posts from channels, parse them and then create its own feed. Was a more complicated flow than I thought. Now I'll see if it actually works
aaronpkjamietanna[m]: yes it's not the same as an ssl cert, it's just a private key and they use this same flow for most of their authentication and signing stuff elsewhere in the platform
[eddie]Ohhhhh, Core ML (Machine Learning) 3 that Apple is rolling out, will be able to use on-device machine learning so you can have custom algorithms that stay on the device you control. One I have the basics built into Remark, I think i’m going to have to experiment with that and see if I can get Remark to help provide some on-device channel (in addition to a Microsubs regular channels) that helps surface stuff from your various channels that you
aaronpkso you're trusting that they're not storing it, but that seems like not a huge risk given that the only thing the key is good for is talking to Apple
[tantek]I have a long delayed post about rel=me updates from the past year (now 1.5 yrs) that I clearly should have already posted since now there's a need to reference what's new with rel=me
aaronpk[fluffy] part of the reason that Atom/RSS looks "easier" to parse than mf2 is because it's much more limited in what can be included in the feed. and even then, go look at any mature parsing library and you'll see it's full of quite a mess of stuff to handle all the variations
[jgmac1106][fluffy] as someone new to much of this stuff I was amazed that all I had to do was put mr2 in HTML and it just worked as a blog in my reader
[fluffy]if you subscribe to the h-feed on the summary page you only get the summaries. In my atom feed I can provide full content and also have a feed-specific layout for things like images etc.
[fluffy]and I also feel like the atom answer of “just include your images/video/etc. as inline content to the HTML” works fine for the cases I’ve considered.
aaronpkmost of the hard part of consuming microformats isn't actually related to HTML at all. it's handling the variations in where things like the author appear in the data structure.
aaronpkthis is a double edged sword, because while it makes it a little harder for consumers, it offers much more flexibility for publishers. and since there are far more publishers than consuming code, it makes sense to optimize for that.
[fluffy]One of my long-term projects is to add authentication/friends-only stuff into feeds and while I think I have an okay idea of how to go forward (which we’ve talked about a bunch in the past, actually your responses to my original E2EE proposal are how I started following the whole indieweb scene in the first place!) I’ve definitely focused on the Atom side of things, and I’m not entirely sure how it might work in an mf2 world.
[fluffy]or in the case of just doing a single thing to indicate that there’s potential auth content, that would just be a stub entry that appears at the top of the screen and does double-duty.
[kevinmarks786]Atom is superficially easier to parse but in practice feedparser covers a multitude of sins. But h-feed does have the summary/content distinction that atom has already. You could hide the full content part with css (or details/summary)
[Rose], [eddie], gRegorLove_ and [grantcodes] joined the channel
Loqih-feed is a microformats2 draft specification with a top level feed object to contain h-entry posts and optionally a common author https://indieweb.org/h-feed
gRegorLove_`u-url u-uid` (canonical url) being different than the feed URL could be a signal, but even then, some h-feeds (mine) include full content, so the extra fetch isn't necessary