#indiewebcamp 2014-03-29

2014-03-29 UTC
#
KartikPrabhu
"i am feeling lucky" ?
#
bret
no google reader had a sort by option called "magic"
#
bret
maybe it is related
#
tantek
hey aaronpk - trying your /talk with benwerd here @MozSF
#
tantek
KartikPrabhu: s/i am feeling lucky/get lucky
#
KevinMarks
I've always wanted a "what people I follow are linking minus what everyone is linking" view
#
aaronpk
tantek: iirc the mozilla network blocks webrtc connections
#
tantek
aaronpk - I'm on the internal Mozilla SSID
#
aaronpk
might still be the case. our internal is actually more locked down than our guest.
#
tantek
that BETTER work or else I'm going to file a serious IT request
#
aaronpk
webrtc needs access to random high numbered ports, which is the kind of thing IT departments don't like
#
bret
mozilla guist wifi blocks IRC
#
bret
irrrc
#
bret
guest*
#
KartikPrabhu
is everything #indieweb now!?
#
bret
i have to ssh into an irssi session
snarfed joined the channel
#
Loqi
snarfed: KartikPrabhu left you a message 2 hours, 25 minutes ago: trying webmention discovery on "https://www.brid.gy/about" gives an SSLError.
#
Loqi
snarfed: KartikPrabhu left you a message 1 hour, 9 minutes ago: also sending a mention to https://snarfed.org/ returned a 404. Seems like somethign the wordpress plugin does in general. also happened on dariusdunlap's site
#
sparverius
bret: as you should because irc was designed for continuous connections
#
sparverius
oh the 80s, where "it's text, good enough for me" was a valid deisgn point
#
KevinMarks
irc was very good at slow discontinuous connections for me
#
snarfed
KartikPrabhu: thanks for letting me know. re SSL on brid.gy, its cert depends on SNI. does whatever you're using to fetch support that? details: https://www.brid.gy/about#ssl
#
KartikPrabhu
just using the usual Python requests lib. unsure of specifics
#
snarfed
re mentions to the snarfed.org homepage, you're right, afaik the wp plugin doesn't support homepage mentions
#
KartikPrabhu
cool good to know! will check on the requests lib thingy
#
bret.io
edited /reader (+508) "/* Challenges */ Added Subscription Maintenance ideas"
(view diff)
#
snarfed
KartikPrabhu: fwiw though, brid.gy/about doesn't advertise a WM endpoint
#
KartikPrabhu
wait really?
#
KartikPrabhu
then that shouldn't have happened in the first place... will dig more
#
snarfed
no, if you linked to it, then fetching it to perform WM discovery is reasonable
#
snarfed
so just figure out the SNI support and you should be ok
#
KartikPrabhu
oh yes i see true
#
aaronpk
needs some work :)
#
bret.io
edited /reader (+567) "Added a section about Google reader replacements"
(view diff)
melvster joined the channel
#
bret.io
edited /reader (-1) "/* Post Google Reader Readers */ Fixed capitalization"
(view diff)
#
KartikPrabhu
bret: nice additions, I can add to a Feedly article if there is some blueprint for these
#
bret
Nope! I'm just going to focus on what I find interesting about each service I have tried
#
KartikPrabhu
alright... will add things with UX screenshots and such sometime in the future
#
bret.io
edited /reader (+21) "/* Post Google Reader Readers */ Added reeder"
(view diff)
#
KartikPrabhu
I have quite a few things I like/hate about feedly
#
bret
i dont trust feedly because its free
#
KartikPrabhu
so was Google Reader
#
bret
I know!
#
KartikPrabhu
there is also a Feedly Pro if you'd like to pay
#
bret
oh cool! didnt see that
#
KartikPrabhu
advertised as better search etc... features
#
kylewm
don't trust it to what? stick around?
#
KartikPrabhu
there seems to be quite a lot of read/write free apps that hae shutdown recently
#
KartikPrabhu
e.g. Editorially, Readmill ...
#
bret
kylewm: it has too much of a startup/vc/aquisition waiting to happen smell
#
kylewm
yes, it does, good point
#
bret.io
uploaded /File:13361683735_62ca09ea53_z.jpg "Reeder Screenshot"
#
bret.io
created /Reeder (+572) "Stubbed out a Reeder page"
(view diff)
#
tommorris
http://tommorris.org/posts/8872 - kind of recorded some history tonight
#
KartikPrabhu
where is this tommorris?
#
KartikPrabhu
nice! good good
#
tommorris
First same sex wedding in England.
#
KartikPrabhu
yeah figured that part... :)
#
KartikPrabhu
would that be England or all of UK?
#
tommorris
England and Wales. Scotland starts later this year.
#
tommorris
And Northern Ireland when they stop being run by fundamentalist bigoted nutjobs.
#
KartikPrabhu
very good very good... progress
#
KartikPrabhu
tommorris: on a diff. topic, would you have any insight into this issue: https://github.com/kartikprabhu/mf2py/issues/29
#
tommorris
Too drunk
#
tommorris
on iPhone.
melvster joined the channel
#
kylewm
tommorris: that's awesome, congrats to England!
#
kylewm
KartikPrabhu: speaking of mf2py, I have made the mistake several times of calling Parser(text) rather than Parser(doc=text), do you think we could make doc/url required named parameters rather than part of **kwargs?
connextrum joined the channel
#
tommorris
thanks kylewm
pasevin, sparverius, melvster, grantmacken and kbs joined the channel
#
kbs
apropos of nothing on this nice friday evening :) idly wondering why vcards point to keys, rather than keys wrapping vcards -- the latter permits signatures on the contact info, which seems handy.
melvster joined the channel
#
bret
In sf for a smidge :)
#
bret
Wish I could stay longer
snarfed and melvster joined the channel
#
KevinMarks
did you see the job I mentioned earlier, bret?
pfenwick, rodlogic, pfenwick1 and ben_thatmustbeme joined the channel
#
bret
KevinMarks: no
#
bret
checking scrollback
paulcp joined the channel
#
bret
Oh for Venture beat, ill check it out
#
bret
I dont know php or wordpress, would need to fake it :/
#
KartikPrabhu
kylewm: The trouble is that Parser(string) could be either a url or doc how do we decide. I think the earlier versions of mf2py did some strange guess work on this. But open to better ideas for handling this.
#
kylewm
KartikPrabhu: it's not perfect, but would it be less ambiguous if the signature were def __init__(self, doc=None, url=None, **kwargs)?
#
KartikPrabhu
but then we are saying that not giving either doc or url is ok, which seems bad
#
kylewm
fair point
#
KartikPrabhu
i thought about this before remving the args parsign logic, but I don't knowa good solution. I have made the Parser(text) thing a bunch of times too
#
KartikPrabhu
is raising a value error if neither doc nor url kwargs are found good?
#
kylewm
hmm, does it make sense to ever provide a doc without a url?
#
kylewm
i guess you might be parsing local data, like Falcon's datastore...
#
KartikPrabhu
yes. also the base element can then give th url
#
kylewm
oh, would the exception be raised if you called Parser(text)
#
kylewm
I like that idea if so
#
KartikPrabhu
yeah if you provide no kwargs with names doc or url then you could raise an exception and give out an error message
#
KartikPrabhu
we can combine that with your signature suggestion. so if both doc and url are None then we raise the error
Kopfstein and cweiske joined the channel
#
KartikPrabhu
anyone here know the ins and outs of CSS units?
#
KartikPrabhu
CSS units question: If I declare font-size = 3cm all browsers on all screens should render the font at 3cm correct? but they don't why?
#
cweiske
because they often don't know their DPI
#
KartikPrabhu
cweiske: then is there a common heuristic that browsers use?
#
cweiske
no idea
#
KartikPrabhu
hmm annoying... seems counter to the spec but oh well
squeakytoy joined the channel
eschnou, LauraJ, pfenwick, friedcell, Sebastien-L and connextrum joined the channel
#
@equivalentideas
RT @caseorganic: "This presentation cannot be opened b/c it is too old." Another reason to host HTML presos on your site. #indieweb http://…
(twitter.com/_/status/449867597530402816)
tobiastom, eschnou, jonnybarnes and melvster joined the channel
#
ben_thatmustbeme
So it seems like don't allow old style commenting directly on your site, bridgy and webmention only. Any thoughts on if this is annoying for readers? I really like the idea for not having to deal with spammers.
#
ben_thatmustbeme
and yes i read the wiki entries on /comment, /comment-policies, /webmention
#
ben_thatmustbeme
but how prevelent is comment spam via webmention protocol?
#
bnvk
ben_thatmustbeme: that is being worked on my barnabywalters at the moment
#
bnvk
I believe he's been posting results on http://indiewebcamp.com/spam
#
ben_thatmustbeme
also, good morning
#
ben.thatmustbe.me
edited /comment (+14) "/* Beware */"
(view diff)
kbs, melvster and pasevin joined the channel
barnabywalters, snarfed, LauraJ and kbs joined the channel
#
snarfed
hey cweiske! saw that you tried bridgy. let me know if i can answer any questions!
#
kbsriram.com
edited /pgp (+201) "Better wiki links, reword the user scenario and reasons for additional checks"
(view diff)
paulcp joined the channel
demis and LauraJ joined the channel
#
aaronpk
ben_thatmustbeme: search google for "turning off comments" and you'll find a lot of people who just disabled their comment form
#
aaronpk
Regardless of whether they even know about web mention
josephboyle joined the channel
#
ben_thatmustbeme
yeah, thats what i've been seeing
bnvk and LauraJ joined the channel
#
cweiske
snarfed, it seems that bridgy is of no use if I don't syndicate my posts to twitter
#
cweiske
I don't plan to do that
#
cweiske
I thought I could get tweets that link to my posts
#
snarfed
cweiske: lol. true! bridgy is only useful for connecting indieweb sites to silos and vice versa
#
cweiske
other people tweet my posts since I'm on planet-php.net
#
cweiske
I don't post myself
#
snarfed
cweiske: got it
#
snarfed
funny. the old, pre-webmention version of bridgy actually did that
#
cweiske
so i'm too late
LauraJ, iangreenleaf, iangreenleaf1, snarfed and ttepasse joined the channel
#
aaronpk
if you run your own twitter search using the streaming api you can use the bridgy urls for the microformats markup
LauraJ joined the channel
#
aaronparecki.com
edited /mention-app (+482) "/* App Functionality */"
(view diff)
tomme86, indie-visitor, josephboyle and LauraJ joined the channel
#
kylewm.com
edited /Red_Wind (+344) "info about switch to flat files"
(view diff)
#
aaronpk
I have lots of friends who still read twitter all the time, and it seems like anyone still reading the whole home timeline is following ~200 people or less
#
aaronpk
i'm following 800 people and my timeline is pretty much useless to me
#
aaronpk
so I'm wondering if I added under 200 people to an indie reader (peoples' own sites, and also some twitter accounts) if I would actually find it useful again
#
kylewm
I follow 127 and it's a full-time job to read it all...
#
demis
out of curiousity, what readers are you guys using for indie sites, and do any of them stand out as being more 'indie friendly'?
#
aaronpk
i seem to have broken something in my selfoss install :(
#
aaronpk
also I have little motivation to use a reader until I can actually interact with the content (reply, favorite, repost, etc)
#
aaronpk
I think that's the main reason interfaces like twitter and facebook have taken over readers
#
demis
I haven't used a reader since Google took thier's offline, and even then wasn't really using it - probably for much the same reasons you mention above.
#
demis
that's one thing that would be great added to Idno - a simple private reader page that collects the latest posts from sites I follow and can be ordered/grouped by time/site. Yes I should see what I can do about that :)
#
kylewm
aaronpk: have you needed to implement write-locking for your file storage scheme?
#
aaronpk
not yet
#
aaronpk
there are few things that actually write to the same file
#
aaronpk
actually the only thing that does is the webmention handler, but I run that as a job queue so there's only ever one thing writing to my webmention log file at a time
#
kylewm
so that may be a weakness in my design :) i'm storing comments inside their parent post
#
aaronpk
I was doing that, and ended up moving them to a separate location
#
aaronpk
it was causing more problems on the git side, not actual filesystem concurrency
pasevin joined the channel
#
kylewm
ah interesting, git problems in that you might change the post locally and it receives comments remotely
#
aaronpk
lots of detail in that issue thread
#
kylewm
now i need to learn about indie archive
#
ben_thatmustbeme
Everyone seems to use a single URL for the webmention handler, but is there any reason not to have a different one per post?
#
aaronpk
ben_thatmustbeme: what would your reason be for using a separate URL?
#
aaronpk
i mean you certainly can, the spec supports it and all the clients will work just fine if you do
#
ben_thatmustbeme
it would make input parsing actually easier on the back end for me, I already am associated to the exact post
#
aaronpk
ah instead of finding your post via the target URL?
#
ben_thatmustbeme
i could technically ignore the target=
#
ben_thatmustbeme
so instead of post/2014/.... i could just do webmention/2014/....
#
aaronpk
hm interesting... might want to still look at the target URL to verify it's something sane
#
aaronpk
you'd need a special case for webmentions of your home page
#
aaronpk
also do you not care about webmentions of other pages that are not posts? like list pages and such?
#
ben_thatmustbeme
yeah, i'm thinking it as a way to cut down on spam too, if they don't agree i just disregard as out of spec
#
kylewm
is the content of IndieArchive something that actually is published somewhere?
#
aaronpk
kylewm: right now it's just a concept, and implemented within individual platforms
#
aaronpk
in theory we could have a shared repository of the content, but that hasn't been needed by anyone yet
#
ben_thatmustbeme
well there shouldn't be a response to a page that isn't a permalink
#
kylewm
ha, that's such an interesting use case for DVCS...would've never thought of it
#
ben_thatmustbeme
and any permalink page would have its own webmention URL
#
aaronpk
ben_thatmustbeme: it depends on whether you care about showing generic links to your pages, not just responses
#
aaronpk
s/links to/mentions of
#
Loqi
aaronpk meant to say: ben_thatmustbeme: it depends on whether you care about showing generic mentions of your pages, not just responses
#
aaronpk
oh and s/showing/knowing about
#
aaronpk
(whether you display those mentions is a different story)
#
ben_thatmustbeme
hmm, true. I suppose the other pages could have a generic handler too. I'm actually thinking more toward a webmention callback for phonebook stuff that I want to work toward
wildtype joined the channel
#
ben_thatmustbeme
yeah, I am not sure on the exact protocol for it, but basically use webmention callback for the author page to say thay I added you to my phone book. You could see who has you added and add them to yours as well (but more importantly figuring out some extension to add more secure info to the author page if you are added back)
#
aaronpk
oh interesting. sounds kind of like how twitter shows you when people follow you
#
ben_thatmustbeme
there are a number of security issues i'd have to look in to, but really, phone number, address, things like that aren't high security, just moderate
#
ben_thatmustbeme
the idea being to have people keep track of their own contact info, and other's can just pull that info in
#
ben_thatmustbeme
but you could always organize people in to groups for access, maybe close friends get my mailing address, but acquaintances just my e-mail and city
#
ben_thatmustbeme
thats why i was interested in kbs's discussion of sharing pgp keys on your home page, gives a way to have secure communication back and forth
#
aaronpk
and you'd want my site to be able to notify yours if my contact info changes
#
ben_thatmustbeme
prevents having to re-pull data every once in a while
#
aaronpk
sounds kind of like a PuSH subscription
#
ben_thatmustbeme
hmm, i suppose that could be a way about it
#
ben_thatmustbeme
i was thinking if it generates a custom URL (with a key) that grants access its a unified interface, webmention just becomes a negotiation of keys
#
ben_thatmustbeme
but I could generate a key for a friend who does not use Indie web
#
ben_thatmustbeme
and then they can either visit the custom URL, or use an app that pulls it in and parses h-card
#
aaronpk
this is intriguing, but I need to get back to refactoring p3k...
#
ben_thatmustbeme
okay, have fun
ttepasse, snarfed, barnabywalters, kbs, tilgovi, bnvk and hober joined the channel
#
demis
hey aaronpk and ben_thatmustbeme, reading your text above brought Bitmessage to mind. Any way to piggy back off the generated identities in that for the contact info?
#
aaronpk
i dunno, sounds complicated
#
aaronpk
my eyes glaze over pretty quick reading https://bitmessage.org/wiki/Protocol_specification
#
demis
lol
#
demis
perhaps the way Bitmessage keys are generated and distributed could provide a framework to help your contact list challenge
#
ben_thatmustbeme
updated my current thinking on it, without they security issues in there yet
#
aaronpk
I should just do this on my own /contact page right now... anybody can sign in to my site using indieauth right now, so it'd be easy to add private data there if you're signed in
#
Loqi
fo sho
#
aaronpk
i already support private messages if yuo're signed in
#
ben_thatmustbeme
seems like a good way about it
#
kbs
pretty interesting idea ben_thatmustbeme
#
aaronpk
I think it would cut out a few steps you've listed
#
ben_thatmustbeme
my concern is just the non-indie web users
#
aaronpk
what do you mean?
#
ben_thatmustbeme
but you are probably right, it would make things much easier for connections
#
aaronpk
you want to be able to give private contact info to people who don't have their own site?
iangreenleaf joined the channel
#
aaronpk
no really, what is your concern?
#
aaronpk
oh I see... "Friends without webmention systems can just ask for a URL that I can e-mail them."
#
aaronpk
ben_thatmustbeme: are you familiar at all with OAuth 2?
#
snarfed
aaronpk: love that you're using ownyougram now
#
snarfed
it looks great on your site!
#
aaronpk
thanks!
#
aaronpk
i'm super excited about it
#
snarfed
i haven't yet carved out the time to implement the req'ts on my site (notably oauth/token server and micropub), but i want to eventually
#
aaronpk
cool. I'm wondering if there is anything I can build to help people out with that
#
aaronpk
like would you be interested in using a token endpoint service, so all you have to build is the micropub endpoint?
#
snarfed
honestly i haven't thought about it enough to say
#
snarfed
the answer to any wordpress question is usually, "there's a plugin for that," so i should look there first
#
snarfed
i eventually will when it bubbles up on my todo list, and i'll let you know
#
aaronpk
ah right wordpress. i doubt there's a plugin yet :)
#
snarfed
…in the meantime, looks like the commenting part of ownyourgram isn't there yet. have they still not gotten back to you?! sheesh
#
aaronpk
no reply from the official channel. I now have someone on the inside trying to push it to the right person.
#
snarfed
thanks
#
snarfed
if you hear (or can ask for) any meta details, e.g. criteria or red flags, i'd love to hear
#
snarfed
on a similar note, i should ask on SO why their like API never works
#
aaronpk
will definitely do so!
#
aaronpk
you've had problems with the instagram API?
#
snarfed
the like api specifically. i've never managed to send a like through it once
#
aaronpk
oh, weird
#
snarfed
i'm guessing it requires the comment whitelisting and they just didn't mention that in the docs
#
aaronpk
what error do you get?
#
aaronpk
hm actually it works for me
#
aaronpk
did you request the "likes" scope during auth?
#
aaronpk
whoa, when I tried posting a comment just now I got a response pointing me to this URL: http://bit.ly/instacomments
friedcell joined the channel
#
aaronpk
that's new, it didn't do that before. just failed with a cryptic error
#
aaronpk
also, come on guys. bit.ly and google docs? you guys have all of facebook at your disposal, you'd think they'd allocate a few resources to make a better developer console for requesting comments access
#
KartikPrabhu
aaronpk : this is instaface?
#
aaronpk
yeah, instagram returns a bit.ly url in the API response which points to a google doc where you can request comment whitelisting
#
KartikPrabhu
wow! so many failure points!!
#
aaronpk
alright, who here is good at naming things?
#
KartikPrabhu
raises hand...
#
aaronpk
I realized that I want to do what I did with instagram, but with a bunch of other silos: strava, last.fm, etc
#
aaronpk
basically PESOS as a service
#
KartikPrabhu
like ownyourgram
#
aaronpk
ownyourgram.com is PESOS as a service but only for instagram
#
aaronpk
rather than making up individual sites for each, I shouldj ust make one service for all of them
#
aaronpk
like bridgy
#
KartikPrabhu
bridgy is an awesome name btw
#
aaronpk
totally
#
KartikPrabhu
so this would be sort of opposite-bridgy which PESOS instead of POSSE
#
KartikPrabhu
Hmm I had a name for a package I was to write that seems apt "Pullback" ?
snarfed joined the channel
#
snarfed
aaronpk: back. interesting that it works for you. i should try again. (yeah, i requested the likes scope)
#
snarfed
and i got the bitly link too, a while back before iwc sf, when i tried commenting w/o whitelisting
#
snarfed
hey, incidentally, when you request IG scope, do you separate scopes with + or , ?
#
aaronpk
oh? strange! it just returned a cryptic error for me when I tried last (after iwc sf)
#
aaronpk
with a space (+). When I tried with , it said invalid scope
#
snarfed
most oauth sites use comma but IG says they use plus: http://instagram.com/developer/authentication/#scope
#
snarfed
good, ok then
#
snarfed
i'll try liking again
#
aaronpk
it was one way in the official oauth spec and it changed before it was finalized
#
snarfed
ah interesting
#
snarfed
hmm. if i get IG likes working, i'd still be reluctant to open bridgy publish signups for IG if all you can do is like
#
snarfed
cross that bridge when i get to it
#
aaronpk
...no pun intended
#
snarfed
re names, a while back i picked a name for (eventually) rebranding activitystreams-unofficial, and i brainstormed a few agriculture related names, due to "silo"
#
snarfed
…none were very good though
#
aaronpk
i could just pick a cute animal name, that usually works
#
snarfed
another theme i thought about was "naked," as in posts on your own site are naked, just you, no silo wrapper
#
snarfed
…probably not worth thinking about too deeply though :P
#
aaronpk
nakify.me
#
snarfed
nud.ie
#
snarfed
off to domai.nr !
#
KartikPrabhu
hmm my head is not going beyond "pullback" :(
CheckDavid joined the channel
kbs joined the channel
#
aaronpk
yeah i'm drawing a blank
#
KartikPrabhu
also "repossess" but it has bad financial connotations
#
KartikPrabhu
i think "pesos" might be good on its own
#
KartikPrabhu
peso = weight or burden in Spanish
#
kbs
is it just me, or has something changed with indieauth.com [eg: insisting that redirect uris be https only?] - seems like login from indiewebcamp.com is no longer working
#
aaronpk
kbs: nothing should have changed
#
aaronpk
kbs: can you click that for me again? i'll watch the logs
#
kbs
well, what do you know. Now it went through
#
kbs
wonder what I did different - hm.
mydjey joined the channel
#
kbs
got the error again - is there anything you want me to pull out while I have the error page on my browser?
#
aaronpk
is your site on github pages?
#
kbs
yes
#
aaronpk
they seem to *sometimes* return a redirect header when visiting your root domain (with no trailing slash)
#
kbs
oh, how weird.
#
kbs
:-) that is weird
#
aaronpk
i need to make indieauth understand relative redirects, then this problem shoudl go away
#
kbs
what the heck is it supposed to do - GET / --> Location: /
tilgovi joined the channel
#
aaronpk
if you then make another request exactly the same, it'll return the HTML page
#
kbs
yep. That's exactly what I get too on my root domain too, every once in a while
#
kbs
GET / -> Location: / doesn't make a whole lot of sense
#
aaronpk
i feel like this is a thing github pages should not do, but I should also probaly handle that case better anyway
#
kbs
yes, it's clearly a bug on their end - thanks for pointing it out - wonder if this will cause issues for other clients. I thought browsers would immediately flag this as a loop, surprised it hasn't shown up
#
aaronpk
it doesn't turn into a loop because by the time you make the second request they return HTML
#
kbs
hm, they don't flag the very first redirect as a loop? interesting
#
kbs
didn't realize that
#
kbs
I wonder linking to http://kbsriram.com is actually better than linking to http://kbsriram.com/ :-) - maybe the browsers are not considering those urls to be equivalent, even though they are making the exact same request
#
snarfed
hey KevinMarks, i'm (finally) looking at that twitter retweet api thing you mentioned a while back: https://github.com/snarfed/bridgy/issues/104
#
snarfed
any chance you remember a specific tweet or response where bridgy clearly used the old wrong way?
#
snarfed
low priority
#
jjuran
Location must be a URL, not a mere path
#
aaronpk
jjuran: I thought so too, but apparently the HTTP spec says it's ok for the Location header to be just a path
#
jjuran
"The soon-to-be obsolete IETF standard requires an absolute URI to follow a Location header, which means it must contain a scheme"
#
jjuran
But it should still be invalid if the client has specified HTTP/1.0
#
aaronpk
pretty sure I've only actually seen http 1.1 at this point
#
aaronpk
also github pages returns http 1.1
#
jjuran
The only HTTP 1.0 client I'm personally aware of is me, using telnet. :-)
#
kbs
clearly, github is also 0.9 non-compliant ;-) http://pastebin.com/m66xEPdw
paulcp joined the channel
#
aaronpk
heading to apex brewery at 12th & division in portland if anyone wants to join me
#
kbs
hm, how does one identify "the" vcard on an indiewebpage, if it has multiple ones?
#
snarfed
kbs: look up the representative h-card algorithm
#
kbs
nice. So the indieweb pages have to be careful enough to remove or reposition rel=me tags that are auto-included on their pages, I guess?
#
snarfed
kbs: which indieweb pages are those?
#
KevinMarks
snarfed: it shows up when a retweet includes a webmentioned page
#
KevinMarks
so if I tweet a link to an indieebcamp page, and you reweet it
#
snarfed
KevinMarks: hmm ok. i'm trying to figure out if or what i'm doing wrong. i don't actually use the content of retweets at all, whether from the outer object or retweeted_status
#
snarfed
ok, i'll look
#
KevinMarks
it will construct a tweet saying RT @kevinmarks: [my tweet truncated]
#
kbs
was actually generally thinking about how people might find ways to sneak in a rel=me link to claim ownership of a desirable domain.
#
KevinMarks
people can claim ownership one way, but the domain needs to link back
#
snarfed
KevinMarks: yes. i understand the general issue
#
snarfed
i'm trying to figure out if/how bridgy specifically is handling it wrong
#
kbs
KevinMarks: right - rel=me to point to [say] a fake twitter account which could then be used to point back to the desirable domain - and then (say) an indieauth login as the desirable domain?
#
KevinMarks
so you shouldn't allow people to post rel=me links on your page, yes.
#
KevinMarks
snarfed: I remember seeing an example, but can't remember where
#
snarfed
KevinMarks: ok. do you remember if the author or content was wrong?
#
snarfed
i don't use content at all, and i believe i'm using the right author (the retweeter)
#
snarfed
anyway. if you remember the example, definitely let me know. thanks!
indie-visitor joined the channel