#indiewebcamp 2015-03-05

2015-03-05 UTC
#
KevinMarks
hm, kyle's RSVP markup is a bit repetitive
#
kylewm
KevinMarks: ohh, how can i fix it?
#
KevinMarks
it says Kevin Marks (www.facebook.com) is going is attending.
#
KevinMarks
looks like you're hitting the bridgy p-name thing again
KevinMarks_ joined the channel
#
kylewm
hmm...yeah the p-name is a good fallback if you just parsed them as regular replies
#
kylewm
Ben Roberts Attending is attending.
#
KevinMarks
so, same problem
#
kylewm
Known just says "Kyle Mahan RSVPed yes"
#
tantek
which example shall we analyze first?
#
aaronpk
I display the green "attending" banner when I see the "rsvp=yes" property
#
tantek
KevinMarks all my RSVPs have been POSSEd to Twitter since the start ;)
#
aaronpk
I then show whatever is in the p-name because presumably people write little comments
#
tantek
(of when I started posting RSVPs)
#
tantek
kylewm: what's your permalink?
#
KevinMarks
it was what we had with likes on webmention.herokuapp.com
#
KevinMarks
so is this bridgy generating the text again?
#
aaronpk
yes bridgy is generating "is attending" text, which is actually good fallback text
#
tantek
KevinMarks: unlikely - those would be from FB
#
KevinMarks
do we each need to mung out the bridgy-made text if it is "is attending" etc?
#
@_crossdiver
Re-built my personal #indieweb site again. Hexo for Markdown compiling into static HTML, Amazoing S3 for hosting. http://ike.io
(twitter.com/_/status/573273756640022528)
#
aaronpk
so how can I know whether the text is fallback or user-entered?
#
tantek
KevinMarks: that should not be necessary
#
snarfed
if it's from bridgy, it's never user-entered
#
tantek
what is Hexo Markdown?
#
aaronpk
without hard-coding things to bridgy
#
KevinMarks
should bridgy mark it with p-fallback or something?
#
tantek
no no no that should be completely unnecessary!
#
aaronpk
wow no
#
tantek
KevinMarks - too much markup, make less.
#
KevinMarks
bridgy is the one creating the text
#
KevinMarks
this happened before
#
aaronpk
the problem is the overloading of p-name (and also p-content)
#
tantek
hmm - I'm not convinced of that
#
aaronpk
we need good fallback text in the p-content and p-name of special post types for consumers that don't know about those types.
#
aaronpk
for example my bike rides have p-content of "Biked 2.1 miles in 14 minutes" http://aaron.pk/m4_ye2Q
#
tantek
we need to look at such special post types on a case-by-case basis
#
aaronpk
so if that shows up in a reader which ignores all the other properties, it'll still look reasonable
#
aaronpk
(actually it also has a photo of the map because I went a little overboard)
#
aaronpk
so the question is then what should it do if I enter text myself in that bike ride?
#
aaronpk
looking for an example
#
KevinMarks
right, thats the point - we need to distinguish between generated fallback and actual comments
#
tantek
kylewm: was this RSVP related? your RSVP?
#
kylewm
sorry what do you mean?
#
kylewm
Hexo is a static site generator
#
kylewm
not sure what Hexo for Markdown is
#
aaronpk
http://aaron.pk/m4YES3L right now my site removes the autogenerated text
#
tantek
you said: kylewm: KevinMarks: ohh, how can i fix it?
#
kylewm
ironically i was more excited about fixing the markup than the copy
#
tantek
kylewm: I always have a hard time finding your rsvps from your home page :)
#
tantek
kylewm: oh this is about other people's RSVPs
#
kylewm
tantek: ah that's a good note
#
tantek
ok we can start with the bridgy example
#
tantek
looks at the raw markup
#
tantek
in the bridgy case, since FB RSVPs have no user-authored text, it makes sense to figure out a way to omit any display text if the only information is the RSVP status
#
tantek
ok first things first we don't need e-content for sure
#
aaronpk
tantek: but for consumers that don't recognize rsvps, it should still show up as a regular reply
#
tantek
aaronpk yes hence fallback
#
aaronpk
okay, will be curious how you suggest this
#
tantek
aaronpk, we must of course start with, what is the desired UX?
alanpearce joined the channel
#
tantek
which has the follow-up of
#
tantek
for RSVP aware consuming code? or RSVP unaware consuming code?
#
tantek
let's take the aware use-case first
#
tantek
what is an rsvp
#
Loqi
An RSVP is a type of post that is a reply to an event post https://indiewebcamp.com/rsvp
#
tantek
nothing about how to present an rsvp
#
tantek
what is an 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
#
KevinMarks
so the bridgy markup does say it twice
#
tantek
KevinMarks you are jumping to plumbing conclusions
#
tantek
without first defining the desired UX outcome
#
KevinMarks
in pure text terms, the bridgy post is duplicative
#
tantek
are you making a storage judgment?
#
KevinMarks
no wonder everyone else gets confused
#
KevinMarks
no, a "what si the text UX we want" one
#
KevinMarks
without parsing any markup, bridy's relected post say "Tantek Çelik is attending. Tantek Çelik is attending."
#
KevinMarks
in fact, if you read the alt text too (as as creenreader would) it says "Tantek Çelik is attending. Tantek Çelik Tantek Çelik is attending."
#
kylewm
the bridgy ui is what you get when you generate microformats html *from* microformats json ... probably not the ideal way to design for user consumption :)
#
kylewm
s/user/human
#
Loqi
kylewm meant to say: the bridgy ui is what you get when you generate microformats html *from* microformats json ... probably not the ideal way to design for human consumption :)
#
KevinMarks
but you and aaron are parsing that HTML
KartikPrabhu joined the channel
#
KevinMarks
bridgy is generating json first then html?
#
KevinMarks
that may be the issue
#
kylewm
KevinMarks: yeah, it converts facebook to activitystreams json to microformats json to microformats html
#
tantek
KevinMarks do you have a URL to how you think RSVP consuming code would ideally display RSVPs?
#
tantek
desperately tries to pull the conversation out from the plumbing depths.
#
KevinMarks
the plumbing is causing the bad markup
#
KevinMarks
s/kye/kylewm/
#
Loqi
KevinMarks meant to say: bridgy sends a webmention to kylewm pointing to this: https://brid-gy.appspot.com/rsvp/facebook/11500459/444025702419735/214611
#
tantek
KevinMarks, it's hard to get to where you want to go if you don't know where that is.
#
tantek.com
edited /event (+688) "how to (sub)head fixes, add why section"
(view diff)
#
tantek
s/so//
#
KevinMarks
shoudl present that wm as ' Tantek Çelik (www.facebook.com) is going'
#
KevinMarks
which is presented with text and links
j12t joined the channel
#
kylewm
so what if someone rsvp'd "I'm <data class="p-rsvp" value="yes">attending</data> HWC and hope to get help setting up my website!"
#
kylewm
I would prefer to show both the "attending" bit and the additional comment
#
KevinMarks
right, which is fine
#
KevinMarks
the problme is that bridgy says it redundantly
#
kylewm
for an implementation that does not recognize RSVPs as special, I would want that to show "kylewm: I'm attending HWC and hope to get help setting up my website!"
#
kylewm
for an implementation that does recognize them as special, "kylewm: [[RSVPed Yes]] I hope to get help setting up my website!
#
KevinMarks
that would require NLP
#
kylewm
hehe, I know
#
kylewm
but is there some way to use the p-rsvp span to "block out" the redundant section
#
kylewm
so that fallbacks would use the whole p-name, but rsvp-consuming implementations could get just the non-p-rsvp part?
#
kylewm
does that make sense?
#
KevinMarks
if it said kylewm is going<cr>I'm attending HWC and hope to get help setting up my website!
#
KevinMarks
that would eb OK
#
kylewm
but that's what is says now?
#
kylewm
is going<br>is attending
#
KevinMarks
yes, I am agreeing that your site si doing the right thing
#
KevinMarks
it is bridgy that is making crappy RSVPs
#
tantek
I still don't see how you can conclude "the right thing" without knowing what presentation you are aiming for.
#
kylewm
tantek: I tried to answer your presentation question
#
KevinMarks
I am aiming for a skimmable list of responses wiht optional comments, just as kyle, aaron and known do now
#
tantek
KevinMarks - can you write that up?
#
KevinMarks
except that bridgy is generating redundant comments that don't help
#
tantek.com
edited /event (+61) "brainstorming"
(view diff)
#
tantek
KevinMarks - that's not state of the art for RSVP displaying
#
tantek
see existing documented research on this: https://indiewebcamp.com/event#Attendees
#
tantek.com
edited /event (+5) "/* Attendees */ More screenshots needed"
(view diff)
#
tantek
"skimmable list of responses" is more like an undesigned approach
#
tantek
lack of RSVP-specific design
#
tantek
kylewm - will try to sift through the plumbing talk
#
kylewm
tantek: haha fair enough, this line and the next: http://indiewebcamp.com/irc/2015-03-04#t1425515628213
#
KevinMarks
even in a non-rsvp specific design I think the bridgy post will still say things twice
#
tantek
I think the problem is in the approach of having your RSVP aware consuming code generate text to display
#
kylewm
i.e. the reply is trying to be a reply and an rsvp, when it should really be one or the other?
#
kylewm
I think that's probably the answer
#
kylewm
they are certainly separate on Facebook
#
tantek
I believe there is precedent for this
#
tantek
feels like the analysis we did for likes
#
tantek
where we decided that u-like-of is sufficient
#
tantek
an u-in-reply-to was unnessary
#
aaronpk
actually an analogous situation woult be if an h-entry has u-like-of then it would not have a p-name or p-content
#
tantek
aaronpk - not at all - it would still have p-name for fallback
#
snarfed
fallback++
#
tantek
and everything has a p-name
#
Loqi
fallback has 1 karma
#
tantek
whether explicit or implicit
#
aaronpk
but I thought we were talking about the problem of not knowing whether it's important to display the name
#
tantek
in the case of a like, we do know
#
tantek
if there's a u-like-of, no need to display p-name
#
tantek
this is why I said it's case-by-case
#
aaronpk
that's because we've decided that "likes" don't have additional content like text
#
tantek
and try to generalize it is at best premature, at worst architecture astronomy
#
aaronpk
but we explicitly do want rsvps to be able to contain text comments
#
tantek
s/generalize it/generalize it to all "special" post types
#
Loqi
tantek meant to say: and try to generalize it to all "special" post types is at best premature, at worst architecture astronomy
#
tantek
aaronpk - right! and the way we do explict text responses is … in-reply-to
#
kylewm
aaronpk: "we explicitly do want rsvps to be able to contain text comments" -- do we? i think maybe we don't
#
aaronpk
i think most examples of indie rsvps contain a comment
#
aaronpk
and facebook certainly encourages it too
#
kylewm
it does??
#
tantek
and typical RSVP UIs have a place to comment at the same time as picking an RSVP state
#
aaronpk
also encourages comments for rsvp no
#
tantek
before you "send"
#
tantek
aaronpk: AKA "declines" in FB
#
kylewm
ok it changes the prompt to "Say why you can't go", but I wouldn't say it encourages you to comment on your decline
#
kylewm
and also the comment is a separate action
#
tantek
kylewm: it used to be one action - not sure when it changed
#
tantek
I feel like the user-visible fall back text inside the <data> element has a role to play here
#
kylewm
sorry benwerd if you just 9000 notifications of me toggling between Going and Not Going
#
tantek
is getting them too :)
#
tantek
if anything we have talked about extending what to RSVP
#
tantek
e.g. RSVPing with maybe "remote"
#
tantek
or RSVPing with a particular time block
#
tantek
does that make an RSVP an event itself? if it has its own dt-start and dt-end?
#
tantek
gets a real world example
#
tantek
ok I'll edit one
#
kylewm
if we give RSVPs a dt-start and dt-end, next they'll start wanting their own fax numbers
#
KevinMarks
there are "pick a time slot" use cases
#
kylewm
Doodle
#
KevinMarks
both in the "lets all agree on a time" case
#
KevinMarks
and the "pick a slot to have your chat" ones
#
tantek
KevinMarks - the "pick a time slot" and "let's all agree" use-case are coordination dependent
#
tantek
The use-case I'm talking about is not
#
tantek
it's more FYI - I'm only going to be there from x to y
#
tantek
or: I won't be there until x
#
tantek
or: I have to leave at y
#
tantek
I have seen all three of those
#
tantek
in text comments as part of an RSVP - not in RSVP UIs
#
kylewm
yeah I can see those as being part of the RSVP, not a separate comment. also "maybe: if i can get a babysitter", etc.
#
tantek
or no: sorry, already have plans!
#
KevinMarks
our three methods are surprise, fear and ruthless efficiency
#
tantek
what are our three methods?
#
tantek
hah - too bad it doesn't work in reverse chronological order
#
tantek
hey aaronpk - can you extend the window for dfns by -3 minutes into the past?
#
aaronpk
so not going there lol
#
tantek
what? time is a continuum. I don't see the problem.
#
aaronpk
searching logs is harder than waiting for a specific bit of text
#
tantek
what is searching logs?
#
tantek
yeah ok this would be mischievous ;)
#
KevinMarks
no-one expects the tantek inqusition
#
@pierreozoux
@BlakeLaPierre @coreoslinux I was considering participating, publishing on my site, and putting a copy on @airpair. #indieweb
(twitter.com/_/status/573289545032212480)
#
snarfed
so re bridgy rsvps…i glaze pretty fast at markup, as usual, but i'll implement anything as long as i don't have to understand it
#
snarfed
feel free to file an issue if/when there's a clear conclusion
#
kylewm
for the markup i was imagining, I think we would need a new property <span class="p-rsvp">yes</span> I'm going to HWC. <span class="p-rsvp-comment">I may be a few minutes late</span>
#
kylewm
(obviously it would need a better name than p-rsvp-comment
#
tantek.com
edited /event (+540) "/* Attendees */ separate silo attendees, start document some of the indie event rsvp display brainstorming"
(view diff)
#
tantek
kylewm: I don't think we need a new property
#
tantek
we already have an explicit way to provide comment content (rather than implied)
#
tantek
and that is p-summary or p-content - depending on whether the publisher wants the entire reply displayed as a comment on the destination, or only a summary
#
tantek
note p-name has *always* been fallback reply/comment content
#
tantek
thus the answer I think is simple
#
tantek
RSVP aware consuming code: if there's a p-rsvp, use it, and only display a comment if it is in a p-summary, or p-content
#
kylewm
I think I need to see an example, typically reply code would use the content property instead of the name property if it was available
#
tantek
RSVP unaware consuming code: (already defined behavior in /comments-presentation ) if there's a p-summary use it, else if p-content use it, else if p-name use it.
#
tantek
there, done
#
tantek
now let's see how existing RSVPs match that
#
aaronpk
that means many RSVP publishers will need to change their current behavior
#
tantek
aaronpk evidence?
#
tantek
not me
#
tantek
I already put my RSVP comment content in e-content
#
aaronpk
looks like I do too http://aaron.pk/r4_22
#
kylewm
tantek, how would RSVP-unaware code display: <span class="p-rsvp">yes</span> <span class="p-content">I'll be a couple minutes late</span>.
#
tantek
kylewm only the latter
#
tantek
span contents
#
tantek
thus that should be published as:
#
tantek
wait - that's not a real example - there's no event being reference
#
tantek
kylewm: URL to a real world example?
#
tantek
meanwhile in bridgy rsvp markup...
#
tantek
ditch <div class="e-content"> and the respective </div>
#
tantek
and instead of <div class="p-name"><a class="u-url" href="https://www.facebook.com/444025702419735#214611">Tantek Çelik is attending.</a></div>
#
kylewm
I don't have a real world example because this rule didn't exist yet!!
#
tantek
use <div class="p-name"><a class="u-url" href="https://www.facebook.com/444025702419735#214611">Tantek Çelik <data class="p-rsvp" value="yes">is attending</data>.</a></div>
#
tantek
hmm can further compact that
#
tantek
by de-duping the author
#
tantek
<div class="p-name">
#
tantek
<div class="h-card p-author">
#
tantek
<div class="p-name"><a class="u-url" href="https://www.facebook.com/214611">Tantek Çelik</a></div>
#
tantek
<img class="u-photo" src="https://graph.facebook.com/v2.2/214611/picture?type=large" alt="" />
#
tantek
</div>
#
tantek
<a class="u-url" href="https://www.facebook.com/444025702419735#214611">
#
tantek
<data class="p-rsvp" value="yes">is attending.</data></a>
#
tantek
</div>
#
tantek
for RSVP aware clients - all they should be using is the p-rsvp, not the p-name
#
tantek
and for RSVP unaware clients - they will fall back to the p-name and display "Tantek Çelik is attending."
#
tantek
so I think that works
#
kylewm
tantek: here's an example. on your post http://tantek.com/2015/061/t3/homebrew-website-club-creamery "only for 17:00-19:00" is the part that should be displayed as the rsvp comment if this is recognized as an rsvp post
#
tantek
that fix to bridgy can be deployed immediately
#
aaronpk
I don't think the author name should be part of the p-name
#
tantek
and should work with existing consumers no problem without the e-content
#
tantek
as well as DRYing out the person and "is attending" text!
#
aaronpk
because then i'll be displaying the author name twice
#
aaronpk
er, not me, but rsvp-unaware consumers
#
tantek
that's reasonable
#
aaronpk
but that's a minor change
#
tantek
kylewm: my intent in that post was to have the entire thing displayed as a comment, not just the "only for 17:00-19:00" part
#
tantek
aaronpk - yes
#
tantek
and makes sense
#
aaronpk
i think the key that will help is having bridgy drop the e-content
#
tantek
like I said earlier, too much markup, make less :)
#
kylewm
I need to see an example with with p-content where p-content < p-name
KevinMarks_ joined the channel
#
tantek
kylewm: if I didn't want that context as part of the comment, then I'd only include in my reply-context
#
tantek
s/include/include it
#
Loqi
tantek meant to say: kylewm: if I didn't want that context as part of the comment, then I'd only include it in my reply-context
#
tantek
kylewm the "general" example for that is when the p-name includes the p-rsvp as well as the p-content\
#
tantek
or is that what you mean?
#
kylewm
that was the example i was trying to give earlier but it was ignored because there is not an example in the wild
#
tantek
let's at least fix Bridgy's markup and move forward
#
tantek
kylewm: yeah - because it didn't read sensible
#
tantek
s/sensible/sensibly
#
Loqi
tantek meant to say: kylewm: yeah - because it didn't read sensibly
alanpearce joined the channel
#
kylewm
<div class="h-entry">In reply to <a class="u-in-reply-to" href="http://aaron.pk/hwc-event">Homebrew Website Club</a>, <data class="p-rsvp" value="yes">I'll be there</data> <span class="p-content">but I may be a couple minutes late</span></div>
#
kylewm
RSVP aware consumers would handle that properly. Kyle [[RSVP Yes]] but I may be a couple minutes late
#
kylewm
RSVP-unaware consumers would only dispaly "but I may be a couple minutes late"
cmhobbs joined the channel
#
aaronpk
but isn't "I'll be there" part of the user-entered text in this case, so should go inside the p-content?
#
kylewm
I thought that was the whole entire problem we were trying to solve?
#
tantek
right aaronpk - I think that's the key
#
kylewm
"is attending" in the content
#
tantek
kylewm: no we are trying to solve the auto-generated text problem
#
aaronpk
but this example is "I'll be there" not "is attending"
#
tantek
digs through bridgy
#
kylewm
so there's no way to have some text as the comment and different text as the standalone-comment
#
kylewm
that's fine
#
tantek
yes I think that's reasonable
#
kylewm
didn't say that well
#
kylewm
no way to have some text as the fallback and only a subset of that text as the rsvp-comment
#
aaronpk
when would you want to do that?
#
kylewm
it's fine
#
kylewm
the example i gave above is when i would want that, but it doesn't bother me if it's not possible
#
aaronpk
i actually disagree with that
#
kylewm
you disagree that that's what i would want?
#
aaronpk
lol well I disagree with that being a desirable outcome. Seeing just "but I may be a few minutes late" seems wrong, since the user entered more than that.
#
kylewm
what if it was [is attending
#
aaronpk
did you type "is attenting" when you made the RSVP post?
#
kylewm
what if it was "[is attending.] I'm really looking forward to seeing everyone"
#
aaronpk
cause that'd be pretty weird
#
aaronpk
i'm not sure what the brackets mean in your example
#
tantek
and that's exactly why we force this back to real world user behavior / examples
#
kylewm
i mean, what if i selected "Yes" from a drop down menu in my UI, and then typed in a comment
#
kylewm
the comment wouldn't necessarily duplicate "Yes I am attending"
#
tantek
snarfed I'm lost in that code - can't find any mention of p-rsvp
#
aaronpk
ah I see what you're getting at now
#
snarfed
tantek: search for rsvp
#
snarfed
i use a parser so i never see 'p-'
#
aaronpk
but I think it's still covered
#
kylewm
or what if the comment was ambiguous. like "I really miss you guys"
#
tantek
snarfed, I'm not following where <data class="p-rsvp"… is actually generated
#
aaronpk
that's a good one. so you type "I really miss you guys" and choose "RSVP yes", that is handled fine for rsvp-aware consumers. but what about unaware?
#
tantek
aaronpk, kylewm I think the first problem is that the DRY violations in the Bridgy output is confusing things more than we need.
#
kylewm
aaronpk: precisely
#
aaronpk
tantek: this case actually has nothing to do with bridgy
#
aaronpk
here is an actual example of an ambiguous comment http://aaronparecki.com/replies/2014/07/01/1/
#
snarfed
tantek: ah. that's https://github.com/snarfed/activitystreams-unofficial/blob/master/source.py#L403 . a bit far away; it's used in the AS object's html content
#
aaronpk
currently that will appear as just a comment saying "I'm not available to host this week" on rsvp-unaware consumers, which does not include the RSVP yes/no information
#
tantek
why is that bad?
#
aaronpk
because people won't see whether it's a yes or no?
#
tantek
hmm - my bad - poor framing
#
tantek
"rsvp-unaware" is really only referring to readers right?
#
aaronpk
no, displaying comments too
#
tantek
where?
#
aaronpk
like before I added code looking for the "rsvp" property on replies
#
aaronpk
on event permalinks
#
tantek
why would you RSVP to a non-event?
#
aaronpk
yes thanks kylewm
#
tantek
wait is that an event post that shows comments but not RSVPs?!?
#
aaronpk
of course, all sites do that until the add explicit support for rsvps
#
tantek
ok I'm confused, why would an event post not display RSVPs?
#
aaronpk
because snarfed can create event posts by entering HTML into wordpress, but his webmention plugin doesn't understand rsvps
#
tantek
a-ha these are events embedded in a blog post, not actual event posts
#
snarfed
aaronpk: that's not true. "is attending" and "is not attending" are rsvps
#
snarfed
i just choose not to style them distinctly
#
aaronpk
oh I didn't realize that, they looked like regular comments and contained the bridgy default text
#
snarfed
oh i see what you mean though. functionally there's not much difference
#
snarfed
there is a little difference in the plumbing, but yeah, you can argue my site is only rsvp-aware underneath, not in presentation
#
aaronpk
i assume "is attending." is actually from the bridgy post and not generated by your site?
#
tantek
snarfed, I can't workout how to rewrite the markup across multiple elements
KevinMarks__ joined the channel
#
snarfed
aaronpk: it's both. my site massages rsvp content. details in https://snarfed.org/functions.php.txt
#
snarfed
tantek: sorry i don't follow
#
aaronpk
snarfed: whoa crazy
#
tantek
snarfed, in that Bridgy RSVP permalink I want to rewrite a *chunk* of markup, not just the line with p-rsvp
#
aaronpk
why do all that work to parse rsvps if you're not going to display them differently?
#
snarfed
tantek: oh, you're saying, you don't know how to refactor the bridgy code to get the markup you want?
#
snarfed
don't worry about that. if you can file an issue with the exact markup you want, we'll make it happen
#
snarfed
aaronpk: not sure
#
snarfed
i think there was a reason
#
tantek
snarfed - ok that's a better way :)
#
snarfed
aaronpk: looking at my code, it just slightly tweaks language and formatting
#
snarfed
but i think the rsvp language actually comes from the semantic linkbacks plugin, not bridgy
#
snarfed
not sure though
#
aaronpk
that ambiguous benwerd rsvp is interesting though
#
snarfed
which?
#
snarfed
looks like he RTed, rsvped no, and replied
#
aaronpk
oh yep, that's not an rsvp. my bad
#
tantek
snarfed: ok I think I wrangled markdown code blocks: https://github.com/snarfed/bridgy/issues/371
KartikPrabhu joined the channel
#
snarfed
thanks tantek!
#
tantek
I think what we discovered is the "name" is the general fallback
#
tantek
or rather, "name" *works* as a general fallback
#
tantek
further question - could/should we use *summary* as a *deliberate* fallback?
#
aaronpk
interesting, I could see doing that for my bike rides
#
tantek
that would free-up "name" to be more name-like if/when needed
#
tantek
whereas "summary" kind of makes sense for the *visual* summary of an RSVP with an icon representing yes/no/maybe for example, or even the auto-generated text for it
#
tantek
just some thinking
#
tantek
I still think summary is a good way to communicate to consuming code that the summary *can* be displayed instead of the full content, with perhaps a way to expand to show the full content
#
tantek
instead of the consuming code having the ellipse the content at an automatic but inopportune breaking point
#
GWG
I have returned
wagle_ joined the channel
#
kylewm
tantek: I'm still a little confused about the previous issue. We have solved it so RSVP-aware implementations can ignore content-less bodies, but not for RSVP-unaware implementations (afaict)
tilgovi_ joined the channel
#
kylewm
maybe that's just an ignorable case. events that don't recognize rsvps would be pretty weird
#
tantek
perhaps because the latter use-case is unclear
#
tantek
hence I asked about readers
#
tantek
like if your reader saw an RSVP post in my h-feed - would it do anything special?
#
kylewm
yeah readers is a good point
yakker_ joined the channel
#
tantek
I think p-name as plain text fallback may be a forced position
#
tantek
as in, just like the <title> element
#
tantek
where you can *only* display plain text
#
kylewm
how should a reader display https://kylewm.com/test/rsvp.html
cphyc1 joined the channel
#
tantek
why test?
Jeena_ joined the channel
#
kylewm
tantek: because redwind can't generate replies where p-content is separate from p-name
#
tantek
you're deliberately asking about an implied p-name?
#
tantek
in which case
#
tantek
not sure why attendingHomebrew is all one word
#
kylewm
I would want that rsvp to show up like "Kyle <img src="is-attending.png"/> Thanks for hosting!"
#
kylewm
but my reader would just show "in reply to Homebrew Website Club ... Thanks for hosting"
#
tantek
er, there's no inreplyto
#
kylewm
argh why not
#
tantek
is confused now and probably needs foo
#
kylewm
please reload
#
kylewm
oh, food > reload
#
tantek
kylewm: because there's no in-reply-to in the source?
#
tantek
right I think that's a reasonable fallback behavior for a reader
#
tantek
it's a reply, and the contents of your reply are "Thanks for hosting."
#
tantek
it might not be ideal, but that's the point of a fallback, they're not ideal
#
tantek
and if you didn't have that content
#
tantek
"I'm attending\n Homebrew Website Club at Quip\n Kyle"
#
tantek
also seems fine, though I suppose you could drop the author name if you put in an explicit p-name in the h-entry
#
tantek
which I think is reasonable for the same reason that aaronpk gave
#
tantek
or heck you could do this
#
tantek
<data class="p-rsvp" value="yes"><a class="p-author h-card" href="https://kylewm.com"><abbr title="Kyle">I'm</abbr></a> attending</data>
#
tantek
<a class="u-in-reply-to" href="https://snarfed.org/2014-06-16_homebrew-website-club-at-quip">Homebrew Website Club at Quip</a>
#
tantek
since presumably your <data> element and author element are auto-generated anyway - you can auto-generate that whole thing
#
tantek
and then no need for explicit p-name inside the h-entry
#
kylewm
does it at least make sense why i was arguing for a more complete fallback than that?
#
tantek
where you want a reader unaware of rsvps to display *all* of that ?
#
kylewm
yeah, "I'm attending HWC. Thanks for hosting"
#
tantek
I think it would/should - in the *title*, e.g. in a list view
#
tantek
which is where you're going to use the entire p-name rather than just the content
#
kylewm
in this case, I would detect the p-name as auto-generated and ignore it (bc content is a subset of name)
#
kylewm
and usually those types of p-names are ugly
#
tantek
I suppose we still need to address that
#
kylewm
we can stop talking about it :) I'll document my concerns on the wiki
#
GWG
snarfed: You are dequeuing the CSS style sheet for syndication links...what about the part that governs what icon goes for what URL?
#
tantek
hey ben_thatmustbeme can you post the IWC header image or make me a co-admin of https://www.facebook.com/events/437461433068146/ and I can?
#
tantek
or aaronpk
#
tantek
wonders if he should bother to design clustering of RSVP posts in a stream.
#
ben_thatmustbeme
tantek, you actually should be a host
#
tantek
right, host, I keep forgetting the terminology
#
tantek
that seems like an event post feature worth documenting
#
tantek
what is an event host?
#
tantek
can any host add a host? or only the creator?
#
ben_thatmustbeme
looks like I can add a host, so i would assume any
#
tantek
keeps refreshing
#
ben_thatmustbeme
it shows you as already being a host
#
tantek
what are badges?
#
ben_thatmustbeme
ahh, you aren't made a host until you actually say you are attending
#
ben_thatmustbeme
sooo is that a both that just tweets random stuff?
#
ben_thatmustbeme
and retweets random things
#
tantek
uh who are all these people invited to indiewebcamp cambridge?
#
tantek
what just happened?
#
tantek
who is Micky Metts?
#
ben_thatmustbeme
i sent out another round of invites to developer friends in the area
#
ben_thatmustbeme
i dind't realize bridgy reports them all to loqi
#
ben_thatmustbeme
as for who Micky Metts is, looking at the photos i'm pretty sure she was there last year
#
tantek.com
edited /2015/Cambridge (-8) "move to do to be part of organizers"
(view diff)
#
tantek.com
edited /2015/Cambridge/Planning (+128) "contact last year sponsors"
(view diff)
alanpearce, jacus, gRegor` and nloadholtes joined the channel
#
bret
is an <ul> of <li>'s really only different than a <div> of <div>'s by its default display style? (and semantics?)
#
acegiak
they're also spelled differently
#
bret
i guess there are some special attributes
#
bret
are semantic tags used for anything?
#
bret
other than convention, and shorthand styling hooks?
#
kylewm
bret: tantek has said in the past basically no
#
bret
why doesn't microformats use them?
#
bret
or have some kind of mapping to them
alanpearce joined the channel
#
kylewm
like <article> always implies class="h-entry"?
#
bret
i mean, I haven't really thought through it
#
bret
but that would be wicked clean
KartikPrabhu joined the channel
#
kylewm
that must have been sort of the dream of them, that a screenreader would say "give me all the articles under main" or whatever
#
bret
has anyone/thing ever tried that?
#
kylewm
"people tried defining <article> based fallbacks to convert to Atom semantics and failed horribly in the HTMLWG in the late 2000s"
#
bret
wonder what the failure was
#
bret
people markup things with mf2 to get access to a machine readable version of the page, people markup with semantic tags based around a loose idea of how they should be used
#
bret
maybe due to lack of an end goal, what you could possible get back from semantic tags is relatively low quality
pwcc joined the channel
#
bret
still, if there was an actual usecase with how semantic are used together, like there is with mf2 classes, (maybe mirroring mf2) people would use them in a useful way
#
bret
anyway, not planning on writing a parser anytime soon
#
aaronpk
i dunno we seem to get by just fine with microformats classes, which have the added benefit of being recognizable in the source code
sanduhrs joined the channel
#
pwcc
MDN has some helpful usage notes on semantics - eg https://developer.mozilla.org/en-US/docs/Web/HTML/Element/article
#
pwcc
That said, reality also kicks in firmly in other places. Be great if browser makers could solve this chicken/egg problem 0- https://developer.mozilla.org/en-US/docs/Web/Guide/HTML/Sections_and_Outlines_of_an_HTML5_document
snarfed, wolftune and KevinMarks_ joined the channel
#
Loqi
slack/snarfed: GWG: good question! that's in my stylesheet, <https://snarfed.org/w/wp-content/themes/snarfed-ryu/style.css> . search for relsyn
halorgium and Erkan_Yilmaz joined the channel
#
KevinMarks_
Bret, we had xoxo which used li and dl tags for json type structures
KevinMarks_, KartikPrabhu, tilgovi, e-lima, cweiske, tantek, alanpearce, sanduhrs, stream7, Jihaisse, loic_m and pfefferle_ joined the channel
#
pfefferle_
good morning all
friedcell, pfefferle, Jeena, KevinMarks__, stream7, krendil and elf-pavlik joined the channel
#
@andkjaer
The Ultimate Guide to Creating a Killer Explainer Video >> http://www.quicksprout.com/?p=30073&utm_content=buffer570a4&utm_medium=social&utm_source=twitter.com&utm_campaign=buffer #SocialMedia Tip from Neil Patel. #SelfPublish #IndieAuth…
(twitter.com/_/status/573402844721758208)
#
cweiske
is anyone using php with gearman? how do you ensure the workers are started when booting the machine?
petermolnar, krendil, petermol1ar, friedcell and adactio joined the channel
#
pfefferle
!tell GWG your likes (https://david.shanske.com/2014/12/26/listening-serial/) are not accessible via RSS/Atom
#
Loqi
Ok, I'll tell them that when I see them next
#
fkooman
cweiske: i added both class1 and class3 cacert root cert, are they both needed, or is the class1 sufficient?
#
cweiske
I'm not sure about this unfortunately
#
cweiske
thanks, it works with my page now
#
cweiske
now I get: "IndieCert was unable to find a matching fingerprint on your claimed URL"
#
cweiske
hey, and here you tell me what to put into the html
#
fkooman
yeah, that was always there :)
#
cweiske
I never came to this point
#
cweiske
but it'd be still nice if the algorithm was documented somewhere
#
fkooman
yeah, working on it :)
#
cweiske
You are now successfully authenticated as https://cweiske.de/!
#
fkooman
it is actually super simple
pfefferle_, Sebastien-L and adactio_ joined the channel
#
@GlennRicePro
So, Hillary Clinton prefers self-hosting, eh? Could the U.S. be getting their first #indieweb President in 2016? ;-)
(twitter.com/_/status/573423381019631618)
#
fkooman
cweiske: also documented the fingerprint generation in the blog post
#
cweiske
can't we simply use the sha256 fingerprint that the browser's cert info popup shows?
dietrich joined the channel
#
fkooman
cweiske: he doesn't i just published it 3 seconds ago ;)
#
elf-pavlik
:D
#
cweiske
erm, I read it yesterday already
#
fkooman
cweiske: oh yeah, but it was not mentioned anywhere, you looked around! :P
alanpearce joined the channel
#
fkooman
cweiske: there is a standard for publishing fingerprints on webpages...so I'd prefer to use that :) I also like the scoping to the particular instance of "IndieCert"
#
wwelves.org perpetual-tripper
edited /IndieAuth (+237) "/* Why is the IndieAuth verification response form-encoded instead of JSON */ Proposal for content negotiation"
(view diff)
michielbdejong, cweiske and LauraJ joined the channel
#
tommorris
fkooman: those all seem reasonable. I hope aaronpk implements ‘em. ;-)
#
fkooman
tommorris: thanks :)
dns53, Acidnerd, KevinMarks_, pfefferle, kerosene, frzn, alanpearce and interactivist joined the channel
#
GWG
Good morning
#
Loqi
GWG: pfefferle left you a message 2 hours, 52 minutes ago: your likes (https://david.shanske.com/2014/12/26/listening-serial/) are not accessible via RSS/Atom
michielbdejong joined the channel
#
GWG
pfefferle: Are you sure?
#
GWG
Do you mean the posts themselves, or the part of it I add dynamically?
#
GWG
I probably need a custom feed template.
alanpearce and modem joined the channel
#
pfefferle
GWG the link I posted has no ”žreal“ content, and so the content field of the feed is completely empty
#
GWG
pfefferle: Yes, so I need to customize the feed. I didn't think of that.
#
pfefferle
GWG I realized that I forgot to subscribe to your feed!!
#
pfefferle
GWG thats why I found the problem ;)
#
GWG
I guess I need to look up how to do that. But will figure it out.
#
fkooman
cweiske: i removed the class 3 CAcert subroot again, it is signed by the class 1 cert, and everything still seems to work...
#
GWG
pfefferle: I've added it as an issue
#
cweiske
yes, I can still authenticate
Acidnerd, friedcell, bupkes and KevinMarks_ joined the channel
LauraJ, interactivist, wolftune and KevinMarks_ joined the channel
KartikPrabhu, KevinMarks_ and interactivist joined the channel
#
GWG
I'm working on styling h-cards again. Why don't we have h-card examples in the wiki? Or am I just missing them?
friedcell, elf-pavlik and stream7 joined the channel
#
ben_thatmustbeme
GWG, you should probably check the microformats wiki for that
#
GWG
ben_thatmustbeme: I meant styling, not markup
#
ben_thatmustbeme
oh, hmm well... screen shot yours and add it :P
#
@anomalily
@HilaryClinton is #indieweb - she runs her own mail server out of her house.
(twitter.com/_/status/573502102313439232)
#
GWG
ben_thatmustbeme: I want to improve mine.
#
GWG
I should be screenshotting other people's.
#
ben_thatmustbeme
ick, mail servers
elf-pavlik and verdi_ joined the channel
#
@adamprocter
a #markdown #golang blog engine with mysql comments (future hope to include webmentions and inline commenting) https://github.com/adamprocter/gomarkdownblog/
(twitter.com/_/status/573505248377073664)
chalettu, friedcell, alanpearce, elf-pavlik, LauraJ, Acidnerd, tantek, pfefferle_ and adactio joined the channel
#
tantek
ben_thatmustbeme, aaronpk about the event hosts feature, at least on FB, it appears only the creator of the event can upgrade people to hosts.
#
tantek
e.g. on this event: https://www.facebook.com/events/444025702419735/ I am a host, but I am unable to make sandro a host
#
tantek
aside: just invited / added you to that event as well ben_thatmustbeme
yakker joined the channel
#
tantek
what is an event host?
#
GWG
tantek: The question I asked a few minutes ago...it seems up your alley. You have any thoughts?
#
tantek
!tell GWG, ben_thatmustbeme yes there are nicely styled hCards documented on the microformats wiki (obviously easy to re-use for h-card) http://microformats.org/wiki/hcard-examples-in-wild#Individuals (see the first subsection "Nicely styled)
#
Loqi
Ok, I'll tell them that when I see them next
KevinMarks_ joined the channel
#
tantek
first example is our very own npdoty
gRegor` and adactio_ joined the channel
#
kylewm
tantek: http://timvandamme.com/ example is no longer an hcard
#
Loqi
ben_thatmustbeme: tantek left you a message 5 minutes ago: yes there are nicely styled hCards documented on the microformats wiki (obviously easy to re-use for h-card) http://microformats.org/wiki/hcard-examples-in-wild#Individuals (see the first subsection "Nicely styled)
bupkes and danlyke joined the channel
#
kylewm
fkooman: in your indieauth proposal, I don't understand why/when you would want the auth and verify endpoints to be different
cweiske joined the channel
#
cweiske
aaronpk, how is a hub supposed to make a thin ping in pubsubhubub 0.4?
#
cweiske
section 7 only talks about fat pings
LauraJ joined the channel
#
kylewm
cweiske: section 7 doesn't say the content of the notification has to be anything in particular
#
aaronpk
good morning
#
aaronpk
lots of catching up to do
#
GWG
aaronpk: How so?
#
Loqi
GWG: tantek left you a message 56 minutes ago: yes there are nicely styled hCards documented on the microformats wiki (obviously easy to re-use for h-card) http://microformats.org/wiki/hcard-examples-in-wild#Individuals (see the first subsection "Nicely styled)
#
aaronpk
cweiske: yes the spec itself doesn't say what the payload contains, so you could just as well interpret that as fat or thin pings
peacekeep3r and KevinMarks_ joined the channel
#
Acidnerd
can someone lend me a hand for a few minutes ?
#
Acidnerd
i wanted to start my indie site from scratch
#
Acidnerd
i followed all the install steps
#
Acidnerd
but now when i visit on browser i get 108.61.190.132
#
Acidnerd
a php file being downloaded ??
#
Acidnerd
i'm probably missing out something very simple, just can't find it
#
Acidnerd
instead of showing the process to finish install i get this in the URL bar : http://_/begin/
#
Acidnerd
anyone ? aaronpk ?
#
aaronpk
Acidnerd: it sounds like your web server is not configured to run php code
#
Acidnerd
that's odd
#
Acidnerd
hmm let me check that
KevinMarks_ joined the channel
#
aaronpk
now reviewing fkooman's post
pauloppenheim joined the channel
#
Acidnerd
aaronpk, it's getting better
#
Acidnerd
but now it's nginx turn to freak out
KevinMarks_ joined the channel
#
bret
getting nginx and php working together on a vps presents a surprisingly steep learning curve when you are just getting started
#
GWG
bret: I didn't find it that hard
#
GWG
It's not Apache, but there were plenty of tutorials
#
Acidnerd
also the wiki part about Known and Nginx is hard to catch
#
Acidnerd
and his config link is dead : https://gist.github.com/rascul/ad665c9df9898c6e5074
#
bret
part of the problem was how opaque linux packaging is. I would install something and have no idea where it would land or be configured
#
bret
and would be inconsistent with the documents I was reading
#
Acidnerd
the official Known doc also don't even mention Nginx once
#
bret
it was a while ago, definitely a frustrating experience. I think not finding good tutorials was part of the problem
#
bret
i should have another go now that I have a much deeper understanding of unix
#
rascul
i wasn't able to get known completely working with nginx
#
Loqi
definitely
#
rascul
i wasn't aware something was linked to a gist of mine
adactio joined the channel
#
bret
GWG: do you have a collection of resources that you recommend to beginners?
#
Acidnerd
it's on the wiki rascul
#
rascul
which page?
#
Acidnerd
i think i'm going to give up and go back to apache & mysql
#
Acidnerd
i also gave up mongodb
#
Acidnerd
it's just impossible, either from repository, or latest nginx repo or even manual install
#
kylewm.com
created /Postel's_law (+631) "stub with definition"
(view diff)
#
rascul
removing it now, i ended up using apache in the end for it
#
rascul.io
edited /Known (-147) "/* Setup Known on nginx */ link was dead and i'm not using nginx"
(view diff)
#
Acidnerd
nobody is i think
#
Acidnerd
but the wiki make it look like it's doable
#
aaronpk
i think i did it at one point
#
kylewm.com
edited /IndieAuth (+276) "/* Why is the IndieAuth verification response form-encoded instead of JSON */ vote to settle on one canonical format"
(view diff)
#
rascul
i had it running but i couldn't get any of the extra stuff working like twitter
#
aaronpk
ugh people don't like the form-encoded response
#
aaronpk
the only reason I could see to switch to JSON response is that the OAuth 2.0 spec says it should be JSON
#
kylewm
are there other examples of APIs returning form encoded data?
#
bret
who cares about form encoded, use the library that every language has
#
aaronpk
remember when every language had an xml parser?
#
bret
oh right, server to server
#
aaronparecki.com
edited /IndieAuth (+232) "/* Why is the IndieAuth verification response form-encoded instead of JSON */"
(view diff)
#
bret
the form encoded lib that every...*
Pierre-O, KevinMarks_ and michielbdejong joined the channel
#
bret
who cares if its json vs form encoded? its results in the same object after parsing either way
#
aaronpk
json_decode vs parse_str in php
bupkes and frzn joined the channel
#
bret
js pass it through JSON.parse
#
bret
form encoding needs a lib
#
elf-pavlik
bret++
#
bret
i guess thats the single, expected advantage
#
Loqi
bret has 52 karma
#
aaronpk
how about other languages?
#
elf-pavlik
maybe worth doing comparison on a wiki?
KevinMarks_ joined the channel
#
aaronpk
ruby `URI.decode_www_form` but doesn't treat [] as arrays and actually turns it into a kind of weird data structure
#
bret
form encoding has the advantage for micropub, since you can do dead simple clients
#
aaronpk
form-encoding for POST requests still makes a ton of sense
#
aaronpk
but this is in the context of responses
#
bret
aaronpk: really? oh no
#
aaronpk
yeah if someone wants to make a wiki page comparing the behavior of parsing form-encoded and json in a bunch of langauges/environments that would be awesome
#
elf-pavlik
why not add xml to that mix :)
#
aaronpk
sure lol
#
bret
libxml--
#
Loqi
libxml has -1 karma
#
elf-pavlik
form-encoded | json | xml
#
bret
nein! XD
#
aaronpk
not like anyone would actually fight for an XML API at this point
#
bret
i think there are some interesting conveniences that both json and form encoding have... right now today.
#
bret
parsing json in JS is super easy and generally well supported in other languages through standard libraries
#
bret
form encoding post requests can be made by raw html
#
aaronpk
bret: ignore the form-encoded requests for this argument
#
aaronpk
this is only about API responses
#
bret
yeah, for api callbacks, I would say json would be easier to deal with
#
elf-pavlik
IMO also more intuitive
#
elf-pavlik
at least as for today ...
#
bret
unless there is an example where the callback needs to be sent from an html page?
#
elf-pavlik
maybe application/x-protobuf ? ;)
#
aaronpk
yeah see this is what i'm talking about
stream7 joined the channel
#
ben_thatmustbeme
yeah, xml reply would be.... horrible
sparverius joined the channel
#
elf-pavlik
aaronpk, if you enable Content Negotiation, would only indieauth.com need to support both from-encoded and json?
#
aaronpk
and any other authorization servers
#
ben_thatmustbeme
actually it would be better to use mf2 reply then as you can just display results without parsing if desired
#
aaronpk
ben_thatmustbeme++
#
Loqi
ben_thatmustbeme has 47 karma
#
ben_thatmustbeme
depending on the situation that may be useful
#
aaronpk
not super useful for the token response, since that is only ever consumed by code
#
elf-pavlik
but not clients
#
ben_thatmustbeme
yeah, but for webmention response that might be quite useful
#
aaronpk
ben_thatmustbeme: my webmention endpoint does that already :)
#
ben_thatmustbeme
especially if you return a status URL
#
aaronpk
look at the webmention form on my site, it makes a POST request to the actual webmention endpoint
#
bret
ben_thatmustbeme: thats an actually pretty interesting idea.
#
aaronpk
if it's coming from a browser then I return HTML
#
aaronpk
with a status URL
#
bret
for micropub at least, not indieauth
#
ben_thatmustbeme
aaronpk, thats a good idea. I haven't done the whole "give me your link" thing
#
elf-pavlik
"convert between mf2 and form-encoded micropub"
#
ben_thatmustbeme
as I would always interpret it as a mention instead of a reply or anything
#
ben_thatmustbeme
hey aaronpk, did you ever try doing the mp_update_field things?
#
bret
elf-pavlik: parse html + mf2 -> json -> form encode that object
#
aaronpk
ben_thatmustbeme: not yet, remind me?
#
aaronpk
is there a link?
#
ben_thatmustbeme
for updating a field, appending, removing, etc
squeakytoy2 joined the channel
#
ben_thatmustbeme
i was thinking at IWC i may try to create a new endpoint for contacts
#
ben_thatmustbeme
and one for chat logs
#
elf-pavlik
bret, can you make gist with simple js script doing it?
#
ben_thatmustbeme
being able to add fields without having to send the whole contact would be better, was going to model it after MP
#
bret
elf-pavlik: sure
#
elf-pavlik
bret, please also make some values arrays...
#
elf-pavlik
e.g. multiple email addresses in h-card
#
elf-pavlik
doesn't know mf that well...
#
cweiske
aaronpk, about the pubsub payload: section 7 says: "... with the payload of the notification. This request MUST have a Content-Type corresponding to the type of the topic. The hub MAY reduce the payload to a diff between two consecutive versions if its format allows it."
#
aaronpk
re-reads
#
cweiske
this totally sounds like I have to send the whole topic/contents of the page
#
cweiske
except when the content type supports diffing
#
cweiske
one could argue about "payload of the notification"
#
cweiske
this could be anything
#
cweiske
but the content-type must be that of the topic....
#
aaronpk
"Notification: A payload describing how a topic's contents have changed, or the full updated content. Depending on the topic's content type, the difference (or "delta") may be computed by the hub and sent to all subscribers."
#
cweiske
ok, so it's diff or full content
#
aaronpk
so if the topic is HTML, then you have to send HTML
#
aaronpk
but "" is also valid HTML :)
#
cweiske
but "" isn't full updated content
#
aaronpk
i guess it doesn't describe how the contenst have changed either
#
aaronpk
I dunno, it's all so undefined
#
cweiske
the diff makes sense in atom or rss
#
cweiske
when you only include the items that are new
#
aaronpk
could also make sense for a list of h-entryes
#
aaronpk
or a list of comments in an h-entry
#
cweiske
if the subscriber supports h-entry
#
aaronpk
p.s. did you see my writeup?
#
cweiske
is reading now
#
aaronpk
it's a subset of the PuSH 0.4 spec that deals specifically with HTML and microformats feeds
#
aaronpk
and attempts to clarify all the ambiguous parts of 0.4
#
cweiske
ambiguous is only how publishers notify the hub
#
aaronpk
the payload is ambiguous too
#
cweiske
the hub I'm writing ("phubb" :)) will be notified by my git hooks
#
cweiske
but having a standard way for http notifications would be fine, too
#
KevinMarks_
Irony: POSSE to medium means my post on sites preferring images gets its image as a preview on twitter https://twitter.com/inmediaref/status/573419916122050560
#
KevinMarks_
Is there a term for retro POSSE? It's like linkslapping but with fat pings
#
aaronpk
cweiske: yeah I think the more that is well-defined the better, makes things easier to implement
#
bret
elf-pavlik: are you familiar with node?
#
aaronpk
back on indieauth... now I'm reconsidering content negotiation
#
elf-pavlik
bret, yes i mostly use JS nowadays :)
#
elf-pavlik
got pretty rusty with ruby because of that :S
#
aaronpk
it's only the authorization server that needs to support it, and only for one method
#
aaronpk
and generating a form-encoded response is pretty much as easy as generating a JSON response
#
aaronpk
so would it really be that bad to let clients choose?
#
elf-pavlik
nope :)
#
bret
ahh damn the form encoder I used is doing something funny
#
elf-pavlik
ration of IndieAuth clients to IndieAuth servers?
#
elf-pavlik
s/ration/ration/
#
Loqi
elf-pavlik meant to say: ration of IndieAuth clients to IndieAuth servers?
#
aaronpk
likely there will be far more clients than authorization servers
#
elf-pavlik
s/ration/ratio/
#
Loqi
elf-pavlik meant to say: s/ratio/ratio/
#
aaronpk
and far more users than servers
#
cweiske
if you want to be total indie you have to host your own indieauth server
#
aaronpk
but "be total indie" is not necessarily the goal of everyone
#
elf-pavlik
http://en.wikipedia.org/wiki/Robustness_principle not sure if fits here...
#
aaronpk
and at any point anyone can switch authorization servers from indieauth.com to their own, and nobody else needs to know
#
aaronpk
so that's pretty indie
KartikPrabhu joined the channel
#
bret
elf-pavlik: the form encoder is working great actually... the %5B%5D just looks funny. thats the [] next to the keys that tell it which array to go into
#
bret
elf-pavlik: is that script what you were looking for?
#
elf-pavlik
bret, please give me few minutes... will take a break from Social IG ACTION-4 and check it now
#
bret
k no rush
#
cweiske
aaronpk, another pubsub question: 5.2. Subscription Validation says "Subscriptions MAY be validated" but a sentence later it's "hub MUST perform the verification of intent"
#
cweiske
I originally thought that VoI is the validation
#
cweiske
but it reads as validation is something different
#
aaronpk
huh yeah that is confusing
#
ben_thatmustbeme
aaronpk, i feel like we need an alternative to indieauth.com just so people get the difference between indieauth and.... what is the actual auth mechanism called? rel-me-auth?
#
aaronpk
ben_thatmustbeme: agreed completely
#
aaronpk
someone go make one please
#
elf-pavlik
fkooman, will deoply one soon
#
aaronpk
i'm almost done with my reply
#
ben_thatmustbeme
actually i was tempted to do that and add in a warning if you are using it to log in when it is not your actual auth provider
#
ben_thatmustbeme
so to warn you are using a sight that isn't actually polling your site for the auth provider, they just hard coded that provider
#
Loqi
[mention] Aaron Parecki commented 'This is a reply to fkooman's proposed changes to IndieAuth protocol. Optional client_id Currently the protocol expects the use of the...' on a post that linked to http://indiewebcamp.com/h-x-app (https://aaronparecki.com/articles/2015/03/05/1/re-proposed-changes-to-indieauth-protocol)
#
aaronpk
there you go
#
elf-pavlik
bret, micropub also doesn't use *properties*
#
aaronpk
properties?
#
elf-pavlik
aaronpk++
#
bret
elf-pavlik: oops, looks like I didnt dig deep enough into the mf2 parsed object
#
Loqi
aaronpk has 727 karma
#
elf-pavlik
can you try to convert html mf into example in http://indiewebcamp.com/micropub#Form-encoded_Microformats_Representation ?
#
elf-pavlik
and get it exactly as micropub expects it?
#
bret
elf-pavlik: yeah, looks like it just pulls out type: "h-entry"
#
elf-pavlik
i guess code exists to get microformat json from such micropub request parameters?
#
bret
so a parsed mf2 object doesn't work as a micropub object
#
bret
you would have to transform it a bit
#
elf-pavlik
bret, you may need to fiddle with this object before converting to form-encede
#
bret
definately,
#
elf-pavlik
any links to code converting micropub form-encoded microformats to microformats json ?
#
bret
elf-pavlik: any gitpub endpoint would have to make that transform
#
elf-pavlik
link to the source?
#
bret
usually it goes micropub post (form encoded) -> language object -> site storage format (whatever you want) -> html
#
aaronpk
there is very little reason to go from micropub -> microformats json
#
bret
i have my endpoint but its not the best example right now
#
aaronpk
microformats json is actually not used very often at all, because the microformats parsers turn the html into native data structures
#
aaronpk
it's just a convenient way to display the parsed microformats data structure when talking about it
#
elf-pavlik
aaronpk, if you render html using template (e.g. handlebar) and pass it microformats json
#
bret
elf-pavlik: heh yeah that was my weekend working on a transform
#
aaronpk
elf-pavlik: do you know of anyone who actualyl does that right now?
#
elf-pavlik
bret, what about params which you didn't expect? just droping them?
#
bret
elf-pavlik: yeah, for now.
#
bret
i gtg riight now bbl
#
elf-pavlik
will need to add something to 'sprincle' mf on top of RDFa ...
#
elf-pavlik
bret, ciao o/
#
aaronpk
cweiske: oh I think this is leaving more un-specified things open for hubs
#
aaronpk
just re-read it again and it clicked
#
aaronpk
"Subscriptions MAY be validated by the Hubs who may require more details to accept or refuse a subscription"
#
elf-pavlik
soon will switch to something like in http://viejs.org/examples/todo/
#
aaronpk
i think it's saying that hubs can do whatever they want to validate subscriptions, like having their own access controls or rate limits or usage limits
#
aaronpk
so you could make a hub that lets people subscribe to 4 topics for free, but requires $$ to subscribe to more
#
elf-pavlik
i understand that among mf html, mf json and mf form-encoded only mf html->mf json has defined algorithm
#
aaronpk
cweiske: does that answer it?
#
aaronpk
cool. that was confusing.
#
bret
elf-pavlik: a useful tool: http://glennjones.net/tools/microformats/ when playing mf2-> json before I go
#
cweiske.de
edited /h-x-app (+161) "/* Aaron Parecki */"
(view diff)
#
elf-pavlik
bret, thx!
stream7, alanpearce, KartikPrabhu, benwerd, krendil and peacekeep3r joined the channel
#
kylewm.com
edited /IndieAuth (+280) "/* Why is the IndieAuth verification response form-encoded instead of JSON */ objection withdrawn"
(view diff)
#
kylewm
elf-pavlik^
#
elf-pavlik
kylewm++
#
Loqi
kylewm has 130 karma
#
kylewm
really interesting to read fkooman's proposal and aaronpk's response.
#
benwerd
I feel bad being commercial here, but FYI: https://withknown.com/pro/ is completely live. Indie Web capable, does IndieAuth and Micropub, has strong connections to brid.gy.
wolftune joined the channel
#
kylewm
benwerd++ congrats on launching known pro!
#
Loqi
benwerd has 62 karma
#
aaronpk
Nothing wrong with being "commercial"! Congrats!
#
@benwerd
Educators and students get a full 50% off Known Pro: https://withknown.com/education/ #domainofonesown #indieweb #edtech
(twitter.com/_/status/573577320532865024)
#
@nitot
RT @benwerd: Educators and students get a full 50% off Known Pro: https://withknown.com/education/ #domainofonesown #indieweb #edtech
(twitter.com/_/status/573577515190632448)
#
benwerd
aaronpk: Thank you!
#
GWG
benwerd: If I was interested in a hosted solution right now, I would strongly consider yours.
#
benwerd
GWG: Thank you :)
#
GWG
benwerd: Well, I like the product.
#
@MarkMorvant
RT @benwerd: Educators and students get a full 50% off Known Pro: https://withknown.com/education/ #domainofonesown #indieweb #edtech
(twitter.com/_/status/573578594775003137)
#
benwerd
Apparenty I need some education of my own. It's 40%.
#
benwerd
de-syndicates ...
#
ben_thatmustbeme
phht, math, who needs thats? :P
#
rascul
maths is hard
tjgillies and wolftune joined the channel
#
tjgillies
indieauth with a sms number in TXT record could trigger an sms auth flow
#
elf-pavlik
benwerd++ congratulations!
#
tjgillies
that seems rather straight forward
#
Loqi
benwerd has 63 karma
#
tjgillies
elf-pavlik: what did I miss?
#
GWG
!tell snarfed I will be deploying several alternate CSS options in your name
#
Loqi
Ok, I'll tell them that when I see them next
#
elf-pavlik
tjgillies, https://withknown.com/pro/
#
tjgillies
nice
#
tjgillies
benwerd++
#
Loqi
benwerd has 64 karma
#
elf-pavlik
benwerd, i understand that you keep *all* the codebase open source and just charge people who don't want to deploy it themselves or use https://indiehosters.net service?
#
gRegor`
You can get a free .withknown.com subdomain. Pro has additional features though.
#
gRegor`
So there's a free version people don't have to deploy themselves
#
benwerd
elf-pavlik: The whole codebase is indeed open source. The sad exception will be Google+ Pages integration, which we legally can't open.
donpdonp joined the channel
#
bret
benwerd: how come?
#
aaronpk
oh wow
#
benwerd
bret: they made us sign an NDA for the API
#
bret
lol oh wow
#
bret
thats lame
#
tjgillies
Google: "the open source company"
#
donpdonp
indieauth feature suggestion, from showing the login process to a friend with a domain name but no interest in hosting a webserver, which seems reasonable.
#
donpdonp
support a DNS 'CERT' record to get a GPG key, then proceed as usual with the gnupg key option. (http://www.initd.net/2010/12/adding-gpg-public-keys-to-your-dns.html)
snarfed joined the channel
#
tjgillies
gpg--
#
Loqi
gpg has -1 karma
#
tjgillies
sms++
#
Loqi
sms has 1 karma
#
bret
donpdonp: what about just having them use a tumblr gh-pages or neocitites?
#
tjgillies
bret: you mean point domain at it?
#
tjgillies
bret++
#
Loqi
bret has 53 karma
#
tjgillies
the think the number of people who would buy a domain and use it for indie auth but have no content is extremely small
#
bret
then they could have their gpg key on their domain and humans can read it too
#
donpdonp
bret: that would be lighter weight than full hosting, i agree
#
donpdonp
that creates a fairly major dependency
#
donpdonp
ie my auth is dependent on neocities being available
#
tjgillies
how often does neocities go down?
#
donpdonp
or blocked from china
#
donpdonp
or shutdown in an fbi raid
#
donpdonp
or simply added lookup time for the neocities dns and http content xfer
#
tjgillies
rational points
#
tjgillies
txt-record-auth++
#
Loqi
txt-record-auth has 1 karma
#
snarfed
so aaronpk tantek_ kylewm etc, i'm thinking about the rsvp markup and presentation discussion yesterday…and i'm wondering how much of it applies to likes and reposts too
#
Loqi
snarfed: GWG left you a message 24 minutes ago: I will be deploying several alternate CSS options in your name
#
snarfed
i'm guessing some at least, if not all
#
snarfed
i'll follow up in https://github.com/snarfed/bridgy/issues/371 and post bridgy's current markup there, we can see
#
snarfed
GWG: interesting! …out of curiosity, why? dequeueing your stylesheet and adding a few rules in mine worked for me
#
Loqi
[mention] Christophe Ducamp commented '@ndebock : #indiewebcamp Germany se tiendra á Düsseldorf les 9 et 10 mai. Participants : https://indiewebcamp.com/2015/Germany/Guest_List ...' on a post https://indiewebcamp.com/2015/Germany/Guest_List (http://xtof.withknown.com/2015/ndebock-indiewebcamp-germany-se-tiendra-dsseldorf-les-9-et-10)
#
Loqi
[mention] Christophe Ducamp commented '@ndebock : #indiewebcamp Germany se tiendra á Düsseldorf les 9 et 10 mai. Participants : https://indiewebcamp.com/2015/Germany/Guest_List ...' on a post https://indiewebcamp.com/2015/Germany/Guest_List (http://xtof.withknown.com/2015/ndebock-indiewebcamp-germany-se-tiendra-dsseldorf-les-9-et-10)
#
snarfed
i expectusers who want to roll their own style will be capable of setting it up, like i did
benwerd joined the channel
#
elf-pavlik
benwerd, did you document this G+ API issue on a wiki?
e-lima joined the channel
#
benwerd
elg-pavlik: No, I didn't, and I can't provide too much detail, unfortunately.
interactivist joined the channel
#
elf-pavlik
even detais about the conditions of using it not funtioning of the API itself?
#
Loqi
slack/snarfed: elf-pavlik it's well known though. see /Google+
#
Loqi
slack/snarfed: <http://WordPress.com|WordPress.com> is another example user
#
Loqi
slack/snarfed: not a surprise
#
elf-pavlik
http://indiewebcamp.com/Google+#API doesn't say anything about signing NDA ...
#
GWG
snarfed: I use SASS. I can roll several style sheets for people to hook onto that are variants. I wanted to do a B&W one anyway.
#
GWG
snarfed: So it wouldn't be any work to maintain.
#
snarfed
GWG: ok!
#
snarfed
elf-pavlik: http://indiewebcamp.com/Google+#POSSE "Make a business relationship with Google so you can use a private/partner API…"
#
GWG
snarfed: Besides, I doubt you are the only one who prefers no color in the icons
#
snarfed
you're right that it doesn't say the word NDA explicitly. feel free to add that!
#
wwelves.org perpetual-tripper
edited /Google_Plus (+103) "/* Activities */ note about signing NDA"
(view diff)
daf joined the channel
#
snarfed.org
edited /Google_Plus (+35) "the normal G+ API doesn't require an NDA, just the writable Pages API"
(view diff)
stream7 joined the channel
#
kylewm
am i interpreting it correctly that Pages API would not give me write access to my user stream?
#
snarfed
kylewm: right
#
snarfed
i think best practice right now for g+ posse is manual, by popping up their sharebox with everything pre-populated so the user only has to click "post"
#
@kevinmarks
“Known Pro lets you connect to an infinite number of social media profiles.” http://stream.withknown.com/2015/introducing-known-pro-the-best-way-to-reach-your-audience #indieweb
(twitter.com/_/status/573598285622996992)
#
@ken_bauer
RT @kevinmarks: “Known Pro lets you connect to an infinite number of social media profiles.” http://stream.withknown.com/2015/introducing-known-pro-the-best-way-to-reach-your-audience #indieweb
(twitter.com/_/status/573599134583779329)
#
@aral
RT @kevinmarks: “Known Pro lets you connect to an infinite number of social media profiles.” http://stream.withknown.com/2015/introducing-known-pro-the-best-way-to-reach-your-audience #indieweb
(twitter.com/_/status/573599146927636480)
#
bret
also.. g+ api stuff aside.. CONGRATS benwerd AND ERIN on known pro
#
benwerd
bret: Thank you!
KartikPrabhu joined the channel
#
bret
donpdonp, i was just trying to address the "dont wanna futz with a webhost" part :p
chalettu joined the channel
#
snarfed
bret, definitely
#
snarfed
benwerd++
#
snarfed
erinjo++
#
Loqi
benwerd has 65 karma
#
snarfed
congrats!~
#
Loqi
erinjo has 6 karma
squeakytoy2 joined the channel
#
benwerd
Thank you!
#
tantek.com
edited /event (+19) "adjust icon to work with toc"
(view diff)
#
@davidpeach1
I have finally got around to starting to build my new website. #indieweb #laravel
(twitter.com/_/status/573606638613061632)
#
ben_thatmustbeme
snarfed, preview button did not seem to like me today
#
ben_thatmustbeme
was getting ' Error: 400 Bad Request The server could not comply with the request since it is either malformed or otherwise incorrect. Missing required parameter: source_key Missing required parameter: source_key' when trying to preview post of 'https://ben.thatmustbe.me/photo/2015/3/5/2/'
#
ben_thatmustbeme
but webmention sending worked
#
snarfed
ben_thatmustbeme: sorry. try shift-refreshing your user page
#
ben_thatmustbeme
didn't even need that, just a refresh seems to have done it
snarfed, stream7, j12t and lukebrooker joined the channel
KevinMarks_ and e-lima joined the channel
KevinMarks_, benwerd, finchd and alanpearce joined the channel
#
elf-pavlik
"The format is not well defined. Serialising complex data, such as a javascript object, is beyond the specification."
j12t joined the channel
#
KevinMarks
it's well defined for key/values , not for nesting