#indiewebcamp 2015-09-23

2015-09-23 UTC
snarfed, gavinc and lukebrooker joined the channel
#
tantek
KevinMarks++ for starting /etag - it's a good start. A step-by-step How To, and perhaps IndieWeb Examples section (is anyone programmatically setting ETag / Last-Modified headers? ) would be even better!
#
Loqi
KevinMarks has 165 karma
#
bear.im
edited /etag (+50) "create howto section and add flask"
(view diff)
#
bear.im
edited /etag (+0) "mediawiki format not markdown - silly bear"
(view diff)
#
tantek.com
edited /etag (+399) "split why, how to, linky linky, indie web examples none"
(view diff)
#
tantek
bear++ thank you! done with editing - please continue.
#
bear
your fast
#
Loqi
bear has 71 karma
#
bear.im
edited /etag (+75) "PHP"
(view diff)
#
bear.im
edited /etag (+300) "nodejs"
(view diff)
j12t joined the channel
#
bear.im
edited /etag (+52) "rework howto text and move examples to bottom of section"
(view diff)
tantek, j12t, snarfed, lukebrooker and mblaney joined the channel
#
mblaney
tantek I recently updated my feed to just be the last 24 hours
tantek joined the channel
#
mblaney
so often it will have 0 posts
#
aaronpk
hah interesting. so you're assuming consumers are polling it at least once every 24 hrs?
#
mblaney
aaronpk yes, mainly because that's part of the rssCloud spec (which I support)
#
mblaney
also I realised aggregators can deal with history
#
mblaney
also about to roll out receiving webmentions on my site! w00t!
#
Loqi
giggles
#
aaronpk
woohoo
#
mblaney
next I'm going to add h-entry support to my feed reader.
#
aaronpk
oh? what feed reader is that?
#
mblaney
my own, but it uses SimplePie.
#
tantek
mblaney++ that's a brilliant observation
#
Loqi
mblaney has 3 karma
#
tantek
I may copy that
#
mblaney
thanks tantek! :-)
#
tantek
past 24hrs of posts sounds much more intelligent than just most recent post
#
tantek
and will likely satisfy like the 99% legacy feed reader use-cases
#
mblaney
so when SimplePie fails at feed discovery I'll add the extra step of parsing the original content for microformats
#
KartikPrabhu
any reason to not show as many posts as on h-feed page?
#
tantek
KartikPrabhu: more abuse (excess polling) of Atom feeds than HTML
#
KartikPrabhu
mblaney: have you added your reader to /reader
#
tantek
based on experience
#
KartikPrabhu
tantek: documenting that would be nice
#
KartikPrabhu
and uswful
#
tantek
already did I think on /Atom
#
mblaney
KartikPrabhu no I'll do that.
#
tantek
mblaney - yes definitely document your technique!
#
tantek.com
edited /Atom (+5) "/* Tantek */ updated limits"
(view diff)
chupacab1a joined the channel
#
tantek.com
edited /Atom (+115) "/* Tantek */ noted bandwidth abuse, link to feed_file#Criticism"
(view diff)
#
mblaney
KartikPrabhu my reader doesn't fit the description on the /reader page as you would call it a "legacy" feed reader at the moment.
#
mblaney
will add it when it supports both
#
KartikPrabhu
what is reader?
#
Loqi
A reader (or indie reader) in the context of the indieweb is the portion/feature integrated into an indieweb site that provides a way to read content from other indieweb sites, possibly including posts from the current site as well https://indiewebcamp.com/reader
#
tantek.com
edited /Atom (+54) "/* Tantek */ in particular feed#Criticism"
(view diff)
#
KartikPrabhu
mblaney: aah I see.
#
mblaney
will get there soon ;-)
#
chupacab1a
hmm why is there a 1 in my name here? lol
#
tantek
mblaney - whoa https://unicyclic.com/mal/reader is very cool
#
mblaney
thanks!
#
mblaney
always happy to get feedback.
#
tantek.com
edited /RSS (+55) "IndieWeb Examples, add subheads"
(view diff)
#
tantek
mblaney - could you add a subsection for yourself in https://indiewebcamp.com/RSS#IndieWeb_Examples and maybe a sentence or two about your RSS feed only having the past 24hours of posts?
#
unicyclic.com mal
edited /RSS (+440) "/* IndieWeb Examples */"
(view diff)
#
mblaney
done, hope that's ok.
#
mblaney
I realised that I set the time to 24 hours so that the author had a chance to edit posts and they could be re-syndicated.
#
mblaney
(ie not due to any other polling assumptions)
loic_m joined the channel
#
Loqi
[mention] Brynn Evans, Deb Schultz, Christian Crumlish, Kara Murphy, Adam Rifkin, Paul Mison, Sabrina Bruning, Doc Searls, Pascale Diaine, Nicole Tollefson, Elisa Jo Harkness, Simon Law, Joël Franusic, Sarah Austin, Matthew Levine, Thor Muller, Colleen Taylor, Joichi Ito, Ben Ward, Katarzyna Babula, Beau Smith, Dan Gailey, Stacey Rivet, Bettina Warburg-Johnson, Matt Harris, Chris Heuer, Cari Levay, Zoe Schiffer, Charles Hope, Olya Lapina, Lawrence Lessig, Jesse Vincent, Zack Fischmann, Jon Pierce, Robin Andersen, Mo Kudeki, Stephanie Vacher, Faruk Ateş, Loic Le Meur, Jordan Harband, Stephen Wyatt Bush, Natalie Downe, Jessica Suttles, Erin Jo Richey, Kevin Smokler, Joseph Smarr, Paul Tarjan, Micah Snyder, Stephanie Haupt Sullivan Rewis, Jeremy Anderson, Brendan Eich, Mary Specht, Jen Bradburn, Laura Forrest, David Baron, Jet Villegas, Anselm Hook, Karen Nguyen, Sarah Dopp, Edward O'Connor, Thomas Vander Wal, Arezu Aghasey, Evan Prodromou, Yan XZ, Paul Hammond, Ian Fung, Larry Halff, Jordan McArthur, Jay Alle
#
Loqi
[mention] Pius Uzamere, Vanessa Naylon, Peter Hirshberg, Yiying Lu, Max Ogden, Shing Wong, Janet DeHart, Amy Muller, and Ann Larie Valentine were invited to https://webmention.io/notification/fvdA7omi23cuwvVkM1U0pQ
#
Loqi
[mention] Arshad Tayyeb, Eddie Codel, Monica Wilkinson, Aza Raskin, Heather Gold, Liane Baskin, Christopher Carfi, Carla Borsoi, Matt Schaar, Lizz Noonan, and Kat Lucky were invited to https://webmention.io/notification/k6PgoX8TR7DL9dIKyTL2XQ
#
Loqi
[mention] Nate Koechley, Cariwyl Hebert, Justin David Kruger, Nima Dilmaghani, Liza Sperling, Mathias Crawford, Daniel Burka, Erin Stevenson O'Connor, and Zibi Braniecki were invited to https://webmention.io/notification/HWyxJpI76_Hcfx0deU2v_A
#
Loqi
[mention] Ariel Waldman, George Kelly, Jason Shellen, Katie Johnson, Martin Atkins, Will Norris, Jeff Hodsdon, and Brittany Lee Lassiter were invited to https://webmention.io/notification/H6XmKT0AGfGC5iGezwSa5Q
#
Loqi
[mention] Ryan Barrett, PJ Khalil, Christine Herron, Zarinah Agnew, Amy MacKinnon, Lonnie Rae, Jeff Rider, Kyle Mahan, and Darius Dunlap were invited to https://webmention.io/notification/ASb51aNhk9RnjB7XZFQLPg
#
Loqi
[mention] Jay Zalowitz, Chris Messina, Brian Behlendorf, Matt Biddulph, Tom Coates, Tony Rai, Rohit Khare, Yoz Grahame, Jeffrey Veen, and Leonard Lin were invited to https://webmention.io/notification/W36hDivDXWxE4yIyNnWuBg
joskar joined the channel
#
@kylewmahan
added support for http request caching to #RedWind and #Woodwind thanks to discussion in #indiewebcamp and example code from @kevinmarks
(twitter.com/_/status/646552503786471424)
#
kylewm
snarfed: noticed thsoe RSVPs aren't backfeeding to my indie event, only to the wiki https://www.brid.gy/facebook/688447421193285#
#
kylewm
I guess because the indiewebcamp Facebook Page doesn't have my personal URL as one if its domains
#
snarfed
hey, yeah, FB is unhappy right now :/
#
snarfed
my fault, #51 etc
arlen and [snarfed] joined the channel
#
[snarfed]
oh, hmm, that could be. although i think they'd still be sent without your domain there, just as mentions instead of full rsvps. not sure
j12t joined the channel
lukebrooker, tantek and snarfed joined the channel
#
tantek.com
edited /Red_Wind (+93) "/* Features */ http request caching with citation"
(view diff)
#
tantek.com
edited /Woodwind (+81) "/* Features */ http request caching with citation"
(view diff)
cweiske, lukebrooker, friedcell, Jihaisse, eschnou, j12t, yakker, nitot, lewisnyman, scor, j12t_, petermolnar, interactivist and kerozene joined the channel
#
petermolnar
good morning
kerozene, eschnou, j12t, fkooman, LauraJ, glennjones, sammachin, sparverius, scor, loic_m and modem joined the channel
#
voxpelli
Oh, looks like http://tantek.com/ disappeared from the internet, "This web site is currently inactive. Please contact billing to reactivate your account."
friedcell, j12t, LanceyWork, kerozene, j12t_, eschnou, nitot, hs0ucy, tantek, scoates and snarfed joined the channel
#
KartikPrabhu
threaded comment UI article: http://www.elezea.com/2015/09/how-to-display-threaded-discussions-on-the-web/ for those trying to do slamntions
j12t joined the channel
#
Jeena
Looks like nobody is comming to the HWC in Gothenburg today, so I'm stopping it and going to our company event driving gocart instead too
#
KartikPrabhu
Jeena: might want to note it on the evnt page that it is cancelled
#
voxpelli
would be cool with an indieweb event "reader"/tracker like Lanyrd – where one could get a notification about such a canceled event :) with some micropub rsvp:ing and favoriting
#
KartikPrabhu
some people do event posts.
#
KartikPrabhu
what is event?
#
Loqi
An event is a type of post that in addition to a post name (event title) has a start datetime (likely end datetime), and a location https://indiewebcamp.com/event
#
voxpelli
yeah, I know, but there's no reader/tracker for them specifically :)
#
KartikPrabhu
i think if you are invited to an event post, you are supposed to get a invite webmention
#
kodfabrik.se
edited /Upcoming.org (+343) "Extending and clarifying the brainstorm"
(view diff)
#
voxpelli
KartikPrabhu: yeah, but on eg. Lanyrd I can both rsvp and track an event – no matter if I'm invited or not – and if I do then I should get updated around it and Webmentions wouldn't be enough for that
#
voxpelli
actually not really sure how one would communicate a cancellation properly – Pubsubhubbub-subscription to the details of a specific event? Unlike an h-entry one is often very interested in keeping track of what happens to an h-event
#
voxpelli
events often add more detail around the event and maybe change the time and/or location – or cancels it completely
friedcell and pfefferle joined the channel
#
pfefferle
good morning all
#
petermolnar
good morning
interactivist joined the channel
#
voxpelli
good morning pfefferle!
#
pfefferle
long time no see ;)
#
voxpelli
pfefferle: stalked you a bit on GitHub now, commenting on your comments on old issues – yay!
#
Loqi
does a happy dance!
#
voxpelli
Loqi: that's the spirit!
#
Loqi
woot!
#
pfefferle
voxpelli just read it ;)
#
voxpelli
pfefferle: do you btw support Salmentions? I'm not seeing an issue in the repo
#
voxpelli
would be pretty neat to get WordPress in on the SWAT0:ing I'm thinking :)
#
pfefferle
voxpelli not really, acegiac started an implementation, but it is very hacky and full of debug messages...
#
voxpelli
I see – as a separate project or?
#
pfefferle
s/acegiac/acegiak
#
Loqi
pfefferle meant to say: voxpelli not really, acegiak started an implementation, but it is very hacky and full of debug messages...
#
voxpelli
the testpinger is updated with Salmention-testing btw ;)
#
pfefferle
voxpelli oh cool!
#
pfefferle
voxpelli that is a motivation to work on it ;)
kerozene joined the channel
#
voxpelli
pfefferle: at least the retrieval part of getting comments of comments is in there :)
#
pfefferle
voxpelli++
#
Loqi
voxpelli has 50 karma
KartikPrabhu, interactivist_ and j12t joined the channel
#
snarfed
wordpresssalmentions++
#
Loqi
wordpresssalmentions has 1 karma
#
pfefferle
snarfed voxpelli we had a long discussion about that in 2013 http://indiewebcamp.com/irc/2013-06-23
j12t_ joined the channel
#
pfefferle
will to think about that... threaded comments makes this really complicated...
#
voxpelli
pfefferle: any specific conclusion form that in regards to wordpress or that differs to the general Salmention flow?
#
pfefferle
voxpelli not really... just keeping up with salmentions ;)
#
voxpelli
is wp threaded by default? it isn't – right?
#
voxpelli
(thinking about my Postgres craziness for supporting the threaded lists in my backend it would be "fun" to do that in WP :P)
#
voxpelli
I wonder if one can do salmentions without threaded comments=?
#
voxpelli
pfefferle: that would help a lot, right?
#
pfefferle
voxpelli yes, but I am using it... and because of "eat your own dogfood"... ;)
wolftune joined the channel
#
@jordan_web
@LukasRosenstock I think he's talking about big websites like, Forbes, TechCrunch, Wired etc. #Indieweb is important for the individual
(twitter.com/_/status/646698655987183616)
kerozene and fiatjaf joined the channel
#
fiatjaf
Loqi: !tell snarfed: the screenshot from your site at https://www.brid.gy/about shows the comments on the side of the post, but I couldn't manage to get them showing like that in your actual site. has something changed? or is my resolution that is too low?
#
Loqi
i'm sure !tell snarfed: the screenshot from my site at https://www.brid.gy/about shows the comments on the side of the post
#
fiatjaf
!tell snarfed: the screenshot from your site at https://www.brid.gy/about shows the comments on the side of the post, but I couldn't manage to get them showing like that in your actual site. has something changed? or is my resolution that is too low?
#
Loqi
Ok, I'll tell them that when I see them next
[snarfed] joined the channel
#
[snarfed]
fiatjaf: lol no, i edited the screenshot to make it lay out better
#
[snarfed]
funny you noticed!
#
Loqi
[snarfed]: fiatjaf left you a message 3 minutes ago: the screenshot from your site at https://www.brid.gy/about shows the comments on the side of the post, but I couldn't manage to get them showing like that in your actual site. has something changed? or is my resolution that is too low? http://indiewebcamp.com/irc/2015-09-23/line/1443022234266
shiflett joined the channel
#
fiatjaf
snarfed: I like the comments on the side of the bridgy screenshot. it has been a source of confusion for me over the last months. every time I see the screenshot I go to your site to see it better, and it is never there.
nitot and Geng joined the channel
#
Geng
Hi there. just to make sure. Today we still have the prewired thing to do in Codebase, right?
danlyke joined the channel
#
Geng
Er... or maybe I should ask if there is anyone online first?
kerozene and j12t joined the channel
#
@nobantu
Now is the time to register for #IIW #21 Oct 27-29! 2ks til LateReg. CUThere: http://www.internetidentityworkshop.com @idworkshop #UMA #VRM #OAuth #indieweb
(twitter.com/_/status/646714501807497217)
j12t_, friedcell, Geng, kerozene, jciv and shiflett joined the channel
#
kylewm
what is Codebase?
#
Loqi
It looks like we don't have a page for "Codebase" yet. Would you like to create it? http://indiewebcamp.com/s/103M
tantek, gRegorLove and KartikPrabhu1 joined the channel
#
tantek
good morning #indiewebcamp!
#
tantek
Jeena - sorry to hear about Göteborg HWC cancellation but understandable - no RSVPs on your indie event or POSSE copy?
#
tantek
was hoping to hear about a SWAT0 demo from you!
cweiske and tantek joined the channel
#
tantek
voxpelli: yes - site disappeared briefly yesterday and a few hours this morning due to exceeding bandwidth limits - nearly all due to bots, nearly all due to bad behaving bots, and a big % of that was excess polling of Atom feed.
cleverdevil joined the channel
#
tantek
interesting discussion about event (post) cancellations and other event info updates (name, location, description, time) !
#
tantek
we do have an existing (proposed) protocol for this for responses in general (and thus would apply to events + RSVPs etc.)
#
tantek
makes sure it's on the wiki and not just a latent undocumented brainstorm ;)
#
voxpelli
tantek: the SWAT0 will still happen, but in a while instead of now :)
#
tantek
I will add that to /event as well since it is non-obvious that this is a (hypothetically) solved problem - I don't know of any implementations yet
#
voxpelli
Oh, WebMentions again :/ So if something gets stuck in the moderation queue or rejected otherwise by publisher then no way for client to subscribe to updates
#
tantek
huh? this is a reverse webmention
#
tantek
if you send a response webmention to an original post, you should be ready to receive a webmention right back to your response from that original post when the original post is updated, WITHOUT moderation
#
voxpelli
If the original post doesn't understand/accept the response, then it won't send any updates
Geng joined the channel
#
voxpelli
It's Pubsubhubbub with implicit subscribe basically :P
#
voxpelli
And for it to work with my endpoint I would more or less have to create a PuSH-like hub that users of my endpoint can ping on updates rather than them pinging a Pubsubhubbub hub
#
tantek
I'm a little confused how this requires PuSH in any way
#
tantek
since it's easier to support receiving webmentions than subscribing to PuSH updates
#
tantek
(as evidence by more implementations and deployments thereof in the community)
#
voxpelli
We have two techs: WebMention and PuSH. One is for upstream notifications, the other for downstream subscriptions. Yet the latter is often attempted to be solved by the former.
#
tantek
webmention is for any kind of peer-to-peer where A links to B
#
tantek
the notion of up/down stream is irrelevant
snarfed joined the channel
#
tantek
PuSH is more for readers
#
tantek
where there is no linking relationship
#
voxpelli
So it should not be possible to subscribe to updates of something without interacting with it? That will force the same polling to happen as killed your site yesterday
#
tantek
no there's no polling with webmentions - that does not follow
#
tantek
also - bad bots that have no curlable presence on the web were at fault - nothing to do with visible servers
interactivist joined the channel
#
voxpelli
To receive a reverse WebMention you first have to do a regular WebMention. So unless you do, you can't receive updates.
#
tantek
that's a tautology. that's like saying, in order for the protocol to work, you have to use the protocol.
#
tantek
and sending a single webmention != polling !
#
aaronpk
i'm not really following all of this but doesn't this solve it? http://indiewebcamp.com/Webmention#Asynchronous_status_notification
#
voxpelli
tantek: the wiki page you provided said to send reverse WebMentions to all comments received – I'm just saying what it says?
#
tantek.com
edited /reply-context (+601) "/* CRUD */ applies to Events too"
(view diff)
#
tantek
you only need to send a webmention back to responses if your original changes
#
tantek
and yeah, like everything else indie web, it is based on incremental implementation
#
tantek
which is OK
#
voxpelli
tantek: how do you know what responses to send it to?
#
voxpelli
tantek: you receive responses through WebMentions, don't you?
#
tantek
all the ones that accepted
#
tantek
that *your site accepted
#
voxpelli
so unless someone has sent you a WebMention about them accepting an event, there's no way to notify them about changes to the event – unlike there would with PuSH – because the only thing needed then is that someone has subscribed to updates
#
voxpelli
Accepter -> WebMention -> Original Event -> Updated -> WebMention -> Accepter
#
aaronpk
if you're talking about getting updates to an event, that's what PuSH is for
#
voxpelli
aaronpk: yes, that's what I'm saying, but tantek seems to not agree :)
#
aaronpk
i mean it doesn't have anything to do with webmention really
#
voxpelli
aaronpk: there's a suggestion that it should: https://indiewebcamp.com/reply-context#Update
#
aaronpk
anyone should be able to come along and say "i want to know when this page has changed", and if that page supports PuSH then great, if not, it can just poll it
#
aaronpk
voxpelli: that's different
#
tantek.com
edited /event (+767) "/* Brainstorming */ Response context CRUD, including what to do if/when you cancel an event"
(view diff)
#
voxpelli
aaronpk: somewhat different, somewhat similar – one can solve the goal of the other in some cases
#
tantek
aaronpk "if you're talking about getting updates to an event, that's what PuSH is for" I disagree - you're jumping to plumbing ignoring the use-case
#
tantek
the use-cases are you've saved, RSVPd etc. to an event, and then get updates (existing real world silo example of FB)
#
tantek
and that's trivially solved with webmention per https://indiewebcamp.com/event#Response_context_CRUD without any PuSH
#
voxpelli
Use case: Privately track URL:s one are interested in, but haven't (yet) RSVP:ed to
#
aaronpk
if I save an event, I usually explicitly don't want to tell the event I saved it, it's like a private bookmark of sorts. I certainly don't want to be listed as someone who has saved it
#
tantek
PuSH is harder than webmention to implement. simpler solution preferred. QED.
#
voxpelli
s/URL:s/events/
#
Loqi
voxpelli meant to say: Use case: Privately track events one are interested in, but haven't (yet) RSVP:ed to
#
voxpelli
tantek: I disagree, I find PuSH to be easier in this case
#
aaronpk
pretty much if you start extending webmention to do this it becomes PuSH
#
tantek
voxpelli: evidence disputes your assertion of ease. more implementatations of receiving webmentions than PuSH.
#
tantek
and simpler! fewer moving parts, less code, less indirection, etc.
#
voxpelli
the salmentions added enough complexity in my endpoint, distributing updates of content itself will be a pain as it's outside of the scope of my endpoint really – it's a task for a PuSH hub
#
tantek
aaronpk - absolutely not - because PuSH has the whole "H" part
#
aaronpk
i'm like 90% sure that if you start actually explaining how to use webmention to do this you'll find out pretty quick it looks like push
#
tantek
webmention works with a simple HTML form, PuSH does not etc.
#
voxpelli
also: it's impossible to unsubscribe from WebMentions – hence WebMentions should only be used for scenarios where you can't be expected to subscribe to something
#
tantek
they are worlds apart in complexity
j12t joined the channel
#
aaronpk
and if you're going to make that arguemnt, then you are actually arguing that PuSH should just be PuS
#
aaronpk
and that the hub is an unnecessary part of PuSH
#
tantek
no S either
#
aaronpk
now what you've done is moved the state managment of keeping track of subscribers into the webmention sender, and now it's no longer a simple protocol
#
tantek
webmention is pure push
#
tantek
no subscribe, no hub
#
voxpelli
tantek: and that's a problem
#
tantek
aaronpk - you're already keeping track of all your responses on a post
#
tantek
there is no additional state management
#
tantek
and your reply-context(s) already keep track of all your in-reply-to URLs
#
tantek
seriously I don't get where you're seeing this additional complexity
#
tantek
it's still a simple protocol. when you change your post, send webmentions to everything you link to. done.
#
voxpelli
popular events can receive thousands of thousands of updates – especially if every new attendee will trigger an update
#
tantek
why is that hard?
#
aaronpk
so if I want to "save" an event and have the event send me updates, I would create a "save" post, and send a webmention saying I saved it?
#
tantek
I believe that's called a /bookmark or an RSVP of "interested"
#
tantek
'saving' sounds bookmark-like
#
aaronpk
okay, but I don't want to be listed on the event
#
voxpelli
and if you unsave it – should you delete the original post to stop updates from being received?
#
tantek
no event posts I know of list bookmarks of it
#
tantek
nor do posts in general AFAIK
#
voxpelli
I save a lot of Facebook events privately
#
tantek
"so and so bookmarked this"
#
aaronpk
but I can't really be sure i'm not going to show up in a generic "mentions" list if I send that
ttepasse joined the channel
#
tantek
you can if your bookmark is a private post perhaps?
#
aaronpk
basically I don't trust that your implementation of events is going to properly understand the "bookmark" or whatever it is and not display it as a mention
#
tantek
hmm - private reply to a public post - that could work!
#
voxpelli
and that's not more complicated than PuSH? PuSH definitely has mor eimplementations than private posts in the community
#
tantek
the right thing for the public post to do would be to keep track of the private reply (perhaps to show original author), but NOT display it to everyone who sees the event
#
tantek
voxpelli: no - very few PuSH receivers
#
tantek
but yes, private posts are still hard
#
aaronpk
i'm pretty sure there are fewer implementations of private posts than PuSH
#
voxpelli
so instead of subscribing using a PuSH-subscribe your basically subscribing by sending a "subscribe" WebMention?
#
tantek
that was more of a strawman hence the "perhaps?"
#
aaronpk
also PuSH doesn't require authentication, whereas private posts does
#
tantek
yeah - for private saving that's more of a toss-up
#
tantek
but as an publisher - e.g. I have no interest in implementing per-post PuSH updates
#
tantek
that's the problem with your proposal
#
tantek
you're asking someone to do work for you who is not benefitting
#
tantek
you've misplaced incentive with where the work must be done
#
tantek
thus, it won't get done, or it certainly won't scale
#
aaronpk
are you more willing to start handling private webmentions?
#
tantek
(scale meaning socially/investment in human time here, not a technical "scale")
#
voxpelli
doing a WebMention POST or a PuSH POST – both are as easy I think – difference is that with the WebMention POST you need to do one per response, but with the PuSH POST the hub can know who has subscribed
#
tantek
heck yeah, private webmentions are FAR more useful personally!
#
tantek
whereas ALL the work to support per post PuSH does not benefit me at all - so I won't do it
#
tantek
it fails the "scratch your own itch" test
#
tantek
so it is thus unlikely to happen
#
tantek
or certainly unlikely to take off
#
tantek
there may be some experiments, but insufficient motivation for it to get much adoption
#
voxpelli
I do my PuSH POST:s through GitHub WebHooks right now – for WebMentions updates I would have to create a similar Hub functionality somewhere to at all be able to replicate it. So much harder for me
#
tantek
voxpelli: that makes no sense. if you support webmentions at all, you're sending them upon publishing to all links in your post
#
tantek
there should be zero extra work
#
rhiaro
skimming backlog, I display bookmarks in theory but nobody has bookmarked any of my posts yet
#
voxpelli
tantek: my webmention sender have no idea about what webmentions I have received
#
tantek
doesn't matter. you link to your responses, thus you can send webmentions to them
#
voxpelli
also: My webmention sender, while still not fully built, will be based on PuSH
#
voxpelli
so my webmention sender needs to poll my webmention receiver – argh :P
#
rhiaro
And I also second the idea of being able to subscribe to / follow any post (including events) without them displaying it. But I guess I'd have to poll for updates.
#
voxpelli
another problem with WebMentions instead of PuSH: One can't detect support for such updates
#
kylewm
for subscribing to individual posts for updates, I'm concerned about the part of PuSH that requires you to periodically refresh your subscriptions
#
kylewm
that's a lot of state to manage
#
voxpelli
with PuSH you subscribe to the hub if you discover a hub and else you poll
#
tantek
kylewm: indeed. plenty of stuff like that in PuSH that is too much overhead, nevermind underspecified
#
tantek
PuSH in its current state is kind of a mess, sorry to say
#
voxpelli
then we should document those challenges and iterate on PuSH
#
tantek
webmention is simple. publishing something at a permalink, send webmentions to everything you link to. done.
#
@tmaslen
@danhind These guys are trying to help us to control our own data https://indiewebcamp.com/ would love to see the BBC get involved in it.
(twitter.com/_/status/646736987869839360)
#
voxpelli
tantek: yes, for that use case webmentions is perfect!
#
tantek
add to the CRUD, update or delete your permalink, send webmentions to everything you link to. done.
#
tantek
s/the CRUD/that CRUD support
#
Loqi
tantek meant to say: add to that CRUD support, update or delete your permalink, send webmentions to everything you link to. done.
#
tantek
voxpelli: that's not a use-case that's the entire webmention protocol!
#
kylewm
bookmarks one of rhiaro's posts
#
tantek
(vis-a-vis verification etc.)
#
rhiaro
hang tight kylewm, need to run my incoming wm wrangling script
#
tantek
with /vouch a bit more work for sure, but even then I don't think it's close to PuSH complexity / state management
#
tantek
the use-cases are receiving responses and displaying them, and then updating responses when the original changes, and lastly salmentions, updating responses when another response changes
#
voxpelli
tantek: how would you know that you don't need to poll a site for updates, because you will receive WebMention pings on all updates?
#
tantek
because you don't bother
#
tantek
if a site chooses to not send you webmentions, it's either broken or they don't want to, so you can respect that and not worry about it
#
voxpelli
well, very few sends such webmentions today, so that would mean many sites will have to be considered broken for quite a while
#
tantek
voxpelli: no, not broken per se, but just partial implementations, which is basically every indieweb site - everyone has different levels of implementations of the building blocks, different IndieMark levels, and that's ok
#
voxpelli
and those who start sending such updates will have a hard time knowing if the receiver is at all interested
#
tantek
same thing with webmentions in general - again, it's ok
#
voxpelli
eg. displaying a reply-context isn't mandatory in any way
#
tantek
nothing presentation related is
#
tantek
assumption of mandatoriness is not why this stuff works
#
tantek
this stuff works because people have incentive to provide good UI
#
tantek
making things "mandatory" typically fails because it misplaces incentives vs work to be done
#
tantek
the only thing long term scalable/maintainable/reliable thing you can do is design protocols based on incentives being aligned with work
#
tantek
no amount of "MUSTS" in a spec will make things actually work in practice'
#
voxpelli
I think we shouldn't rule out PuSH in this case and see which wins in the long run when we have had more implementation experience
#
tantek
voxpelli: sure - if you want to implement PuSH on every post - go for it - not saying not to!
#
tantek
I'm saying it's a methodologically flawed approach since it has a misalignment of incentive vs. who does the work
#
tantek
but that shouldn't stop you if you feel like doing the work yourself on your own site :)
#
voxpelli
biggest problem we have right now imho: Too few readers with too little diversity inbetween them, thus making it hard to dogfeed certain scenarios
#
tantek
what is dogfeed?
#
voxpelli
s/biggest/a big/
#
Loqi
voxpelli meant to say: a big problem we have right now imho: Too few readers with too little diversity inbetween them, thus making it hard to dogfeed certain scenarios
#
Loqi
It looks like we don't have a page for "dogfeed" yet. Would you like to create it? http://indiewebcamp.com/s/103N
#
tantek
voxpelli: you're right about not enough indie readers - still waiting to see someone implement inline reading on their own site with the post /create UI at the top
#
tantek
like 2006 era /Twitter
#
tantek
and current FB etc.
Geng joined the channel
#
rhiaro
(also your other reply to that post doesn't show up because of another bug i have)
#
voxpelli
what originally sparked this event discussion was the idea of a Lanyrd like reader for indie-events and it was in the context of that PuSH was btorught up
#
rhiaro
oh no wait it does, huh
[aaronpk] joined the channel
#
[aaronpk]
You know, at some level, PuSH is actually a lot like webmention where the webmention endpoint is on a different server
#
tantek
voxpelli: except on Lanyrd when you "save" an event it's public, you're listed as such, and it's what RSVP "interested" was designed based on
#
tantek
hence why https://indiewebcamp.com/event#Response_context_CRUD solves the Lanyrd save event use-case
#
[aaronpk]
I wonder if we could make a simpler PuSH that's based on webmention
#
tantek
aaronpk: I think Sandeep wanted to
#
tantek
hopes sandeep is doing well
#
voxpelli
tantek: hence I said Lanyrd _like_ ;)
#
kylewm
rhiaro: sorry I apparently wasn't sending webmentions to bookmarks, fixed now
#
tantek
voxpelli: I prefer to solve the real world use-case because it's been demonstrated as somethign users like/want
#
voxpelli
tantek: Never event solutions, like Facebook, has private saved events so very much a real world use case...
#
voxpelli
and I can't spell, Never = Newer
#
kylewm
I suspect a "simpler PuSH based on webmention" will end up being very very similar to PuSH
#
voxpelli
aaronpk: main difference is who the advertiser of the public endpoint is – in PuSH it's the sender, in WebMention it's the receiver
#
tantek
voxpelli: perhaps you can document how FB "private saved events" work here? http://indiewebcamp.com/event#Facebook
#
tantek
(like add a new subsection)
#
tantek
(with screenshots would be nice! - I say this having documented a bunch of the existing stuff there)
#
voxpelli
tantek: probably not the right place – Facebook allows everything to be saved privately – so not specific to events
#
tantek
really? I only saw "Save" as a button on their events
#
[aaronpk]
voxpelli not quite, imagine that you want to receive updates from a page or feed, with PuSH you look for the hub endpoint, if you were to do it with webmention you'd look for a webmention endpoint
#
loqi.me
created /Prewired (+58) "prompted by tantek and dfn added by kylewm"
(view diff)
#
voxpelli
tantek: I saved a group post earlier and have saved many pages as well
#
tantek
whoa fascinating
#
tantek
and maybe create a separate stub page for it
#
tantek
(with those examples of event, group post, pages)
#
voxpelli
[aaronpk]: if you wanted to send a "subscribe" WebMention, yes, and then that site would do a lookup on that page and find the callback URL as the WebMention endpoint of that URL – so instead of explicitly sending a callback URL, one embeds it in a "subscribe" webmention
#
kylewm.com
edited /Prewired (+49) "add one sentence summary"
(view diff)
#
[aaronpk]
With webmention subscriptions you'd basically create a URL that acts as your verification of your subscription, and it's easier to manage that because the subscription would be active as long as that URL exists
#
tantek
makes sense
#
[aaronpk]
To unsubscribe you'd delete your subscription URL and resend the webmention
#
voxpelli
[aaronpk]: so it would be like an infinite PuSH subscription that one would have to explicitly unsubscribe from – something one could make PuSH do as well, but which I imagine there has been a specific reason to not do
#
@slicknet
Considering eliminating comments from my blog. Just not seeing enough value from them these days. Thoughts?
(twitter.com/_/status/646708303217602560)
#
tantek
(note the responses)
#
aaronpk
no not infinite, because the sender would stop sending webmentions if it notices the subscription url is gone
#
tantek.com
edited /comment_problems (+111) "cite tweet for a couple of the summary problems"
(view diff)
#
voxpelli
so rather than the subscriber doing regular resubscribing, the publisher will do regular polling?
#
voxpelli
I think I rather have subscribers resuscribe than publishers polling – the ratio between publishers and subscribers tend to be that there are a lot more of the latter than the former
#
aaronpk
not even polling, because in order to send that webmention, the publisher has to *discover* the webmention endpoint *from the subscription url*
#
aaronpk
so it's built in
#
voxpelli
[aaronpk]: it only has to discover it once, then it can cache it
#
voxpelli
and expect to receive another WebMention if anything changes
#
aaronpk
that's an optimization the publisher could make
#
aaronpk
afaik only bridgy is doing that right now
#
voxpelli
I do it as well
#
rhiaro
subscribes webmention... /post? something like http://rhiaro.co.uk/2015/04/1430099340
#
voxpelli
in my Salmentioning
#
aaronpk
if the publisher is going to make that optimization, it would have to account for edge cases created by that optimization
#
voxpelli
ok, so the publisher should expect to be polling
#
rhiaro
But I don't necessarily want to delete that post to unsubscribe. I'd rather make an unsubscribe post, for record-keeping..
#
aaronpk
no it's not polling
#
voxpelli
I don't see why that is better than doing a resubscribe every X time where X is a well communicated time frame that all parties know about
#
aaronpk
"is a well communicated time" is false though
#
voxpelli
[aaronpk]: fine: not polling, "rediscovery" of webmention endpoint
#
voxpelli
[aaronpk]: hub.lease_seconds is a requirement in a PuSH Hub's verification: http://pubsubhubbub.github.io/PubSubHubbub/pubsubhubbub-core-0.4.html#verifysub
#
tantek
as opposed the thing kylewm mentioned about PuSH reverifying subscribers every so often? just slower polling?
eschnou joined the channel
#
aaronpk
PuSH subscriptions expire at a time decided by the publisher, so yes it is effectively slower polling
#
tantek
btw re: feeds and stuff
#
tantek
mblaney++ for documenting his very clever minimal feed approach: https://indiewebcamp.com/RSS#Malcolm_Blaney
#
Loqi
mblaney has 4 karma
#
tantek
I'm thinking of implementing that + 1 feed item minimum
#
tantek
thus allowing subscribers to *only* poll once a day
#
voxpelli
[aaronpk]: the subscriber can request a specific hub.lease_seconds, but the publisher can choose to ignore it, but will always communicate the time frame
#
aaronpk
basically waht i'm saying is that's not much of an optimization and adds complexity on both sides
#
voxpelli
to know whether you need to resubscribe every day/hour/week?
#
aaronpk
rhiaro: re: wanting to create an ubsubscribe post, I think that'd be fine too. it would need a reference back to the original subscription URL
#
kylewm
i thought there was some minimum for what the lease would be?
#
voxpelli
kylewm: didn't see any when skimming through the specs – feels like something the spec shouldn't care about
#
aaronpk
i guess i don't see the problem of having to discover the webmention endpoint when you go send another webmention
#
aaronpk
seems easier than caching it!
#
aaronpk
since you know, you don't have to do anything extra
#
aaronpk
also fun side effect: subscribers can change their webmention endpoint without having to re-subscribe
#
kylewm
voxpelli: you're right...don't know what i was thinking of, maybe just the in practice minimum of the three 0.4 hubs we use
#
snarfed
aaronpk: for small services and individual sites, sure, caching is more code/complexity and not much value
#
voxpelli
[aaronpk]: my current Salmention implementation made the caching case pretty apparent to me :P
hs0ucy joined the channel
#
snarfed
for high volume services like bridgy, the optimization matters
#
aaronpk
sure, and my point is that if you're going to optimize, you need to also handle the edge cases created by that optimization
#
snarfed
sure! of course.
#
aaronpk
point is you *can* optimize if you want, but you don't have to
#
aaronpk
whereas PuSH requires a certain level of complexity off the bat
#
KevinMarks
snarfed: would If-Last-Modified help with the facebook-atom duplication problem?
#
snarfed
sorry, i jumped into the conversation without context :P
#
gRegorLove
(Scrolling back) A subscribe and unsubscribe could still be the same URL. Have it return HTTP 410 when it's unsubscribed, and set the updated date to the unsubscribe date.
#
snarfed
KevinMarks: huh. maybe? honestly i don't fully understand even the problem yet. i haven't dug in; been focusing on https://github.com/snarfed/bridgy/issues/456 and #51
#
snarfed
i'll switch back to fb-atom soon
#
aaronpk
gRegorLove: good point
#
KevinMarks
kylewm has implemented Last-Modified handling in woodwind
#
aaronpk
when I said "delete" i actually meant tombstone with a 410
#
KevinMarks
so if you send the header he will echo it back to you on next fetch
#
snarfed
KevinMarks: it definitely could help. jump in and send a PR? :P
#
tantek
aaronpk - yeah that's what /delete means :)
#
gRegorLove
Right, just to capture voxpelli's desired history tracking
#
aaronpk
that was rhiaro but yeah
#
voxpelli
yeah, I'm just saying PuSH twice a sentence
#
voxpelli
kylewm: yeah, as all the current hubs have the same use case they probably have all set similar minimum times (wonder if they have a shared max as well?)
#
KevinMarks
the point of the /etag conversation was to help minimise polling
#
KevinMarks
or rather, polling's overheads
#
aaronpk
now i'm tempted to write up a spec for subscribing to things via webmention
#
voxpelli
oh dear :P
#
aaronpk
would be curious to see if people are less intimidated by it than they are by PuSH
#
voxpelli
can't wait for the fun of being forced to become the PuSH-replacement of KevinMarks and others...
#
aaronpk
I even wrote this up, but it doesn't seem to have helped it take off much more here http://indiewebcamp.com/how-to-push
#
voxpelli
aaronpk: it layers a lot of stuff on top of WebMentions which makes it increasingly hard to implement a full WebMention endpoint
#
voxpelli
it was simple before Salmentions, then Salmentions complicated things and this will add even more and then we have the Vouch things etc etc – eventually it will be too much to selfdogfood
#
KevinMarks
my cunning plan is to eventually make voxpelli manage my annotations web-wide
#
aaronpk
really? I think it's mostly just doing the same things
#
KevinMarks
as I can webmention anything via voxpelli's app, and it will keep them for them until they decide to ask
#
voxpelli
aaronpk: Salmentions made my endpoint much much more complex – from one HTTP request to 3-4 per ping
#
aaronpk
i meant the subscription part. i also haven't implemented salmentions yet
#
voxpelli
aaronpk: the subscription part means I would have to start accepting pings for content updates – like any PuSH Hub would
#
aaronpk
but you should be doing that already for regular webmentions
#
voxpelli
I don't care about the contents of a webpage until I'm supposed to deliver a salmention to the pages it mentions
#
snarfed
not to derail the conversation, but i'm going to take advantage of you all since you're here and use you as guinea pigs to test out https://github.com/snarfed/bridgy/issues/456
#
snarfed
i think it's finally ready, and not falling over :P
#
voxpelli
[aaronpk]: do people ping webmention.io every time they update a post that receives comment through tha?
#
aaronpk
webmention.io also hasn't implemented salmentions if that's what you're asking
#
snarfed
the top two responses on https://www.brid.gy/twitter/kylewmahan#responses are good examples: bridgy searched twitter for kylewm.com and found and backfed mentions from other people
#
snarfed
(ie not just responses to kylewm's original tweets)
#
voxpelli
aaronpk: but you said "but you should be doing that already for regular webmentions"
#
aaronpk
i was talking about sending webmentions when you've updated the post
#
voxpelli
aaronpk: so why should KevinMarks be notifying me of his new posts today when all I'm doing is receiving comments on his behalf?
#
aaronpk
i'm getting lost. the term "me" is getting used to mean both your personal site and your webmention service, and i can't follow
#
voxpelli
I only receive comments, the WebMentions sending is typically out of scope of my endpoint (although salmentions forced me to do add it partly – and with me I mean my endpoint)
#
voxpelli
and thinking about it I should probably also defer the salmention-pinging to an external service
Geng joined the channel
#
snarfed
pulled the trigger, hold on to your butts...
#
snarfed
oh dear don't do that
#
snarfed
no good will come of it
#
snarfed
actually #51 is worse
#
snarfed
51 + 456 turned into way way way more work than i expected
#
tantek
waits for KevinMarks to complain about more spam or troll "replies" to his posts
#
snarfed
ah well!
#
ben_thatmustbeme
snarfed++ for Jurassic Park reference
#
Loqi
snarfed has 141 karma
#
tantek
thinks this is snarfed subtle way of nudging /block implementations
#
KevinMarks
should I write something inflammatory and see what happens?
#
voxpelli
KevinMarks: did you ever sort them out? ever really heard about whether you wanted me to remove any?
#
snarfed
lol KevinMarks sure!
#
snarfed
welcomes stress tests
KartikPrabhu joined the channel
#
snarfed
tantek: oh man, hopefully not! i still don't plan to prioritize it in bridgy itself at least
#
snarfed
we'll see
#
@jordan_web
@LukasRosenstock Perhaps, but small groups & individuals have a more focused following so can survive in an #indieweb environment, - I hope.
(twitter.com/_/status/646751018852687873)
#
KevinMarks
well, I only have a couple on my main site, as I didn't POSSE the most comment-provoking tweet
#
tantek
KevinMarks, no need to be inflammatory, just thank the XOXOFest folks for how well they sorted/filtered out all the trolls, especially in person and in Slack, e.g. (insert your favorite GG hashtag here)
#
voxpelli
snarfed: did any of you had any feedback on the approach proposed in https://github.com/snarfed/bridgy/issues/473 though? might be that I implement it before you do I'm thinking
#
@kevinmarks
When @femfreq speaks it is impossible to summarise her, because she speaks in perfect paragraphs that go from analysis to action #xoxofest
(twitter.com/_/status/642927875465719809)
#
snarfed
hey voxpelli i noticed your twitter acct isn't signed up for bridgy. intentional?
#
snarfed
totally fine if so, just curious
#
snarfed
ah, 473! sorry, honestly i haven't tried to fully grok the proposal there
#
KevinMarks
If we want to get serious about stress testing, Christina Warren is interested in indieweb
#
snarfed
it was pretty deep
#
snarfed
unlikely that we'll do it in bridgy any time soon, so you probably would get there before us
#
voxpelli
snarfed: yes, because I don't fully POSSE yet, so it misinterpreted URL:s and stuff – hence my ticket about wishing to lock things down to just those URL:s with a explicit u-syndication
#
snarfed
KevinMarks: go for it!
#
snarfed
ahhh right
#
voxpelli
snarfed: the proposal was basically: Don't do anything hard on your part, let me do all the hard stuff for you ;)
#
snarfed
voxpelli: #51 actually gets halfway there; many things now will only be sent as mentions, not actual replies/likes/etc
#
snarfed
voxpelli: lol ok. i like that spirit! i'll actually read and reply soon.
#
KevinMarks
hm, so I'd need a public block list for voxpelli to see to filter
#
voxpelli
KevinMarks: well, public or private, so proposal to snarfed was to export your Twitter block list as a private block list you could link to from your profile and/or import directly into my endpoint
#
tantek
KevinMarks: nah, you just need some sort of block UI. whether that's with a list, public or not - those are details.
#
tantek
can an app that has Twitter read-only OAuth into your account read your Twitter blocklist?
#
voxpelli
I don't want to save such data in my endpoint – that would make me silo:ish :P
#
tantek
Delegate all the data!
#
tantek
Loqi ^^^
#
Loqi
dude
#
tantek
Loqi, seriously
#
KevinMarks
you want #social loqi for that tantek
#
KevinMarks
as twitter is the source of the webmentions, it makes sense to keep the blocklist there too
#
tantek
gives Loqi a memeworthy pattern.
#
Loqi
runs away from the memeworthy pattern.
#
tantek
nice!
#
tantek
what is a block list?
#
Loqi
It looks like we don't have a page for "block list" yet. Would you like to create it? http://indiewebcamp.com/s/103Q
#
voxpelli
yeah, that format hasn't been definied
#
snarfed
ok bridgy #51 (smarter posse vs mention detection) and #456 (search all tweets for mentions of your site) are now on for everyone, with hopefully final logic for both
#
voxpelli
was discussing a reverse of XFN
#
tantek
voxpelli: it's not a format, it's a concept
#
snarfed
everyone, please ping me if you see any anything weird with your bridgy accounts. thanks in advance!
#
tantek
think user-centric - what does a user think of as a block list?
#
aaronpk
fun times
#
voxpelli
tantek: a follow list, like a blog roll/XFN, is a whitelist – eg /Vouch – and a block list is a blacklist that's the opposite of that
#
tantek
A block list is a list of accounts that a user has blocked on a service, like [[Flickr]] https://www.flickr.com/people/tantek/contacts/ignore/ and [[Twitter]] https://dev.twitter.com/rest/reference/get/blocks/list .
#
loqi.me
created /block_list (+240) "prompted by tantek https://indiewebcamp.com/irc/2015-09-23/line/1443033705563 and dfn added by tantek"
(view diff)
#
tantek
voxpelli: again, that's too plumbing-centric a definition / viewpoint.
#
tantek
a follow list or a blog roll *can* be used as a whitelist for some particular use-case, like Vouch but doesn't have to be either.
#
tantek
Flickr even calls it "Your Block List" literally in the UI
#
tantek
hence the term is preferred over something like "blacklist"
#
voxpelli
whitelist as in "people I prefer to see more from", blacklist as in "people I prefer to see less from"
#
KevinMarks
ah, bridgy doesn't know that that tweet was a POSSE
#
voxpelli
and I meant it from the perspective of the actual format being XFN like and that the things are two sides of the same coin, rather than as a kickass definition
#
snarfed
KevinMarks: if you add known.kevinmarks.com to your twitter profile, it will
#
KevinMarks
I cna just do that in the text, right?
#
snarfed
i'll add all that to the docs
#
snarfed
(oh KevinMarks you'll need to re-auth twitter on bridgy afterward to make it pick up the new domain)
#
tantek.com
edited /block_list (+724) "why, how to, indieweb examples none, silo examples section expand with prose"
(view diff)
#
tantek
challenge thrown down: https://indiewebcamp.com/block_list#IndieWeb_Examples "None so far." - who will be the first?
#
snarfed
KevinMarks: right, it's on the front page
#
tantek
voxpelli: that's the thing they are not two sides of the same coin
#
tantek
that's a super-inaccurate oversimplified artificial dichotomy
#
tantek
when you compare the full range of use-cases from /report_abuse to /block to /mute to /follow etc. there is a big spectrum
#
tantek
not to mention close friends etc. - concepts that FB has deployed and people apparently use
#
kodfabrik.se
edited /block_list (+194) "Referencing implementation discussions"
(view diff)
#
KevinMarks
a "changed profile? update now" link on that page would be a bandaid
#
KevinMarks
so it still says "No post links found"
#
aaronpk
low tech solutions ftw
#
voxpelli
tantek: and XFN captures all of the positive relations of the positive side of coin, so the block list / mute / blacklist counterpart would capture the negative relations on the negative side of the coin
#
KevinMarks
do I have to use bridgy publish to let it know?
#
tantek
voxpelli: XFN does not capture all the positive relations
#
tantek
e.g. FB's "close friends" is not in XFN
#
snarfed
KevinMarks: no
#
voxpelli
s/all/many/
#
Loqi
voxpelli meant to say: tantek: and XFN captures many of the positive relations of the positive side of coin, so the block list / mute / blacklist counterpart would capture the negative relations on the negative side of the coin
#
snarfed
thinks through the current logic...
#
KevinMarks
I need to make you fetch the post somehow
#
tantek
also "mute" is typically more neutral than positive or negative
j12t joined the channel
#
tantek
e.g. on twitter, if you mute someone, you can still get DMs from them
#
voxpelli
tantek: is that because XFN is fundamentally broken or because no-one has felt a need to define that yet?
#
tantek
so it isn't all positive or all negative
#
tantek
hence disproving the coin dichotomy
#
tommorris
I can't see myself ever publishing my block list.
#
aaronpk
also muting implies not telling the other person they are muted, whereas block shows them they are blocked
#
tantek
voxpelli: XFN has been (deliberately) slow to evolve
#
aaronpk
tommorris: agreed, i am never going to publish my block list, probably also not my mute list
#
voxpelli
I think you're misunderstanding me – the coin wasn't the core part of what I was suggesting – but whatever
#
tantek
if you want to add to it - there's a bunch of ideas here: http://microformats.org/wiki/xfn-brainstorming
#
tantek
voxpelli: point is that framing matters, and using a dichotomy like coin is more misleading than helpful
#
tantek
it's a spectrum at a minimum
#
tantek
may even be multidimensional
#
voxpelli
lets rephrase it then: XFN has some cool relations (evidently not all). To make it useful in relation to block lists it could be extended with some more.
#
tantek
it could be yes. in other ways I've grown to be skeptical about putting more human meaning like that into (mostly invisible) rel attributes
#
tantek
worthy of documenting brainstorming / experimentation especially with use-cases
#
voxpelli
tantek: remember one thing: That framing something so that it accurately reflects the message that one wants to convey is something that comes with the skills of a language. So we who are not native english speakers might sometimes make mistakes there and it's kind of frustrating when you then focus on that.
#
tommorris
(funnily enough, today I've been getting my head around how to do friends only/private posts.)
#
tantek
voxpelli you are absolutely correct and I apologize.
#
voxpelli
tantek: apology accepted :)
#
tantek
voxpelli: TBH your english skills are certainly good enough to easily fool native speakers into thinking you are native speaker.
#
tantek
so despite knowing you're not, it is easy to forget :)
#
tantek
*a native speaker
#
voxpelli
:) I guess it becomes mostly apparent when it comes to figure of speeches and similar
scoates joined the channel
#
snarfed
KevinMarks: still investigating. thanks for the test case. :P
Geng joined the channel
#
KevinMarks
I should add that to my wiki page: "want edge cases? ask Kevin"
#
Loqi
definitely
#
tantek
huh wat
#
tantek
does that mean they invited others?
#
tantek
time to experiment with an RSVP to an Eventbrite event
#
snarfed
changes in those invites may be due to the new posse vs mention logic (https://github.com/snarfed/bridgy/issues/51)
#
snarfed
e.g. kylewm may no longer gets rsvps to his event posts - or just as mentions - since the IWC FB page doesn't have kylewm.com in its profile
#
snarfed
(haven't fully investigated that)
#
tantek
snarfed - wondering if the IWC FB event page should be considered a POSSE copy of the IWC wiki event page
#
tantek
since the wiki event page can receive webmentions
#
tantek
how would the IWC wiki sign-up to get Bridgy webmentions from a FB POSSE copy?
#
snarfed
i believe the new logic should consider the event a POSSE, and the wiki is already "signed up" through the IWC FB page
#
snarfed
not sure what's going on with the invites yet
#
snarfed
i'm going to add "Indie events posted at kylewm.com." to the page's profile
#
tantek
sure!
#
tantek
(assuming kylewm is up for that responsibility ;) )
#
kylewm
go for it!
j12t joined the channel
#
snarfed
actually more work to do for this than i thought. https://github.com/snarfed/bridgy/issues/483
interactivist joined the channel
#
tantek
huh no Bridgy issue for that
#
tantek
nor wiki guidance ...
#
tantek.com
edited /event (+38) "/* POSSE */ see also RSVP POSSE"
(view diff)
kerozene and loic_m joined the channel
#
kylewm
does anyone who POSSEs likes to Twitter show a u-syndication url on the original?
#
kylewm
annoying bug in redwind caused by u-syndication = the URL of the liked post instead of the URL of the like itself (because likes in twitter don't have a url of course)
#
kylewm
trying to decide between adding a made up #fragment like bridgy, or just leaving off the syndication url
camerongray and gavinc joined the channel
#
tantek.com
edited /rsvp (+3356) "add brainstorm POSSE similar to /event#POSSE"
(view diff)
#
tantek
indie->silo RSVP exploration: indie RSVP to an Evenbrite event http://tantek.com/2015/266/t1/science-hack-day-san-francisco-sign-up manually POSSEd to the Eventbrite event, but automatically POSSEd to Twitter. Documenting thinking here: https://indiewebcamp.com/rsvp#POSSE
#
tantek
and related Bridgy Publish request to potentially automate POSSEing RSVPs to Eventbrite https://github.com/snarfed/bridgy/issues/484 - feel free to comment or add your indie RSVP examples to Eventbrite events (on either the github issue or https://indiewebcamp.com/rsvp#POSSE wherever you think is best!)
#
snarfed
tantek++
#
Loqi
tantek has 242 karma
#
tantek
also, yay science!
#
tantek
if you enjoy IndieWebCamp and hacking/building things - you'll be delighted at Science Hack Day too - strongly encourage you to check it out if you can make it out to SF 2015-10-24..25.
j12t and kerozene joined the channel
#
snarfed
!tell KevinMarks ok the constant dupes in https://facebook-atom.appspot.com/ are annoying me now too. feel free to follow https://github.com/snarfed/facebook-atom/issues/11 for progress
#
Loqi
Ok, I'll tell them that when I see them next
#
tantek
braindumps a bit of how non-trivial it might be to automate POSSEing RSVPs to Eventbrite, in aforementioned Bridgy issue 484.
#
tantek
staying optimistic that someone (more) clever may figure something out :)
#
snarfed
tantek: is there an eventbrite rsvp api? mind adding a doc link to that issue if so?
#
tantek
oh they call them "Orders"
kerozene, friedcell and shiflett joined the channel
#
tantek
snarfed, bad news :(
#
snarfed
ah well. thanks for documenting
#
tantek
going to ping an Eventbrite contact
j12t joined the channel
#
tantek
so weird. FB lets us RSVP via API, but not create events. Eventbrite lets us create events, but not RSVP.
#
tantek
(via API)
#
gRegorLove
That's surprising. I thought Eventbrite had a really robust API
kerozene joined the channel
#
gRegorLove
Wait, getting ExactTarget confused with Eventbrite
#
gRegorLove
Don't mind me
#
tantek
what is ExactTarget?
#
Loqi
It looks like we don't have a page for "ExactTarget" yet. Would you like to create it? http://indiewebcamp.com/s/103R
#
gRegorLove
Probably not wiki worthy. "Salesforce Marketing Cloud (formerly known as ExactTarget) is a provider of digital marketing automation and analytics software and services."
#
tantek
huh - yeah not unless someone here is using them for their Indieweb site?
#
gRegorLove
What is Eventbrite?
#
Loqi
It looks like we don't have a page for "Eventbrite" yet. Would you like to create it? http://indiewebcamp.com/s/103S
#
tantek
Eventbrite is an [[event]] hosting silo that can be used for free and paid events. https://www.eventbrite.com/
#
loqi.me
created /Eventbrite (+137) "prompted by gRegorLove https://indiewebcamp.com/irc/2015-09-23/line/1443043105756 and dfn added by tantek"
(view diff)
KevinMarks and kerozene joined the channel
#
tommorris
reads scrollback and sees mention of Salesforce and "Marketing Cloud" and winces.
#
tommorris
reaches for painkillers.
[kevinmarks] joined the channel
#
[kevinmarks]
interesting hack I just saw in another slack
#
Loqi
[kevinmarks]: snarfed left you a message 56 minutes ago: ok the constant dupes in https://facebook-atom.appspot.com/ are annoying me now too. feel free to follow https://github.com/snarfed/facebook-atom/issues/11 for progress http://indiewebcamp.com/irc/2015-09-23/line/1443041101775
#
[kevinmarks]
create a google calendar for an arbitrarylongstring@gmail.com account
#
[kevinmarks]
then you can add it in google calendar as a 'friend's calendar'
#
[kevinmarks]
and post to it by inviting the email address
kerozene joined the channel
#
[kevinmarks]
I suppose an ics <-> hfeed would make sense too
#
[kevinmarks]
I wonder if I know anyone who works on google calendar
j12t joined the channel
#
gRegorLove
Sorry for the trauma, tommorris
scoates, arlen, kerozene, aaronpk_ and ben_thatmustbeme joined the channel
kerozene, j12t and [kevinmarks] joined the channel
#
[kevinmarks]
Just realised we have a HWC night on the day Marty, Doc and Jennifer arrive
lukebrooker, kerozene, KevinMarks_ and j12t joined the channel
#
tantek.com
edited /events/2015-10-21-homebrew-website-club (+72) "/* Notes */ aside per kevinmarks in IRC with wikipedia citation"
(view diff)
KevinMarks, KevinMarks_, j12t and j12t_ joined the channel
#
snarfed
when a URL redirects, is there standard terminology for the before vs after URLs?
#
snarfed
original/final are good, but i'm writing original post detection code, so "original" conflicts
kerozene joined the channel
#
snarfed
source/target similarly conflict w/webmention params
#
snarfed
after/before are meh but at least conflict free
sparverius, lukebroo_ and tantek joined the channel