#indiewebcamp 2015-09-29

2015-09-29 UTC
cmhobbs, lukebrooker, snarfed, nitot and tantek joined the channel
#
tantek ownyourresponses++
#
Loqi ownyourresponses has 1 karma
#
tantek what is ownyourresponses?
#
Loqi ownyourresponses is a project to enable PESOS as a service for likes, replies, reposts, and event RSVPs from silos to your own indieweb site https://indiewebcamp.com/ownyourresponses
#
tantek KartikPrabhu: that's negative reasoning too ;) re: http://indiewebcamp.com/irc/2015-09-28/line/1443474230521
#
tantek KartikPrabhu: alternatively, if you have a solution, it can be helpful to suggest it along with reasons why it may be better than existing approaches.
#
aaronpk I might actually work on my website tonight \o/
davidmead joined the channel
#
tantek or, with or without a solution, sometimes asking questions is more helpful than offering criticism
#
tantek aaronpk: \o/
#
tantek kylewm++ for https://kylewm.com/2015/10/sf-homebrew-website-club and the FB POSSE copy!
#
Loqi kylewm has 246 karma
rhiaro_, Guerillero|BNC, snarfed and nitot joined the channel
#
KevinMarks !tell snarfed I still can't get the oauth-dropins virtualenv thing to work. is there another way to use it?
#
Loqi Ok, I'll tell them that when I see them next
#
aaronpk anyone want to chat collections right now?
KartikPrabhu joined the channel
#
aaronpk I feel like I could use some outside opinions
lukebrooker, snarfed and nitot joined the channel
#
aaronpk !tell tantek I would be curious to see your thoughts on implied post types written up as a spec, for the purpose of deciding which posts to show in my different "collections/feeds"
#
Loqi Ok, I'll tell him that when I see him next
#
snarfed KevinMarks: ugh, sorry for the trouble. it kinda does need the virtualenv. what error are you getting?
#
Loqi snarfed: KevinMarks left you a message 1 hour, 39 minutes ago: I still can't get the oauth-dropins virtualenv thing to work. is there another way to use it? http://indiewebcamp.com/irc/2015-09-28/line/1443492039884
#
KevinMarks pip is installing into /usr/local/lib/site-packages not into the venv
#
KevinMarks not sure how I made it do that
#
@tedcurran DAEMON Sync Shares Files Over Local Wi-Fi, No Internet Required http://lifehacker.com/daemon-sync-shares-files-over-local-wi-fi-no-internet-1732976961 #indieweb (twitter.com/_/status/648708178658463744)
#
KevinMarks ah, pip need to escape the spaces too.
nitot and [snarfed] joined the channel
#
[snarfed] KevinMarks: woo, sounds like progress? is there something in setup.py or the readme i should update?
snarfed and wolftune joined the channel
#
KevinMarks this 2011 bug in virtualenv https://github.com/pypa/virtualenv/issues/53
#
KevinMarks fixed by this patch https://github.com/pypa/virtualenv/pull/611/files
#
KevinMarks yakshaving is one thing, people keeping their yaks shaggy on purpose is another
#
KartikPrabhu one person's shaggy yak is another one's sweater?
#
KevinMarks well, the real yakshaving here is the absolute hashbang paths. Why they can't be relative in virtualenv i don't know
#
KartikPrabhu python overlord
[snarfed] joined the channel
#
[snarfed] oh wow that is pretty obscure
#
[snarfed] KevinMarks++for finding it!
#
Loqi KevinMarks has 168 karma
#
[snarfed] KevinMarks: I'll happily include a warning in the o-d readme if you want to send a PR!
nitot, loic_m, wolftune and cweiske joined the channel
#
@tinokremer RT @benwerd: You can't solve the #redecentralize / #indieweb problem without figuring out where the money is coming from. (twitter.com/_/status/648746503704547328)
tantek joined the channel
#
tantek good evening #indiewebcamp
#
Loqi tantek: aaronpk left you a message 2 hours, 56 minutes ago: I would be curious to see your thoughts on implied post types written up as a spec, for the purpose of deciding which posts to show in my different "collections/feeds" http://indiewebcamp.com/irc/2015-09-28/line/1443497824862
yakker joined the channel
#
tantek aaronpk: drafting here https://indiewebcamp.com/post-type-discovery
#
tantek going to start with simple more widely implemented types
Jihaisse and fkooman joined the channel
#
M-kegan lurks on the post types discussion
rosetree joined the channel
#
M-kegan Hmm, I can't really say much without knowing why explicit post types are "being abandoned by post creation UIs"
#
M-kegan I assume there is a reason for it, but I don't see that anywhere
rosetree joined the channel
#
M-kegan But implied types trigger my knee jerk DONOTWANT spider sense. It becomes hard to do anything more exotic than what was considered when doing the implied algorithm
#
M-kegan And that's ignoring the combinatorial explosion of "location clobbers pic unless there is foo in which case bar" which then becomes really hard to grok. Explicit types make it easier to reason about in general which I much prefer. Hence wanting to know a good reason for considering it
#
M-kegan it = implied types
friedcell, mlncn and Geng joined the channel
#
tantek.com edited /post-type-discovery (+443) "/* Brainstorming */ list which types I plan on writing up an algorithm for first" (view diff)
#
tantek M-kegan: better UX is >>> "easier to reason about in general"
glennjones joined the channel
#
tantek in fact, so much so that "easier to reason about in general" is a very poor way to design anything where a user in is involved. since "reason about" implies data-centric = plumbing-centric thinking which is an anti-pattern.
#
tantek as to why explicit post types are "being abandoned by post creation UIs", you'll have to ask the common popular silos, however one aspect is, lower cognitive load for creating content.
#
tantek re: combinatorial explosion, hypothetical concern, does not occur in practice
#
tantek meaning we can specify something arbitrary and it works just fine
interactivist, ttepasse, loic_m, fkooman, petermolnar and squeakytoy joined the channel
#
M-kegan I reject the notion that "data-centric = plumbing-centric" - if you want to encode display information in the request along with representing the post in a way which is usable beyond display (searching, subscribing to post type subsets) - the most flexible display is arguably HTML which is a pita to parse, the most pure data form being explicit types with properties relevant to the type. To me, it sounds like the
#
M-kegan implied types is a way of trying to "compromise" one or the other, rather than keeping them completely out of each other's hair (e.g. with an html key and allowing template style {{vars}})
#
M-kegan I also don't buy "does not occur in practice" as a general rule - of *course* you don't think it will occur in practice, the whole point is to design a solution which solves all the things you think *do* occur in practice. My argument is that you don't know what the end-user will do and making assumptions about how types are composed sounds like Trouble as end-users develop bodges to work around assumptions baked into
#
M-kegan the implied-type algorithm
#
M-kegan e.g. it's feasible that I *want* to tie a p-location to a u-photo (because the location is NOT a check-in but is a "this photo was taken at this location") rather than the assumption that it is tied to the entire h-entry
pfefferle, ttepasse, ttepasse_ and elf-pavlik joined the channel
#
Zegnat Good morning IndieWeb
nitot joined the channel
#
pfefferle good morning folks
fkooman joined the channel
#
petermolnar good morning
#
Zegnat anyone working on something special today?
dns53, acegiak, friedcell, eschnou, scoates, petermolnareu, LanceyWork, pfefferle, nitot, hs0ucy, keroberos, rosetree, kerozene, rosetree1, snarfed and tantek joined the channel
#
tantek.com edited /post-type-discovery (+124) "/* Brainstorming */ note other types being considered, and refine criteria" (view diff)
#
tantek good morning #indiewebcamp
#
rosetree1 good afternoon from germany
#
tantek hello and welcome rosetree1!
#
rosetree1 hm … i thought i registered with rosetree. maybe login did not work due to network problems
#
voxpelli good morning!
#
tantek !tell M-kegan the counterpart of "does not occur in practice" is designing what *does* occur in practice, and deprioritizing (not worrying about) what there is no evidence for, or what "will occur". Also "all the things" is not the point, but rather, solving for the 80/20 *today*, and continuing to gather data, and iterating.
#
Loqi Ok, I'll tell them that when I see them next
#
tantek !tell M-kegan you're right about worrying what the end-user will do - but that involves designing UX first e.g. /create - and "data model" only secondarily based upon that UX. for more see: http://indiewebcamp.com/design#UX_Before_Infrastructure
#
Loqi Ok, I'll tell them that when I see them next
#
tantek !tell M-kegan current thoughts on checkins are here: https://indiewebcamp.com/checkin - there are some recommended practices however no convergence in practice as of yet - would be great to see how you publish checkins on your personal site!
#
Loqi Ok, I'll tell them that when I see them next
#
tantek good morning to you too voxpelli!
#
tantek is still struggling with iOS9 wifi support being nearly totally broken.
#
kylewm hmm, it is unfortunate that the micropub spec is using a mix of - and _ in properties... e.g. "in-reply-to", "syndicate-to[]" vs. "place_name" and "access_token"
#
voxpelli kylewm: "in-reply-to" comes from the mf2 parsing spec
#
tantek the properties in general come from h-entry
#
tantek I don't know what "place_name" is for.
#
tantek "access_token" sounds like ephemeral protocol plumbing rather than content
#
tantek and I thought all the mp specific properties were mp-* prefixed
#
voxpelli "access_token" is just for those cases where you can't set a header – like if you're submitting a HTML-form
#
tantek kylewm: probably worth capturing as an issue in /Micropub#Issues
#
kylewm tantek: OK will do
wolftune joined the channel
#
kylewm place_name is from OwnYourGram, probably just an artifact of it being one of the first implementations
#
voxpelli tantek: kylewm: they should all be "mp-" prefixed, yes: http://indiewebcamp.com/micropub#Reserved_properties
#
tantek kylewm: any thoughts on how to post (discover) a photo with a location, vs. a checkin at a venue with a photo?
loic_m joined the channel
#
kylewm sounds like the photo with a caption vs. note with an attached photo problem
#
kylewm I don't off hand, will give it some thought
#
rosetree1 Would you recommend making replies a specific post type (with special URL) or a post (article, note, photo, …) with an in-reply-to link?
#
tantek kylewm: it's similar except that in practice (in /create UIs) as soon as a user uploads a photo, there is consensus that it's a "photo" post, regardless of whether they'd started writing a note or not.
loic_m_ joined the channel
#
tantek I think we've figured out /photo vs /note pretty well - thanks to lots of experimentation and iteration
#
tantek We're less far along with figuring out photo+location vs. checkin+photo
#
tantek need to document more existing UIs for posting photos vs doing checkins
#
tantek checkin UIs are a challenge period
#
tantek for everyone, including silos, given waning checkin usage in both Swarm (Foursquare), and even Facebook.
#
tantek ironically many of my Swarm/4sq checkins are via photo posts from IG with location / venue info attached
#
tantek so what does that mean when a UI for photo+location posting is used to cross-post checkin+photo?!?
#
Zegnat Didn’t Instagram change their location policy not too long ago to make it harder for your location to pop-up on maps? I wonder if that will negatively effect the number of checkins you are seeing.
#
aaronpk good morning
#
aaronpk kylewm: yeah place_name is definitely an artifact and I would happily change it
#
aaronpk and tantek is correct that access_token is plumbing stuff, not really part of the micropub request, since it's from OAuth 2 and is only used if a header is not sent
#
aaronpk rosetree1: my current site treats replies as their own post type with their own URL structure and now I regret doing it that way
#
aaronpk compare http://aaronparecki.com/replies and http://aaronparecki.com/notes, they show different sets of things
#
rosetree1 aaronpk: thanks for your answer. have you explained somewhere why you regret that way? what would you do instead?
#
aaronpk the problem is when I want to include a photo with the reply, (which I didn't think of at first), I'd have to update my reply template to be able to display photos
#
aaronpk i'm switching now to this "implied post type" idea, where everything starts as a post, and posts can have various properties like a photo, an "in-reply-to", etc, and the code knows how to display the posts depending on what's in it
#
aaronpk i'm still working out the details, but I do plan on writing everything up
nitot joined the channel
#
tantek rosetree1: re: your question, I have both /article and /note replies on my site, and what makes them replies is the presence of an in-reply-to link. I have also seen /photo replies e.g. in comment threads (even on github) where literally the only thing in the /reply is a photo (like a meme image).
#
tantek could easily see that expanding to video/audio replies etc.
#
rosetree1 I’m currently thinking about differentiating between articles and notes. Just to make sure visitors can subscribe to multiple feeds (long posts, short posts, everything).
#
tantek aaronpk: what do you think of changing "place_name" to just "location" ? which can be a string (name), or a URL (to a /venue )
#
tantek rosetree1: those two are not connected
#
aaronpk my new approach is that I have these "collections" or feeds, where any post can be in any feed regardless of type
#
tantek that is, you can provide a UI / archives of "multiple feeds (long posts, short posts, everything)" *regardless* of how you differentiate internally
#
rosetree1 tantek: sorry, i’m not really sure what you mean by connected
#
aaronpk last night I started writing some rules that automatically add posts to different feeds based on their implied post type, but I can always override that if I want
#
tantek e.g. right now I have explicit distinction between articles and notes in my URL structure, but I'm leaning more towards getting rid of that (making everything "notes")
#
tantek becuase I can infer from a note whether it is more article-like
#
tantek aaronpk: will be interesting to see how your rules and the post-type-discovery spec overlap
#
aaronpk I definitely support the idea of publishing multiple feeds with different kinds of content (long posts, short notes, photos, checkins, etc)
#
aaronpk tantek: I suspect they will be identical!
#
tantek aaronpk, I like your optimism :D
#
aaronpk well, at least identical for the post types that are in common. I have a few more feeds I'm planning on publishing than what's on the wiki page now
#
rosetree1 I think I’m going to drop my idea of a specific post URL. I don’t completly understand why and couldn’t explain it, but it seems to be better :)
#
rhiaro I'm going to expand my different types of feeds soonish, and work out rules for automatically adding to a feed, plus ui for manually indicating feeds at post-time
#
rhiaro But I'm particularly interested in how to do feed discovery from the homepage
#
rhiaro As not everything will be in main homepage feed
#
rhiaro So if you want to subscribe to be you get options of different sets of stuff to follow.. Be good if a reader could offer this, rather than relying on someone knowing what they want to subscribe to
#
aaronpk totally!
#
aaronpk i want exactly that as well
#
aaronpk also I think it's funny that I'm ending up back here after having designed+built exactly that feature into a twitter clone the year after twitter launched
#
rosetree1 I guess a static site is somehow a disadvantage to achieve such goals.
#
aaronpk nah it just means the build process is more complex
#
rhiaro Ideas on how to do feed discovery yet, aaronpk?
#
rhiaro What was the twitter clone?
#
Loqi It looks like we don't have a page for "twitter clone" yet. Would you like to create it? http://indiewebcamp.com/s/103e
#
aaronpk rhiaro: I think some people have already experimented with feed discovery, doesn't Woodwind support it?
#
aaronpk http://indiewebcamp.com/feed_discovery
#
rhiaro Don't know off the top of my head
#
rhiaro Cool, thanks
#
rhiaro On mobile currently, only half here
#
rosetree1 aaronpk: too complex for me, currently.
#
rhiaro I'll add some stuff to that page soon
#
kylewm ya, Woodwind checks for h-feed or a collection of h-entrys on the home page, and also looks at rel=feed and rel=alternate
#
aaronpk sure, static sites can be a challenge for sure
#
aaronpk rhiaro: it was called "Untied". i archived some screenshots here https://www.flickr.com/photos/aaronpk/albums/72157633041808362
#
rhiaro rel=feed seems like a good place to start
#
kylewm.com edited /IRC_People (-2) "/* Nicknames */" (view diff)
cleverdevil, mlncn, ttepasse and shiflett joined the channel
#
rosetree1 so i have to leave. thanks for your answers.
#
@ChrisAldrich The future is #indieweb anyway. https://indiewebcamp.com/ https://twitter.com/jessalbert/status/648894000187899904 (twitter.com/_/status/648898497333215233)
gavinc joined the channel
#
tantek.com edited /post-type-discovery (+2482) "draft with license, methodology, algorithm, next steps, references, see also" (view diff)
#
aaronpk !
snarfed joined the channel
#
tantek.com edited /post-type-discovery (-1) "/* Algorithm */" (view diff)
snarfed joined the channel
#
M-kegan !tell tantek I agree that UX driven development is good. I also agree that solving the problem for the majority of use cases is pragmatic and iterating for "the 20%" is a good strategy.
#
Loqi M-kegan: tantek left you a message 2 hours, 32 minutes ago: the counterpart of "does not occur in practice" is designing what *does* occur in practice, and deprioritizing (not worrying about) what there is no evidence for, or what "will occur". Also "all the things" is not the point, but rather, solving for the 80/20 *today*, and continuing to gather data, and iterating. http://indiewebcamp.com/irc/2015-09-29/line/1443537703049
#
Loqi M-kegan: tantek left you a message 2 hours, 30 minutes ago: you're right about worrying what the end-user will do - but that involves designing UX first e.g. /create - and "data model" only secondarily based upon that UX. for more see: http://indiewebcamp.com/design#UX_Before_Infrastructure http://indiewebcamp.com/irc/2015-09-29/line/1443537845769
#
M-kegan !tell tantek I see the problem as two distinct issues : "how do I represent a complex type post in such a way that I can subscribe sensibly to all checkins say" and "how do I represent this complex type post on the UI".
#
Loqi M-kegan: tantek left you a message 2 hours, 28 minutes ago: current thoughts on checkins are here: https://indiewebcamp.com/checkin - there are some recommended practices however no convergence in practice as of yet - would be great to see how you publish checkins on your personal site! http://indiewebcamp.com/irc/2015-09-29/line/1443537942461
#
Loqi Ok, I'll tell him that when I see him next
#
Loqi Ok, I'll tell him that when I see him next
#
M-kegan !tell tantek to do that in a way that works for "the 80%" I feel requires these issues to be resolved in isolation where possible
#
Loqi Ok, I'll tell him that when I see him next
wolftune joined the channel
#
Kongaloosh tbrb: do you have webmentions going yet?
#
tantek.com edited /post-type-discovery (+116) "feedback" (view diff)
emceeaich and snarfed joined the channel
#
emmah.net edited /IRC_People (+98) (view diff)
#
emceeaich waves
#
aaronpk hi!
rabble joined the channel
#
tantek.com edited /post-type-discovery (+238) "add explicit Participate section to header with links to Feedback, Issues, IRC" (view diff)
#
tantek.com edited /post-type-discovery (+396) "/* Next steps */ explicitly note plan to to use W3C wiki, github, irc if spec is accepted as an editor's draft in the Social Web WG" (view diff)
#
tantek aaronpk: since you previously expressed interest, could you review https://indiewebcamp.com/post-type-discovery#Algorithm for 1) what have you implemented, and 2) what you agree/disagree with (whether conceptually or in code!)
loic_m joined the channel
#
Loqi tantek: M-kegan left you a message 56 minutes ago: I agree that UX driven development is good. I also agree that solving the problem for the majority of use cases is pragmatic and iterating for "the 20%" is a good strategy. http://indiewebcamp.com/irc/2015-09-29/line/1443546843092
#
Loqi tantek: M-kegan left you a message 56 minutes ago: I see the problem as two distinct issues : "how do I represent a complex type post in such a way that I can subscribe sensibly to all checkins say" and "how do I represent this complex type post on the UI". http://indiewebcamp.com/irc/2015-09-29/line/1443546854304
#
Loqi tantek: M-kegan left you a message 55 minutes ago: to do that in a way that works for "the 80%" I feel requires these issues to be resolved in isolation where possible http://indiewebcamp.com/irc/2015-09-29/line/1443546927240
#
tantek !tell M-kegan I agree /checkin is challenging in many ways (UI, representation etc.) hence have not included it in the current (first) explicit version of Post Type Discovery. Feedback welcome! https://indiewebcamp.com/post-type-discovery
#
Loqi Ok, I'll tell them that when I see them next
#
tantek !tell KartikPrabhu you expressed +1 interest in /post-type-discovery - please take a look and provide feedback! http://indiewebcamp.com/post-type-discovery
#
Loqi Ok, I'll tell them that when I see them next
nitot joined the channel
#
aaronpk reviewing...
#
tantek thanks!
#
tantek.com edited /post-type-discovery (+10) "/* Other Types Under Consideration */ jam per request from KevinMarks" (view diff)
#
aaronpk my filtering by implied post type that I needed yesterday is actualy slightly different than the explicit algorithm here
#
tantek.com edited /jam (+277) "FAQ / Why ♫ - per question from KevinMarks" (view diff)
#
aaronpk what I was trying to do is decide "given this post, which feeds should it be included in"
#
snarfed wow, interesting. http://stream.withknown.com/2015/known-announces-dreamhost-as-exclusive-web-hosting-sponsor
#
snarfed especially given benwerd's posts yesterday
#
aaronpk whoa!
#
cleverdevil may have had something to do with that....
#
cleverdevil ;)
#
snarfed i wonder if cleverdevil was involved at all
#
snarfed haha jinx
#
snarfed congrats!
#
cleverdevil gotta find ways to support awesome open source projects.
#
cleverdevil and, more importantly, get more people to *use* them.
#
snarfed totally. much appreciated!
#
cleverdevil you bet.
#
cleverdevil happy to do more, too.
#
cleverdevil we're more successful in a world where people own their own data, domains, and are hosting on open platforms rather than in silos.
#
tantek aaronpk - you could do that in two steps
#
tantek 1) what type of post is this (apply post type discovery) ? 2) matrix of post types -> feeds
#
tantek the second step is purely presentational and custom to your site
#
tantek however it is still useful to agree on 1) because that affects the presentation of each post regardless of what feed(s) you put it into
#
aaronpk the problem is when a post has multiple properties that fall under this post type discovery
#
aaronpk such as a reply that includes a photo
#
tantek that's a reply
#
aaronpk then it wouldn't show up in my photo stream
#
tantek and I'd say in practice that's what you want. e.g. you reply to a github issue with some random meme image - did you really want that in your "photo stream" ?
#
aaronpk hm good point
#
tantek the converse is, I totally want the freedom to reply to people with WHATEVER, and NOT have it pollute the streams of "my stuff" that people subscribe to
#
aaronpk checking my list for other things..
#
tantek but this is an EXCELLENT FAQ
#
aaronpk i think that is the only composite post that falls under the current algorithm
#
tantek RSVPs with comments are another interesting case
#
tantek as discussed previously with kylewm
#
aaronpk rsvps also have to have in-reply-to URLs right?
#
tantek correct
#
aaronpk so the question is whether I would want my RSVPs with comments to also be shown in the list of my replies
#
KartikPrabhu Loqi wassup?
#
Loqi KartikPrabhu: tantek left you a message 24 minutes ago: you expressed +1 interest in /post-type-discovery - please take a look and provide feedback! http://indiewebcamp.com/post-type-discovery http://indiewebcamp.com/irc/2015-09-29/line/1443550347900
#
tantek the empirical rephrasing of that question is, do RSVPs with comments show up in the list of replies on existing Event permalinks?
#
aaronpk not quite, more like can you see RSVPs with comments on someone's profile page
#
tantek.com edited /post-type-discovery (+967) "FAQ / What about a photo reply" (view diff)
#
tantek aaronpk - depends I believe you can see "So and so is going to event name and said ..."
#
tantek or rather I believe I have seen
#
tantek now I can't recall what silo that was
#
@melissagira @courtneytrouble swear to god so much indieweb that's still sticking around today was built from tori (and related) fandom (twitter.com/_/status/648929303116255232)
#
aaronpk whaaaaat http://recode.net/2015/09/29/twitter-plans-to-go-beyond-its-140-character-limit/
#
tantek lolol
#
snarfed do we have guidelines or an algorithm for how to present webmention sources with multiple "types?" e.g. if the target link has both in-reply-to and like-of?
#
snarfed example: https://github.com/idno/Known/issues/1072
#
snarfed i didn't see anything on /comments-presentation, /comment, or /multiple-reply
benwerd joined the channel
#
aaronpk i think the h-entry vocabulary says that a post with "like-of" can't have other content
#
aaronpk so either a) it's an invalid post, or b) it's just an in-reply-to according to the discovery algorithm
emceeaich joined the channel
#
tantek oh boy snarfed, deliberately making trouble ;)
#
snarfed you inspire me tantek!
#
tantek touché!
#
snarfed aaronpk: aww that's sad
#
tantek actual example: https://snarfed.org/2015-09-29_known-announces-dreamhost-as-exclusive-web-hosting-sponsor :)
#
aaronpk "like with comment"
#
aaronpk is that a new post type?
#
benwerd I'm going to fix it for now by prioritizing replies and then I'll think about multiple actions from a single mention later
#
tantek snarfed: in the current post type discovery your reply like is treated as a reply.
#
tantek nice to hear benwerd came to the same conclusion
#
aaronpk i suppose it's not unreasonable to treat that as two actions and display it accordingly. I would display both the profile pic in the "likes" facepile, and also show the comment
#
tantek aaronpk exactly! and note that that display heuristic does not require any knowledge of explicit typing!
#
aaronpk which brings me back to my original question, what is the purpose of implied post type discovery then!
#
tantek in other words, you can go directly from properties -> affects on display
#
tantek having some use-cases does not mean you must use it in all use-cases
Stephan_ joined the channel
#
aaronpk what's the point of saying "this post is a _____" when what you'd actually be doing is looking at the properties on the post and deciding one or more things to do with it
#
kylewm cleverdevil: wow you graduated from ga tech two years before me!
#
cleverdevil :D
#
tantek or rather, if your use-case needs an explicit type, then use post type discovery :D
#
cleverdevil kylewm: awesome 😀
#
tantek otherwise just use the properties directly
#
tantek btw here's a degenerate use-case: to synthesize an artificial AS(1,2) stream from an h-feed
#
aaronpk that's what the page needs, a "use cases" section
#
tantek nice
#
benwerd aaronpk++
#
Loqi aaronpk has 963 karma
#
tantek.com edited /post-type-discovery (+1) "/* What about a photo reply */ ." (view diff)
#
benwerd I strongly agree with "take a look at this post and figure out what to do based on its properties"
#
benwerd I want to kill post types in Known for the same reason
#
KartikPrabhu new comment silo: https://civilcomments.com/
#
tantek part of the point of this spec is for the implementers that apparently think "but I MUST know what the post type is because ..."
#
snarfed tantek: feel free to try out the existing AS from h-feed implementation on https://granary-demo.appspot.com/ :P
#
Loqi [mention] Kyle Mahan wrote a post that linked to an event: "Homebrew Website Club Meetup" https://indiewebcamp.com/events/2015-10-21-homebrew-website-club https://webmention.io/notification/G6rnB9JrVF7wAdhHGwROjQ
#
tantek.com edited /post-type-discovery (+145) "put algorithm first, methodology second as explanation for aspects of it" (view diff)
#
aaronpk tantek: that's a totally valid reason for having the spec IMO :)
#
kylewm huh, who sent that webmention
#
aaronpk kylewm: haha what
#
@kevinmarks “The social era fragmented the networks by media type.”-@bhorowitz https://medium.com/@bhorowitz/now-let-me-holler-at-ya-partner-spit-this-game-in-you-ear-for-a-minute-quit-complaining-bout-70435c14e3d7 fixing that http://indiewebcamp.com/post-type-discovery (twitter.com/_/status/648934768562712576)
#
tantek oh boy (again). thanks kevinmarks :P
#
KevinMarks his points about medium's UI are strong
#
KevinMarks "2. Blogging on Medium feels like a new media format native to the web and that could only come post social. The best Medium posts are a combination of collaging and blogging. You include tweets, gifs, pull quotes, graphics, and videos along with your text. The social era fragmented the networks by media type. Medium gives you a way to pull it back together."
eschnou joined the channel
#
kylewm aaronpk: this webmention, https://webmention.io/notification/G6rnB9JrVF7wAdhHGwROjQ it's correct, but it should have been sent 2 days ago
#
kylewm i'm not sure why it just showed up now
#
aaronpk might be some weirdness with the clustering delay
#
KevinMarks my tweets from hwc were showing up in reverse order last week
emceeaich joined the channel
#
aaronpk that's a twitter search glitch
#
tantek snarfed, interesting! I tried it on my sites, and it thinks my notes are articles!
#
tantek even though displayName === content !
#
snarfed tantek: interesting, thanks for the report
#
tantek which is even more than necessary
#
snarfed it's very alpha and doesn't actually implement many algorithms like post type discovery
#
tantek.com edited /post-type-discovery (+700) "Use cases / Synthesizing explicit type formats" (view diff)
shiflett joined the channel
#
snarfed issues always welcome. thanks for trying it!
#
tantek that is, see steps 6 onward here: https://indiewebcamp.com/post-type-discovery#Algorithm
#
snarfed if/when i get to features like that, kylewm's mf2util is pretty great, i'd probably want to reuse it
#
snarfed (i think it implements some if not all of those algorithms)
#
tantek indeed I need to look into what else kylewm had to do to distinguish notes from articles
#
tantek I remember him saying he had to do a bit more
#
KevinMarks this leads towards having examples and tests to converge
#
tantek !tell kylewm would appreciate your feedback on post-type-discovery, especially https://indiewebcamp.com/post-type-discovery#Algorithm step 6 onward!
#
Loqi Ok, I'll tell them that when I see them next
#
snarfed tests++
#
snarfed testing++
#
Loqi tests has 15 karma
#
Loqi testing has 10 karma
#
tantek aaronpk: per your request: https://indiewebcamp.com/post-type-discovery#Use_cases
#
aaronpk \o/
snarfed joined the channel
#
tantek.com edited /post-type-discovery (+516) "Implementations section / Granary" (view diff)
#
tantek snarfed++ thanks for Granary! Included: https://indiewebcamp.com/post-type-discovery#Granary
#
Loqi snarfed has 144 karma
#
snarfed aww welcome!
#
tantek I hope the characterization / description is accurate!
#
snarfed i'll tweak a bit, but broadly yes, thanks
#
snarfed.org edited /post-type-discovery (+33) "/* Granary */ tweak language" (view diff)
cleverdevil and mlncn joined the channel
#
kylewm tantek: I actually do step 11 the opposite way. if content ⊆ name, (and just contained in, not necessarily the prefix) with the idea that name will either the be == the plaintext equivalent of content, or it was auto-generated and will be the content + all the other stuff around it
#
Loqi kylewm: tantek left you a message 43 minutes ago: would appreciate your feedback on post-type-discovery, especially https://indiewebcamp.com/post-type-discovery#Algorithm step 6 onward! http://indiewebcamp.com/irc/2015-09-29/line/1443554176788
#
kylewm checking whether name is a prefix of content would identify Emily Dickinson poems as notes :P
#
tantek people do post poetry on Twitter so that would work
#
tantek poems are just notes with whitespace preserved right?
#
tantek from a post presentation perspective obv ;)
#
kylewm hehe, sure
#
tantek aside: https://twitter.com/emilydickenson lol
nitot joined the channel
#
KevinMarks https://twitter.com/kevinmarks/status/648381075937366016
#
@kevinmarks The darkness cannot comprehend the light I couldn't even see the moon tonight @pentametron https://twitter.com/kevinmarks/status/648381075937366016/photo/1 (twitter.com/_/status/648381075937366016)
#
Jeena If you don't have a good and flexible application structure from the start, you gonna have a hard time to fix that later. But if you wait until you have a good structure before you start, you never gonna write anything anyway, so refactor all the things!!!
nitot_, gRegorLove, KartikPrabhu and snarfed joined the channel
#
Jeena If I have a note A and I reply myself to that with a comment B then with salmentions I would need to resent webmentions to all the comments associated with note A, does it mean I should send a webmention to comment B too because it is now associated with note A? Those salmentions are a complicated beast.
camerongray_ joined the channel
#
tantek Jeena, start with doing less work and see what breaks :)
#
@benwerd I do not give Facebook permission to use my pictures ... #indieweb http://werd.io/2015/i-do-not-give-facebook-permission-to-use-my-pictures (twitter.com/_/status/648954997506359296)
#
gRegorLove Jeena: I thought Note A would only need to send an updated webmention to whatever it was in-reply-to (if anything), not all the comments on it.
#
gRegorLove What is salmention?
#
Loqi Salmentions are a way to propagate comments upstream by sending a webmention from a reply post to the original post when the reply recieves a comment https://indiewebcamp.com/salmention
#
Jeena but those comments want to know about other comments from 3rd parties
#
gRegorLove Why?
#
Jeena if you take a picture of me and tag it, then I write a comment and later tantek writes a comment too then I want to know that he wrote that comment
Geng joined the channel
#
gRegorLove Ah, hm.
#
Jeena the only way for me to know that is if you send me a webmention (to my comment) that tantek commented too
#
gRegorLove Yeah, person-tag seems like a special case there.
#
gRegorLove But if I comment on your Note A, I don't necessarily want to get a salmention everytime someone else comments on it.
#
Jeena ok forget the person-tag, you write a note, I comment on that note, later tantek comments on this note too, I still want to be informed that he did that
#
Jeena I think you want to get a salmention, you perhaps don't want to do anything with it but you want to have the posibility
cleverdevil joined the channel
#
gRegorLove Sounds like a side-mention of sorts, "subscribe me to the discussion on this note"
#
gRegorLove But I understand better what you mean.
#
tantek indeed, notification of follow-up comments is common behavior on silos e.g. FB
#
tantek or likes for that matter
#
gRegorLove Is anyone using salmentions like that currently?
sparverius joined the channel
#
Jeena I thought that was the whole idea behing salmentions ^^
#
gRegorLove I typically find that default-subscribe-to-thread on FB annoying
#
aaronpk no that's not the original goal of salmentions
#
Jeena but if it's not there you will miss most of the discussion
#
aaronpk swat0 describes the original goal
#
aaronpk the author of a post mentions you, so then you start getting updates when people comment
#
gRegorLove My understanding is that it's about sending webmentions up the stream of in-reply-to URLs.
#
aaronpk which is ever so slightly different from you commenting on a post and then getting updates when others comment
#
tantek they are distinct use-cases yes and worthy of being documented as such
#
Jeena so salmentions are only for persontagged posts?
#
aaronpk no, that's just where it originated
#
gRegorLove No, sorry for adding confusion by bringing up person-tag
#
loqi.me created /architecture_astronaut (+35) "prompted by tantek and dfn added by tantek" (view diff)
#
Jeena hm, ok, I tried to read the definition really slowly on http://indiewebcamp.com/salmentions and I guess gRegorLove is right, it seems only to cover the in-reply-to chain of events
#
gRegorLove kylewm and acegiak's examples on there might help too
#
Jeena I don't know, I guess I got confused about how many comments there are on https://kylewm.com/2015/06/yeah-comment-threads-are-sort-of-a-todo-for-the-indieweb#kyle_mahan-4565 so I thought it is the whole tree
#
tantek Jeena - if you come up with additional use-cases, please document them on the wiki!
#
aaronpk it sounds like this is a potential new use case, but I think it can be treated separately!
#
aaronpk as in, you can implement one without the other
Geng2 joined the channel
#
kylewm Jeena: wow that comment thread is a mess!
#
Jeena and it is our best example of threaded comments ;)
#
aaronpk i like the last comment is a facepalm avatar
#
Jeena hehe
#
kylewm I usually de-duplicate comments... wonder why there are so many with the same permalink in this thread
#
Jeena but ok, just doing the in-reply-to stuff with salmentions should make things much easier
#
Jeena I mostly want to be able to do swat0 and am not getting done anything because it got overly complicated how I understood the whole system
elf-pavlik joined the channel
#
Loqi [mention] Kevin Marks wrote a post that linked to a photo: "post-type-discovery" http://indiewebcamp.com/post-type-discovery https://webmention.io/notification/kBNXaejRbI3CQi4vzf590g
nitot joined the channel
#
tommorris I heard a rumour that Twitter are now planning to remove the 140 limit. Didn't have a chance to confirm it. This makes POSSE interesting.
#
kylewm makes me think of how many person hours spent on writing ellipsizing algorithms :)
#
Zegnat I saw that too, tommorris, on The Next Web, I think
#
kylewm http://recode.net/2015/09/29/twitter-plans-to-go-beyond-its-140-character-limit/
#
tommorris http://www.slate.com/blogs/future_tense/2015/09/29/twitter_s_140_character_limit_isn_t_going_away_here_s_what_s_really_happening.html
#
Zegnat http://thenextweb.com/twitter/2015/09/29/twitter-is-reportedly-ditching-the-140-character-limit/
#
Zegnat The ReCode link seems to be the original source though
#
tommorris tech journalism: reporting inconsistent bullshit rumours since INT_OVERFLOW
#
aaronpk tommorris++
#
Loqi tommorris has 87 karma
#
kylewm also, how far behind will the API be in getting this feature
#
tommorris API? Pah. That's for loser developers and their loser third-party app users.
#
Zegnat It all sounds like it will be a secondary product/service though … maybe a new try for them to become profitable? Want to post news on Twitter, need more than 140 characters, pay up.
#
Zegnat Also, would they close down API access of current services that allow for “longtweets”?
#
Zegnat Much to look forward to. Or not.
#
gRegorLove TwitterPlus
#
tommorris "current services that allow for "longtweets"" == indieweb. :|
#
tommorris also this is amazing site-death - http://blog.plug.dj/2015/09/its-time-to-say-goodbye/ - announcing shutdown and saying that it'll stop working the SAME DAY
#
Zegnat Most IndieWeb is just linking to your own site, though. While those other services are specifically made for everyone to use them to post tweets over 140 characters. And if Twitter wants to run a service like that, well, I do remember what they did to Twitter clients
#
gRegorLove Heh, plugdj still has this tweet pinned from 9/14 https://twitter.com/plugdj
#
@iNavegar I do not give Facebook permission to use my pictures ... #indieweb http://werd.io/2015/i-do-not-give-facebook-permission-to-use-my-pictures (twitter.com/_/status/648973274404274176)
#
gRegorLove What is plug.dj?
#
Loqi It looks like we don't have a page for "plug.dj" yet. Would you like to create it? http://indiewebcamp.com/s/103g
#
gRegorLove What is turntable.fm?
#
Loqi It looks like we don't have a page for "turntable.fm" yet. Would you like to create it? http://indiewebcamp.com/s/103h
shiflett joined the channel
#
gregorlove.com created /turntable.fm (+456) "prompted by gRegorLove https://indiewebcamp.com/irc/2015-09-29/line/1443563407783" (view diff)
snarfed and gavinc joined the channel
#
gregorlove.com created /plug.dj (+236) "prompted by gRegorLove https://indiewebcamp.com/irc/2015-09-29/line/1443563197542" (view diff)
#
gregorlove.com edited /site-deaths (+580) "/* 2015 */ plug.dj" (view diff)
#
snarfed hey aaronpk your site's wm processing is usually pretty fast, right?
#
snarfed curious about this one from ~3h ago: http://aaronparecki.com/webmention/DoHZTbJQ
#
aaronpk usually it's pretty much instant
#
aaronpk if not, probably means the script has crashed
#
aaronpk or, in this case, locked up
#
aaronpk ooooops
#
aaronpk this should be fun
#
snarfed heh sorry
#
aaronpk here comes a nice flood
#
aaronpk only for me luckily
#
aaronpk oh good only goes back to yesterday
#
snarfed how long was it wedged?
#
snarfed ah ok
#
aaronpk i was kind of wondering why i didn't get any webmentions today
#
snarfed i was curious to see how you'd present bridgy's new mentions
#
snarfed e.g. the one that's now on http://aaronparecki.com/notes/2015/09/29/2/spotify
#
aaronpk new..mentions?
#
aaronpk oh right!
#
aaronpk oh awesome
#
snarfed for pure mentions, my site currently just shows "X mentioned this," but ideally i'd show the full mention content if it's small enough
#
snarfed maybe i can harass acegiak/pfefferle into doing that for me :P
#
aaronpk i am on the fence about showing full text
#
aaronpk i think i might compromise and show a subset of the text for notes, and only the name for articles
#
snarfed agreed, that was basically my thought too
#
aaronpk like... as much as i appreciate tantek's really long notes that link to my posts, i don't want the full text of those showing up as a comment :)
#
snarfed definitely! hence "if it's small enough"
#
snarfed do you always just show the url right now, for mentions?
#
aaronpk yeah
#
aaronpk or wait i might show the name for articles
#
snarfed got it
#
snarfed at least you used u-url, that's nice
#
aaronpk yeah, and that's why some posts look totally broken, because my "article detection" code isn't right
#
aaronpk e.g. https://aaronparecki.com/articles/2012/07/29/1/oauth2-simplified
#
aaronpk mostly broken by proxy of the mf1 legacy parsing
emceeaich joined the channel
#
KevinMarks aaronpk: did you explictly set up twitter card permission for https://twitter.com/petterpalander/status/648931812786483201
#
@petterpalander Constraints is a good thing: http://aaronparecki.com/notes/2015/09/29/2/spotify re: @Spotify Discover Weekly (twitter.com/_/status/648931812786483201)
#
aaronpk KevinMarks: yeah, long time ago
#
KevinMarks so you already have longer that 140 char tweets
lukebrooker joined the channel
#
aaronpk the twitter card still truncate, just slightly longer than 140
cmhobbs, wolftune and nitot joined the channel
#
kylewm I'd be kind of interested to see that Civil Comments model applied to all of twitter. before you can write something, it shows you three random status updates (+ your own) and you have to rate them as constructive before they are published
#
KevinMarks oblig https://xkcd.com/810/
#
kylewm wow I didn't make the connection, they literally built the system described in this comic
sandro-, MMN-o_, michel_v_, hober2, shalkydri, Phyks, paulproteus, lordabdul, tantek, lukebroo_ and KartikPrabhu joined the channel