#[pfefferle]the question was: how can I provide a list of different feeds (articles, notes, replys, …) and maybe how to prioritize or categorize them, that a feed reader can handle them proper…
#[tantek]When you follow someone, you should be able to both pick up front what you want to see from them (maybe default to just their default stream), and later be able to mute particular things
#[tantek]And none of that should depend on how many gazillion variants of feeds they publish because that’s unsustainable both for publishers (high publisher expectations always fail), and no one actually wants a “feed management per person” UI (except maybe OPML fans and stalkers?)
#[tantek]There are likely some lessons to be learned from existing silo UIs related to limiting the stuff you see from those you follow
#[pfefferle]because of POSSE, many blogs/sites became very noisy and I often have to search for the correct feed to subscribe, that does not include for example POSSEs…
#Loqimute is the ability to hide posts in your reader that have specific words, or from particular publishers, optionally with an automatic expiration, after which newer posts are visible again https://indieweb.org/mute
#[pfefferle]and they does not always exclude them from their “default” feeds
#aaronpkright, the fact that that "breaks" existing models of blog feed readers isn't a problem of their personal websites, it's a problem with the legacy feed reader model
#[tantek]Why should their default feeds show anything less than everything they personally author?!?
#aaronpkI want my default feed to show specifically chosen things and those things are most of the time not just blog posts
#[tantek]Again that kind “they don’t exclude” puts the burden on the wrong person
#[pfefferle]because it might force people (like me) with “legacy” feed readers to unsibscribe
#[tantek]People should be encouraged to publish (including to feeds) everything they want to publish
#[tantek]Certainly not asked or expected to “exclude” anything
#aaronpkExactly, and the reader software should adapt to that
#[pfefferle]I agree, but there is a transition phase
#[tantek]Silos are already solving this problem so social readers should too
#[pfefferle]and in that phase I have a lot of noise in me reader
#aaronpkThe hen/egg problem describes a situation where both sides require work for something to change. In this case it's only the reader side that needs to change
#ZegnatThey just require potential subscribers to put in the effort to write a reading application? :P I think [pfefferle] is saying that it is valid to say that it should not be the publisher, but at the same time it should not be reader.
#[tantek]Sounds like feature requests need to be filed on those readers then
#[pfefferle]yes, but I as a reader have a problem then
#[tantek]No Zegnat, to file feature requests for things like “mute”
#ZegnatAnd until then they should just not subscribe to my site? I think that is the issue at hand here. What is the transition period
#[tantek][pfefferle] this is part of the larger pattern of legacy readers have stubbornly fallen behind modern social media reading UIs
#[pfefferle]but I fear that I might crawl my current readers (that use legacy readers) in the meantime
#[pfefferle]so it is a problem for me as a publisher!
#[pfefferle][Zegnat] yes… at the end it might be a “problem” for both
#sknebelI think having options for filtered feeds for such reader applications makes sense, but the additional markup to "categorize" such feeds would also need implementation work on the reader application side, and I wonder if that wouldn't be better invested in filter/mute features
#[tantek]Also there is also room for documenting best practices for what to put in your default feed as a publisher
#[tantek]Both aaronpk and myself for example are very deliberate (differently) about what goes in our default feeds (which btw is definitely not “everything” 😂)
#[tantek]Eg my home page stream shows all my posts including likes and replies, whereas my default atom feed only shows my original posts
#[tantek]also due to the “dumbness” of legacy feed readers and their poor handling of overcaching / failing to update, I have a delay before anything I publish makes it into any legacy feed files
#sknebelmight be worth collecting what readers offer right now filter wise, both traditional ones and the ones from our circles
#sknebel(e.g. I don't know of the top of my head what kind of filtering Aperture has either)
#[tantek]Definitely. Worth adding such examples to /mute
#[pfefferle]hmmm, but then I am (for now) deciding what the reader sees and what not (at least if he subscribe to the “main” feed)
#sknebelnothing wrong with having multiple feeds IMHO, just not sure I need more than a human-readable description on them
#GWG I would like to markup my main and secondary feeds differently
#aaronpkAperture has some filtering but it's on the channel as a whole not per feed
#sknebelyeah. for exposing more meta-data about a feed I think some experimentation would be needed what's actually useful
#sknebele.g. maybe offering update frequency or sth is helpful, but I wouldn't start marking that up without having something to show its useful
#sknebelthis could very well also be helpful for new reader apps. I guess apps could trial it by fetching a bunch of a feed and compiling the information themselves - if it turns out to be helpful marking it up is an optimization
#sknebel[pfefferle]: I very much feel the pain of these two worlds too ;)
#[tantek]mute << silo example: Twitter has an option to “not show retweets” from someone you follow (needs screenshot). This is essentially a “mute retweets” feature even though they don’t call it that.
#Loqiok, I added "silo example: Twitter has an option to “not show retweets” from someone you follow (needs screenshot). This is essentially a “mute retweets” feature even though they don’t call it that." to the "See Also" section of /mutehttps://indieweb.org/wiki/index.php?diff=73673&oldid=71773
#aaronpkhrm i think i'm gonna make some breaking changes to my indieauth client library. it's gotten kind of messy
#vilhalmerI'm three revisions behind now, I guess I know what I'm working on next
#aaronpkHopefully these changes mean you'll be able to remove some code
jeremycherfas, [Raphael_Luckom], geoffo, nolith and reed joined the channel
#aaronpkside effect of deleting all my redirect handling code is my client now provides whatever the user entered as the "me" value to the authorization endpoint
#aaronpkso if i enter "aaronpk.com" into a client i'm taken to my website with that as the "me" rather than previously the client expanded it to "https://aaronparecki.com/"
#aaronpkeverything works fine for me because my website is a single user website and was just ignoring that parameter anyway
#aaronpkso i took out all the code that i had to keep track of the URLs found in the redirect chain, so i'm only doing the optimization of "if the entered URL and profile URL are an exact match then skip verifying the authorization endpoint"
#GWGI need to get back to my 1.1 update PR for WordPress
[KevinMarks] joined the channel
#[KevinMarks]is tagging a way to do this feed filtering? Or has that now become balkanized by platform?
#Zegnataaronpk: I think we specifically wrote in the updated spec that the me parameter given by the client to the AS might be as provided by the user, so non canonicalised?
#aaronpk"me - (optional) The URL that the user entered"
#ZegnatOf course the fact that we wrote that in the spec does not mean all AS would work with it. But I would not have expected it the break many implementations either. Last we checked, not a lot of implementations were even using the me provided by the client, they just always returned their own hardcoded value
#aaronpki just wanted to check if it might break things in the wild
#ZegnatSending a me at all there is an extension to OAuth and excluded using generic OAuth libs. Which IIRC is why we watered it down a bit, if that is how we could call it
#aaronpknow... what project should i update to the new indieauth client first
#ZegnatTelegraph? If that one is still expecting form encoded as it used to, it is prime for an update
#aaronpki kinda want to make telegraph do the indieauth flow itself unless there's no authorization endpoint, then use indielogin.com only for the fallback provider options
#Loqicodepen is a silo for hosting HTML, CSS and JavaScript snippets to demonstrate the capabilities of the client-side Web https://indieweb.org/codepen
#@AndreJaenisch@cassiecodes@jh3yy@smashingmag Good to know! Look, I like to host code I write on my site (eventually). I would like to document the process of making and my ideas behind it. The only way I can do that in CodePen is through comments (or a section in the HTML). Then, you cannot see the progress over time. (twitter.com/_/status/1331900348752138240)
#sknebelpetermolnar: your home internet doesn't happen to be A&A? :P
[Raphael_Luckom] joined the channel
#[Raphael_Luckom]it's more complicated that "aws is centralized" Their outage in US-east messed up my workflow yesterday, but my friends down the street are using us-west and didn't have any issue. Both aws
#aaronpkit accepts any request to validate authorization codes and always returns the same user URL
#ZegnatOh, yeah, so, there might be one like that currently in production already ....... :P
#aaronpki'm trying to test the new requirement of verifying the authorizatio nserver
#ZegnatI think Sink works a bit like that. But because it is mainly internal and for debugging purposes anyway, I do not care
#ZegnatMy new AS to replace Selfauth for me does just straight up have a text field where you can tell it what to put in the me. So that one could be used to check implementations as well
#aaronpkyou might want to add a note there that says you should only fill in a URL there that links to this as its authorization endpoint
#ZegnatIt still requires a password, so nobody but me will actually be able to use it. But I liked the idea of being able to have the URL be an open text field.
#aaronpkwoohoo telegraph is now doing the indieauth flow itself and only uses indielogin.com if there's no authoriziation server!
#[tantek]mute << silo example: Instagram: on the iOS app when viewing someone’s profile that you’re following, you can tap on the [ Following v ] button which reveals a slide up menu from the bottom of the screen which has a "Mute >" item, tapping it slides left two sliders (off by default) for "Posts" and "Stories", which you can set independently. (screenshots needed)
#Loqiok, I added "silo example: Instagram: on the iOS app when viewing someone’s profile that you’re following, you can tap on the [ Following v ] button which reveals a slide up menu from the bottom of the screen which has a "Mute >" item, tapping it slides left two sliders (off by default) for "Posts" and "Stories", which you can set independently. (screenshots needed)" to the "See Also" section of /mutehttps://indieweb.org/wiki/index.php?diff=73678&oldid=73673
#ZegnatStraight out of a Mcsweeney article, feels like ;) But a good one none the less, lots of nerdy and interesting links into my reader
#[tantek]alright for those of you interested in the feed reader filtering conversation, I added a few more mute feature Silo examples for you take a look at and consider in terms of UX: https://indieweb.org/mute#Silo_Examples (note that none of those require the publisher to do anything special)
#[tantek]I'm pretty convinced that's the general flow that will work for lots more people:
#[tantek]1 follow a person (for whatever reasons) to see their posts in your reader
#[tantek]2 mute posts from that you don't care about or feel like are noise
#[tantek]much better than trying to a priori judge from a bunch of random feeds with random names according to what the publisher might have thought when they created them
#[tantek](if the publisher even bothered to create special / separate feeds, which 99%+ don't bother with, and maybe some other small % have automatic "tag" or "Category" feeds that they don't know about)
#sknebelif someone goes through the effort of providing selective feeds for topics replicating that with filters is really tricky I've found, especially when it comes to microblogging/tweets/... and I've tried doing that a bunch
#sknebel(but that's a different problem from just post-types)
#[tantek]it's a lot of work to write/author/publish in a way that automatically slots into those selective feeds, so very few will bother
#[tantek]people can barely be bothered to use hashtags
#[tantek]that's the closest you're going to get to broadly adopted selective feeds (hashtag-based feeds)
#[Raphael_Luckom]I looked at the feed reader page and the social reader page, but I couldn't get a sense of 1) whether there are clear favorites among the readers 2) whether there's much activity in making new readers.
#[Raphael_Luckom]Not saying those belong on the wiki (would probably get stale if it was) but I'm having a hard time following the conversation.
#aaronpkfor example someone can subscribe to my curated feeds or subscribe to /all and filter themselves
#[tantek]overfocusing on an option that only a handful will do is replicating the elitism of the old RSS/Atom ecosystem
#[tantek]so sure, both *can* exist, but don't bother evangelizing the few/elite option as if it will solve the problem for most people
#[tantek]like yes, I myself have an "articles-only feed" (originally built for Planet Mozilla use-case), and have on my to-do to make a "photos-only page" (along with feed for that), but I'm not going to suggest anyone else bother doing so
#[tantek][Raphael_Luckom] there's "legacy readers" which are literally read-only passive readers, old school sidefiles etc., and then there are "social readers", which as the term "social" implies, are two-way *by default* (assuming you sign-in with a Micropub enabled personal site). concentrate your effort on the latter as they are actually evolving their UI to displace social media usage etc. the former are kind of like email UIs except you can't
#sknebelfairly sure I've heard that one before. probably from you I'd guess :D
#[tantek]lol I'm pretty sure I haven't said it before. would have to do a search to see if I can find a variant of someone else saying it
#sknebelI've seriously been tempted to try rig some classification code from a spam filter or something to categorize posts, but on tweets even that fails
#sknebel(don't get me started on people tweeting screenshots of tweets :D)
#LoqiNIPSA is an acronym for Not In Public Site Areas, and an admin feature of Flickr that allows their support staff to mark an account such that posts from it will not be shown in search results and other similar public views of posts https://indieweb.org/NIPSA
#[tantek]^ back in the day, people got NIPSA's for posting only or mostly screenshots to Flickr
#[tantek]so there's some prior art / precedence for that
#sknebelyeah. but really at some point the answer is probably going to be "don't use twitter" or "only follow people that don't tweet much about things that annoy me" instead of "writing more code". (well, maybe my attitude to coding in my free time will improve again at some point ;))
#sknebelbut yes, that's a valid idea for a filter. detect text in images
#aaronpkupdates his website to reflect some more indieauth changes
#sebbui wonder how hard it'ld be to make a wiki plugin for micropub, and allow to edit pages (articles?) from it
#sebbu(and first, you'ld need to know which markdown flavor it uses, with which extensions)
#sknebelsebbu: on the mediawiki API that wouldn't be too hard if you are just thinking editing the native wiki markup (converting mediawiki markup back and forth to other formats is ... "interesting")
#lahackeri got my own markdown implementation to pipe through pandoc to mediawiki for an automated POSSE to indieweb wiki.. works for basic pages, clobbers complicated large pages; zegnat recommended i look into section editing which sounded promising
#sebbuskalnik, the issue is that i'm using dokuwiki
#aaronpklahacker: i usually do the first draft with Quill, bringing in photos and doing basic formatting. i post as a draft to my website and then do some final tweaks there