#gRegorLoveIt's been one of those days where my schedule is all off. Haven't missed anything, thankfully.
#tantekgRegorLove: I hear you. time and even light sensing is altered for me since YxYY this past weekend
KartikPrabhu, KevinMarks_ and snarfed joined the channel
#snarfedso i'm trying to understand a bridgy bug/feature request from someone who wants to use fragment urls for entries inside a feed as permalinks to backfeed to
#snarfedi honestly don't fully understand the request, and it doesn't help that i can't find his markup example on his site...but i'd love input if anyone knows of precedent or has ideas!
#tantek!tell KevinMarks greetings and welcome to MozSF!
#aaronpkbut lately i've been working on a spiritual successor to RSS, or another way of thinking about iti s a successor to md5
#aaronpkif you're distributing content online people use rss.xml or md5 files to get the feed or verify the contents
#aaronpki've been writing up a data specification for a data sync project
#aaronpka file format for how you'd put an md5 hash of a data set, and you want people to access it randomly, so instead of downloading the whole thing people could download the metadata and then download only the range they want
#aaronpkif you only get the metadata then you can get the pieces over a p2p network
#aaronpkall built on top of http so you can fetch it with a regular javascript http fetch
KartikPrabhu joined the channel
#aaronpkdownload the metadata and hashes from the primary webserver, but then you can download the actual content from other web servers and you can still verify it
#Ruxtonthe metadata should break large files into multiple pieces, so you can http partial-content and range request updates to them
#aaronpki stopped transcribing but yes that's the idea
#aaronpkand each piece is verifiable from the md5 hashes you get from the source
#sknebelrelay message: don't use MD5 for new stuff, read up on better hash functions
#Ruxtononce you have those, the specifics are dependent on the game (ingress vs. pokemon go for example)
#tantekthe "server" creates pokemon at specific locations, or designates venues as either blue spinny thing "stop" that produce items, or a "gym"
#tantekclients can pickup a pokemon, or use the spinny thing that generates items that you can optionally pickup, or you can install a lure module on a stop
emmak joined the channel
#aaronpkThe cool part is all the locations were created by players of the previous iteration of this game, Ingress
#tantekcreating a pokemon is kind of like creating a geofenced post
#Ruxtonactually the locations came from Field Trip first
#tanteke.g. I could have /venue posts on my site that you could view anywhere any time (like indie equivalent of foursquare venues)
#tantekthen each venue could indicate, hey there are geofenced posts at this venue!
#tantekthen when you view my indie venue while being physically near that venue (via my site using W3C Geolocation API), then that indie venue post reveals the additional geofenced notes!
#tantekthen if you bookmark one of those geofenced notes, my geofenced note when receiving the bookmark-of webmention, could decided to disappear from the geofence, while maintaining its permalink, thus only allowing one person to "collect" that geofenced note
#tantek(would have to verify that the bookmark-of post is itself made by the person at that venue)
#[chrisaldrich]GWG: there isn't a chance that one of the IndieWeb plugins for WordPress throws in <link rel="webmention" href="http://sitename.com/?webmention=endpoint" /> into headers to automatically identify a webmention endpoint is there?
#GWG[chrisaldrich]: Yes, the webmention plugin does it automatically.
#[chrisaldrich]A ha! I was losing my mind trying to find where it was coming from as I thought I'd put it in manually. I'm guessing it also throws in <link rel="http://webmention.org/" href="http://boffosocko.com/?webmention=endpoint" /> too?
snarfed joined the channel
#[chrisaldrich]I'd had put in <a rel="webmention" href="https://www.brid.gy/webmention/wordpress"></a> manually ages ago and it was apparently causing issues with Brid.gy sending webmentions properly since the endpoint was specified multiple times.
#[chrisaldrich]snarfed: You might consider updating the brid.gy FAQ for WordPress.ORG use, or if you prefer I can do it
#GWG[chrisaldrich]: I think pfefferle figured most people wouldn't do it themselves.
#[chrisaldrich]Smart thinking for moving forward, but be aware in the near term that it may cause people with the bridgy snippet above to have webmentions failing sproradically until it's fixed.
#GWGI think it is a Bridgy issue more than a webmention plugin issue, in my opinion. Will be aware.
wolftune joined the channel
#snarfed[chrisaldrich]: funny, you think you added that manual endpoint link ages ago?
#snarfedi only added it to the docs (https://brid.gy/about#endpoint) maybe a few weeks ago. before then, the only place it showed up was in the wordpress.com signup flow
#[chrisaldrich]It's been a really long month... I know the code actually looked really familiar and I'd discussed something similar with you a while back about it in relation to the FAQ page.
#rMdesmy setup is 200 tweets at each sync, every 5 minutes. you can sync 1000 tweets every 5 minutes, but this is where your data plan is going to spike
#rMdesalso tweaking it to only download full image on wifi is worth selecting.
#lukasrosI assume the price of that app keeps the total number of users way beyond Twitter's access token limit.
#[shaners]Which is why i’m “shaners” and not “veganstraightedge”
#[ben_thatmustbemyeah, i had it hitting the max length already, which was why it was ben_thatmustbeme and not ben_thatmustbe_me
#gRegorLoveI thought I'd seen his full nick from slack before, but I guess not.
nitot joined the channel
#voxpelliaaronpk: how would one make it simpler for people to build apps? isn't it already fairly simple to build an indieweb reader app?
snarfed1 joined the channel
#voxpellilike the sound of giving Woodwind some pwa magic :)
mlncn_ and friedcell joined the channel
#aaronpkin my experience with Monocle, parsing peoples' feeds was not actually that simple, lots of edge cases
#aaronpkI would hate to have to go through all that each time I build an app
#aaronpkI started bundling some of the parsing logic into XRay so that it's all handled there and exposes a simplified/normalized version of peoples' posts
#voxpelliyeah, getting a good presentation will probably take some while and is somewhat of a moving target, but getting content in there at all should be fairly simple :)
#aaronpkyou'd think so, until you actually start parsing peoples' sites
#aaronpkthe good/bad of microformats is that it lets authors be somewhat flexible about how they arrange things in the HTML
#aaronpkmy main concern is that you show someone who's building an app the mf2 JSON (not even the HTML), and there's so many different ways you have to look for the equivalent information there, that it's way offputting compared to say the twitter API
#voxpellitrue, Superfeedr kind of services would be good, but should perhaps also try to see if there can be some best practice or convergence for at least the data needed for a a minimal reader experience
KevinMarks1 joined the channel
#voxpellithat plus some kind of ACID-like test for readers I guess
#aaronpkthere is rough convergence in those three ways
friedcell joined the channel
#aaronpkfrom the other side, it's not reasonable to ask the publisher to change their HTML in order to support a different resulting mf2 data structure
#voxpellino, not if there's something much more major than just moving a class from one place to another
#voxpelliaaronpk: do you have a test suite for xray with sample mf2 data? like I got with my testpinger?
#voxpelliaaronpk: if one could somehow select a subset of that test html that would be just standalone h-entries or similar webmentionable content, then I could pull it into my testpinger right away
#voxpellithat would make them be imported into my webmention endpoint's test suite
#KevinMarks1hm. If we're all doing this kind of normalisation, maybe converging and documenting that is what jf2 should be about?
#aaronpkmf2 is fine for what it is, a not domain specific representation
#aaronpkbut as soon as i'm building a reader app, or whatever, then i'm in the domain, and i end up doing the same algorithm/transforms of the mf2 again and again
#tantekI think domain specific stuff makes sense as utility functions as I think barnabywalters has done with phpmf2
#aaronpksure but that needs tests, and test data, and we might as well all be using the same test data which means we might as well be using the same representation of the transformations
#aaronpki'm trying to make an argument for making a thing, (whatever it's called, possibly jf2 but maybe it's something else), that multiple implementations can use
#tantekis this like the "post streams" concept I'd floated earlier?
#snarfedif there are indeed common utils for a domain like readers, sounds like aaronpk is advocating for doing them once, lang independent, instead of once per lang
#aaronpkor at least making test data that is independent of the language
#snarfed(ie basically jf2, but maybe more domain specific)
#aaronpkand having test data also implies there is essentially a spec that it's based on
#voxpelliI think the first step would be test data and then having each implementation themselves figure out if they are successfully parsing them or not
#snarfedah. common test data sgtm too, but is smaller than something like jf2
#tantekok that makes sense as a way to make jf2 immediately more useful / usable for those use-cases
#voxpelliaaronpk: the testdata I have in testpinger is real worl data and is not implying a spec but rather implying that the world is far from perfect and clients needs to expect that :)
#aaronpkvoxpelli: right but what i'm saying is that there can be a thing (spec with implementations) that converts the real world mess into something actually useful for readers and consumers
#tantekaaronpk: your challenge will be trying to avoid having more flexibility just creep into jf2, due to publisher demands
#tanteki.e. what happened with RSS, Atom, and eventually ActivityStreams
#tantekwhich then made it once again harder on consuming code
#KevinMarks1so is the thing we need 'list of entries in json'
#tantekeven just, maybe the author is in an entry, maybe it's just at the top of the feed
#aaronpkmy point is in all 6 cases, they represent the *same* list of posts
#voxpelliI'm a bit hesitant on inserting an intermediary step and saying that some interpretation is more correct than others
#tantekso you'll have to come up with a pretty tight definition for what jf2 means and what it is for in order to have it maintain that kind of "easy to consume" design
#voxpellibut exposing test data will make plenty of people implement their own solutions and then, when they find a need, they will maybe converge on sharing some solutions between them
#tantekotherwise you'll get feature creep again, by people who think the feed is the thing, instead of realizing the feed is just a summary from what's on the HTML
#KevinMarks1if we converge what webmention.io and webmention.herokuapp.com and mention-tech return, is that a way to make this concrete
#aaronpkKevinMarks1: those aren't based on consuming feeds though
#voxpelliaaronpk: we could converge on data for testing the parsing of individual h-entries first though
#aaronpkyes that's much simpler. i have way more tests in XRay for permalinks
#tantekI'm wondering if the mf2 / jf2 structures and parsing/consuming code discussion is better for #indieweb-dev
#tantekstarting to feel more plumbing (but useful) centric
begriffs joined the channel
#aaronpkthe problem with not having this consistent mapping of mf2 feeds to what you need when actually rendering the posts in a reader is that everyone ends up doing it slightly differently and then learning the edge cases individually instead of having this all documented
#voxpelliyeah, hard when a discussion transforms :) I think we've reached a good conclusion though – starting on sharing some test data for h-entries, if aaronpk can take the lead there that would be great :)
#aaronpkif I want to make a page that looks like that, but is generated by parsing the data on tantek.com, how do I find what profile info to include for the page?
#tantekin that case, they question is (should be) one of - does this page have a representative h-card
#tanteknot sure how you want to capture that in the gh issue
#voxpelliif the h-feed is a child of an h-card then just like an h-entry is connected to the h-feed by being a child of that, wouldn't the h-feed be connected to the h-card? else why is it a child?
#aaronpkvoxpelli: sure but that's just one possible arrangement of the pieces. turns out there's many more ways it happens in the wild.
#tantekaaronpk, according to http://microformats.org/wiki/h-feed you can look for a p-author on the h-feed, and if there is none, then there is no explicit author of the feed. that seems pretty simple.
#tanteka feed is *just* a set of posts, each post -> /authorship, you're done
#voxpellitantek: so if it shows up without an avatar in a Twitter like indieweb reader then that's how you prefer it to be shown? because most people would assume that to be an errenous presentation of your posts
#tantekpublishers are doing just fine not putting an explicit author on a feed, so you're not going to force them to
#tantekand frankly, I think shows the entire flaw of the name "feed reader" and assumptions that brings
begriffs_ joined the channel
#voxpelliso basically: all consuming code should always fetch the u-url of he h-entries and rely on the data there as the data in the feed should not be considered enough for a good presentation of that feed?
#KevinMarks(that's like the opposite of atom, where an entry without an author gets the author of the feed)
#gRegorLovefavicon seems a fine fallback for a feed avatar, if no feed author is listed.
#voxpellitantek: but you're saying that an h-feed that's a child of an h-card that someone follows doesn't inherit the profile data of the h-card? so it doesn't matter if I follow a profile or a feed?
#tantekperhaps the heart of where the RSS / Atom wars went wrong was what they agreed on, the very battleground they were fighting for, was the wrong battleground
#aaronpkbut in that case each entry has a different author, and it doesn't really make sense for the feed to have its own author since ethe feed isn't anything except a collection of other posts
#tantekaaronpk: but not the 80/20 case, certainly shouldn't be the source how you think or deesign
#voxpellitantek: I'm not complaining, I'm just saying that I would expect there to be an avatar as that's the expectation that has been set by current social media clients
#tantekvoxpelli: right! social media clients follow *people*/profiles, not feeds
#aaronpkvoxpelli: i'm saying you get that by interpreting the tantek.com home page as a *profile* not as a feed
#aaronpkand if you follow /authorship starting at any of his entrys, you'll get that same profile data for each post
kbs joined the channel
#voxpellithat the fact that the feed is a child of the h-card doesn't convey any meaning is a bit weird though :/ so a feed can be a child of one h-card but the entries within it would still inherit author from the representative h-card on the page
#gRegorLoveconsiders making a people reader. A flying purple one.
#voxpelliwould be happiest if no h-feed are made childs of h-cards then – if that relationship is meaningless
#aaronpkvoxpelli: that's the kind of normalization I am going to do in XRay
#gRegorLoveBarnaby's mf-cleaner lib for php-mf2 makes it easier, too.
#kbsKevinMarks / (anyone else in the know) was talking in a recent twig podcast about unifying various content-addressable storage protocols - any references to ongoing efforts here?
#kbsKevinMarks - interested as a consumer of it - but not sure yet about concrete projects. It seems potentially a handy way to publishing binary transparency logs (basically, merkle trees with published hashes of binary artifacts) - the nodes of the tree are content-addressable in principle..
#KevinMarksFollowing people is right, but sometimes you want to follow aspects of people, which may show up as separate pages or sites
#aaronpki have many feeds on my site, such as tag feeds and the somewhat type-specific feeds
#KevinMarksConverging Merkle DAGs is a possible next step with this thinking, kbs, but converging hashes of discrete content seems both a good entry point and something practically useful
#kbsmakes sense :-) and will poke around your notes and links (thank you for identifying the problem + motivating efforts towards consolidation!)
#KevinMarksRight, Aaron. Feeds as lists of entries area a useful thing to represent more than just a person's entries. You have consumption feeds as well as production feeds, (what am I reading, who replied etc) so multi author is common
nitot joined the channel
#tantekKevinMarks, agreed about aspects, but that should be a UI where you pick that from the profile, not where you offload "feed management" onto the user (as feed readers do)
#voxpelliWe really need more reader apps to iron it out I think
#beari'm liking this conversation about feeds and authorship
Gold joined the channel
#bearbecause I feel the past (atom, rss, whatevs) was all about publishing in one direction
#bearand today we are about curating of content - which can be single source or mutlple source
#bearand the curator should be the one with the onous of providing citation, which with webmentions is built-in
#KevinMarksWell, a lot of feed publishing was about sharing links to others posts too
#Loqimarked safe is a notification feature that tells you when those you follow have marked themselves as safe in an area affected by a disaster or other physical emergency within a geographical area https://indieweb.org/marked_safe
KevinMarks1 joined the channel
#KevinMarks1message was one of medium's first magazines, when it was a platisher
#aaronpkneeds to make an order form for this stuff :)
#[chrisaldrich]aaronpk, my address is in my h-card on www.boffosocko.com; let me know the price, shipping, and preferred method of payment and we'll send you everything with loads of extra appreciation
#[chrisaldrich]I looked around and didn't see any HWC logos, has anyone considered making one? I do love the fact that google image searches for the term give back lots of photos of people though.