#bearit can handle the webmention indirectly - my thought was that it would be the library that any webmention work is done for another small code that is a http listener - does that make sense
#bearprocessing incoming webmentions is two parts: first you need something that is running under an http server to receive the POST and then you need all the code to decode that payload and process the webmention
#snarfedhey bnvk, saw you signed up for bridgy. welcome!
#bnvkbut no biggy if it's not really setup that way
#aaronpkbnvk: no it'll work fine, you can use that URL for any site
#bearthe second part is what I thought ronkyuu would do
#KartikPrabhubear: oh I see. So ronkyuu is doing the later with findMentions... :)
#snarfedbnvk: looks like bridgy had webmention discovery for bnvk.me cached as no endpoint found. i deleted that, so it should see your endpoint in the future
#bnvksnarfed: ah interesting, so it will re-poll whenever the job runs again?
#snarfedyeah, it re-polls every 15m, but it won't retry any of the existing responses that it marked as complete because it thought bnvk.me didn't have a WM endpoint. it'll send future responses though
#bnvkhehe, PDX coffee is a lot stronger than Iceland kaffi
#aaronpkis at fresh pot on hawthorne! always seem to get a lot of stuff done while i'm here!
#bnvkme too, perhaps the large quantity of books emits a certain frequency conducive to thinking deeply
snarfed joined the channel
#aaronpkok! indieauth.com is now on the new server!
#aaronpkit should be faster, and should also not have the twitter ssl problem anymore!
#kbsaaronpk: am I correct in thinking that when checking rel=me links from the proivder, you're directly parsing [and then following through any redirects] rel=me links from them? (as opposed to using an api call offered by the provider)
#aaronpkyes, it fetches the HTML page, not using the provider's API
#aaronpk(that's why facebook doesn't work, cause they don't return any sensible HTML on public profile pages)
#acegiakaaronpk: my webmention comments on your posts are only displaying as hyperlinks to the mentioned posts. is this normal behaviour for your system?
#acegiakaaronpk: the p-name is the title of the post? the post has both e-content and p-summary on the actual note content?
#bnvksnarfed: maybe I don't have any content that has the right kind of activity on it? I'm even trying to manually enter post URL and not seeing anything https://brennannovak.com/notes/410
#snarfedbnvk: re responses, yeah, sorry, more caches to clear. guess i didn't design it to selectively clear the cache. i'll just flush
#aaronpkacegiak: looking at your post in my browser, I don't really see what would be the name of the note
#bnvkfriggin caching... isn't there some quote bout caching be the devil
#aaronpkbnvk: it's one of the two hard problems in computer science
#aaronpkalong with naming things and off-by-one errors
#snarfedand it should have also handled the bnvk.me redirect to brennannovak.com and found that domain ok, instead of complaining. i'll take a todo for that too
#bnvkI wish IRC was smart enough / nice UI that it would auto break off threads between two people (e.g. me and snarfed) but keep it in the same public room view
#acegiakaaronpk: the way I create posts that are replies/reblogs/likes I have a form that prefills the title of the post with the format [POSTERNAME]:[POST TITLE] which I can then change if I want to
#aaronpkacegiak: interesting, is that a convention you picked up from somewhere?
#bnvkacegiak: brains, yes, but requires much scan time / brain cpu cycles to read msgs not for me ;)
#acegiakaaronpk: I think probably tumblr? I'll see if that's how they do it
#aaronpkbnvk: lol that was me that said that. point taken. :)
#acegiakaaronpk: the title thing in mine is the corssover between the way tumblr handles these things and what I already had built for when whisperfollow was soley rss based
#aaronpkso this may mean comments-presentation is underspecified, or we need another guideline for how to display "mentions" that are not explicitly in-reply-to
#aaronpkbut separately, your post does appear to be a reply, so should probably have the in-reply-to class
#bnvkwhich is odd cause there's a conditional before that which asks if(is_array($headers['Link']))
#KartikPrabhuaaronpk: mf2py does what tantek said. so does php-mf2 afaik
#aaronpki have a test suite, i should add your html to it
#aaronpkok wait i'm gonna go back to this p-in-reply-to issue
#aaronpktantek: are you saying I should see the same h-cite object in both a "like-of" and "in-reply-to" property? (parsed JSON for reference http://pin13.net/8sV )
#acegiakwell I've just fixed that issue with the h-* properties
#tantekdrat - 504 Gateway timeout on indiewebcamp.com :(
#aaronpkacegiak: no, the "u-*" properties make the parser go look for an "href" attribute for the URL
#tantekacegiak - no one should be looking for both u-like-of and p-like-of - that doesn't make any sense
#snarfedbnvk: also figured out why bridgy publish didn't follow the bnvk.me redirect. it doesn't (yet) know about the Refresh header. not sure if i'll add that, but feel free to file an issue if you want it
#aaronpkthe one-letter prefixes are for the parser, not the actual consumer of the data
#acegiaktantek: the question was whetehr they would be looking for one or the other
#tantekthere is only the "like-of" property - the prefixes tell the parser where to parse for it
#bnvkaighty gents, time for this lil hacker to unplug- g'night
#tantek.comedited /Federated_Social_Web_Summit (+573) "/* Portland 2010 */ make articles its own subsection, add post re: registerprotocolhandler, expand search for posts into actual list of posts" (view diff)
#KartikPrabhuaaronpk: fixed to return JSON. url parameter thingie postponed for the time being
#tantekman I hate GoogleGroups permalinks. much markup. so slow to load.
#KartikPrabhuagrees. have started avoiding GG links when I search for code help
#KartikPrabhuits all the JS trickery that Google seems to love nowadays
#tantekaaronpk - looks like www.indiewebcamp links are broken
#tantekit's in Düsseldorf - who here is Germany again?
#tantekalright - I'm going to put it in the /events list - as we don't have any IndieWebCamps in May anyway - and this seems like a good enough substitute
#aaronpkdidn't they pop in here a while back asking about something?
#tantek.comedited /Events (+403) "add Decentralize Camp as it looks similar / complementary to IndieWebCamps, and heck, we didn't have any camps planned in May anyway" (view diff)
#tantekaaronpk - I think they did, and we asked a bunch of questions, and they went back to go work on it
#tanteksince then they've signed up some speakers including adactio and one of the Hoodie guys
#KartikPrabhuaaronpk: I would eventuallylike to supersede Django's admin interface for posting. I feel like struggling with modifying Django would be futile and I should just use HTML+JS to do some micropubbing
#aaronpkcause they use it as a regular word in their own docs
#KartikPrabhuaah so there is my second suggestion "scrobbler" or "scrobby"
#aaronpk(trademarks can only be adjectives modifying a generic noun)
#aaronpkso by using forms of "scrobble" in their docs like "scrobbling" and "scrobbles" and using it as a verb, they prevent themselves from being able to trademark it
#aaronpkhah yeah, github registered their trademark in clothing!
#aaronpki would send you a link, but the uspto site doesn't work taht way
#snarfedben_thatmustbeme: g+ is either manual, private api via biz dev relationship (e.g. wordpress.com and jetpack), or screen scraping (e.g. nextscripts snapp, maybe mailchimp social)
#snarfedafaik the moments api doesn't really surface to users yet
#ben_thatmustbemethats what i figured, its rather rediculous that they haven't opened that API yet
#ben_thatmustbemei can imagine they don't want to get innundated by auto-post junk though
#KartikPrabhubear: I wrote a Python object (long back) to help send webmentions to some post. Automatically discovers mentions by scanning content of post using ronkyuu. Functions to send mentions to combinations of past and present mentions. Is this something that belongs in ronkyuu itself, I am happy to add it there.
#KartikPrabhu"Look Facebook, all we’re saying is that we wanted to share a joke about chicken wings, not ponder astrophysics." that is just bad complaining it I ever saw one.
#KevinMarksI need to write this up, but the mapreduce worldview assumes you cna create databases from documents
#KevinMarksand that databases are inherently volatile and iterative
#KevinMarkswhereas the SQL worldview sees databases as ideal and pure, and documents as volatile reports you generate from them
rtaibah joined the channel
#KartikPrabhuKevinMarks: They can both be inter converted of course. Which on is "pure" and which is "volatile" depends on what is using it and how I think... Would be an interesting blogpost
#KartikPrabhuas a complete aside saw the Dave Winer wiki page. Man the talk page is contentious
#tommorrisKartikPrabhu: to be fair to Dave Winer, the contention on the WP talk page is undeserved. Most of it was drama stirred up by a user who was banned.
#tommorrisKartikPrabhu: I’d forgotten I was involved in that ban. ;)
#KartikPrabhuoh of course. not blaming Winer at all for that
#tommorrisNot only was I involved in the ban, I actually dropped the block. Funny how quickly I forget the history on that.
eschnou, barnabywalters, KevinMarks, dns53, bnvk, pfenwick, friedcell, cweiske, voxpelli, jonnybarnes, rtaibah, melvster and Jihaisse1 joined the channel
#bnvkis there a page on the wiki regarding migrating one's indie web site to a new domain?
scor joined the channel
#voxpellibnvk: any particular challenges in regards to moving a site to a new domain that you are thinking about?
#bnvkvoxpelli: specifically the ability for others sites to update my syndicated data, should I send them some sort "hey, my home has moved from domain1.com/note/213 to dom2.com/note213
#voxpellibnvk: Sending a HTTP 301 status back from the old domain with a location-header pointing to the new one would be enough as I see it
#voxpellibnvk: for each page on the old domain that is
#bnvkvoxpelli: right a 301 from the old domain seems like the right first step
#bnvkjust curious if we should document this somewhere so to start thinking about the flow / approach for notifying sites of this 301 so they can handle it properly & I don't have to keep paying for old domain indefinitely
#voxpelliif I were to have done all of my tweets indie web style then that would mean thousands of notifications of I ever were to move to a new domain
#bnvksurely one wouldn't want to say update 10K+ messages NOW, but a que that churns through it
#barnabywaltersin theory an update webmention could be sent for each one, which returns a 301 redirect, and some piece of software goes through and sends them all
#voxpellibnvk: I think that in the case of a pubsubhubbub subscription it makes lots of sense as the other sites has handed the responsibility of notifying it over to you
#barnabywaltersvoxpelli: yeah, I keep an archive of everything I link to so one thing I considered was recrawling periodically, and replacing the links with links to the archived versions if the original 404s
#voxpelliin the other cases I’m not so sure – a link is a link – no matter if it was made through webmention or made manually
#bnvkseems like there could be 2 approaches to this... sites crawl their own comments detecting the health of domains, or sites send notifications to others sites to "update me"
#bnvkI mean, I think this sort of thing would be useful even if one does not change domain, but say changes publishing platform
#voxpelliI mean – people likely links to you outside of your webmentions to them – those links would ideally have to be updated as well
#brennannovak.comcreated /migration (+1250) "Created page with "Ideally people will create their IndieWeb sites and their data will live at the same URL forever and ever into eternity. In reality, it seems this is unlikely, as people do cha..."" (view diff)
#ben_thatmustbemeif you run that last one and change ay /gi to ay/gi you have it
mgarrido and Sebastien-L joined the channel
#bnvk!tell snarfed would you ever consider hacking Bridgy to be it's own app that's easy to install a given users server a la http://indiewebcamp.com/store
#Loqisnarfed: bnvk left you a message 28 seconds ago: would you ever consider hacking Bridgy to be it's own app that's easy to install a given users server a la http://indiewebcamp.com/store
#snarfedhonestly, i don't really expect scaling issues. it's pretty efficient right now, and scaling would only cost money. still, happy to have people try porting it
#snarfedmost of the work would be devops, not writing code. if someone else wanted to drive, i'd be happy to hep!
#bnvksnarfed: also IIRC, Thinkup has a really heavy front-end focused on providing a social-media dashboard
#snarfedbut adoption is arguably the hard part. it arguably doesn't matter that ostatus/activitystreams exist, sadly, because no one who matters uses them
#bnvkThinkup is quite cool & really helpful for accounts with large followings, but seems quite different than the task of just saving my tweets + facebook activity to disk that I can then render in a conversation thread on my site
#voxpellisnarfed: yeah, so sad to see many new initiatives focus on the technical aspects of it – as if that ever has been the problem over the last couple of years
#snarfedvoxpelli: eh, we're hackers, it's what we do. we have a hammer. :P
#voxpellisnarfed: hackers hack – hackers don’t talk ;) we build then standardize – not standardize, then re-standardize
#snarfedvoxpelli: heh, true! but in this space, neither hacking nor standardizing drives adoption, necessarily
#bnvkvoxpelli: agreed, it is sad, but like I tinkered with activity streams on my Social-Igniter project for ages, I never found the community materialize / other sites to talk to + the whole OSTATUS stack was ridiculously complex
#bnvksnarfed: awesome!!! didn't realize you had done sooo much in that space
#snarfedthanks! and fwiw, activitystreams-unofficial *does* include responses. would be pretty easy to write a minimal main.py and stick it in a cron job to grab your own stuff plus responses. not polished for others, but something
#bnvki'm borrowing this line of thinking as per the unix philosophy of have one piece of software do one small task really well, figure out how to make those components work with other components
#snarfedbtw, the archive downloads that FB and G+ and Twitter all offer are now really good. more for end users than hackers, and not set up to regularly sync, but still
#bnvkvoxpelli: there's no reason we can't, in fact h-entry would probably make a lot of sense
#snarfedbnvk: we can, but i'm not sure we should. it's a big and different problem. activitystreams-unofficial would be a perfect building block though!
#bnvkI definitely don't have the bandwidth cycles to work on this at present except for advising / UX flows stuff, but seems like a really worthwhile thing
#gRegor`KartikPrabhu: It's not huge and I'm not sure how busy / noisy it might be Wednesday at 6:30
#tantekgRegor`, KartikPrabhu your indieweb usage doesn't really deserve to be hidden in an anti-pattern article. I'm going to make a /MySQL article and move your usage there
#tantekgRegor`, KartikPrabhu what's the name of the software/project CMS's that you're running on your sites? I left ??? in place here: http://indiewebcamp.com/MySQL
#KartikPrabhutantek: I'm just using A custom Django installation. No name other than that :P
#tantekKartikPrabhu: that's ok - you can say that - or you can name your "project in the making"
#tantekactually I thought I saw you did - like on your User: page...
#kbsKartikPrabhu: I think I would offer https://github.com/kbsriram/kbsriram.github.io if I didn't also feel that my ragtag of random scripts would probably drive most people off the use-documents-vs-database principle :)
#KartikPrabhugRegor`: are we trying to sync with the SF/PDX homebrew website club meeting?
squeakytoy, benprew, snarfed and j12t joined the channel
#Loqij12t: tantek left you a message on 3/18 at 10:27am: are you still using Open Photo / Trovebox "running a private instance for … family pictures" ?
#kylewmok thanks, that's encouraging :) i'll put that on my todo list. someone at IWC SF said something along the lines of 'first do it manually, and then once you've figured out if you want to do it a lot, worryabout automating it' ... lots of fun stuff you can do manually
#tantekKevinMarks - not sure where I saw that eat24 thing - maybe randomly on Twitter?
#tantekI dipped in briefly to see if my suspicions about too many "comment" tweets was still true
KartikPrabhu, Kopfstein, pasevin and kbs joined the channel
#kbsso, I have a tiny graph problem puzzle and a silly solution :) you can poke holes it [or if you prefer, solve independently and compare if you feel inclined]. First, the puzzle (and it's related to my current playing with rel=me links)
#kbsSo I have a directed graph, with a special source node - I know this source node is connected to all the other node via some path. Okay? that's the property of this graph
#kbsnow I want to find the strongly connected component of this graph, that contains this node [ie, every pair of vertices of this subgraph is connected]
#kbs(the context being essentially, a proof graph that locks all these sites as being connected via rel=me links, if you're curious)
yaf joined the channel
#kbsSo I have a pretty dumb solution, that I can't convince myself is wrong - but figure you folks might want to have a bash at a solution first, so let me know if any of this sounds interesting to you :)
#kbstantek: ah, ok. This was more in the nature of a puzzle really, and more to see if someone else arrived at the same dumb idea as me :), I'll poke there to see how they did it
#tantekI think an abstract puzzle like that is a math problem / theoretical.
#tantekIn the context of rel=me, there's a lot more information you can use to develop simpler solutions.
#kbshm, identengne must be doing something different from what I'm thinking, clearly
#KevinMarkswe could ask bradfitz, who wrote the rel-me crawler for google
#kbswhat I'd love to know actually, is if doing a dfs on the reversed graph is correct (more curious about the algorithm actually :)
#kbsie, assuming I've crawled from site "source", doing a dfs on the reversed graph of rel=me links gives me all the sites that have a chain of connections both ways to "source" - more just a puzzle I happend to stumble upon, etc...
paulcp and bnvk joined the channel
#kbsI'm probably missing something in how identengine works, but as far as I can see so far it is (was, I guess :) mostly delegating the query to the social graph api maybe?
#KevinMarksthe social graph api would start form a url and return all linked urls and whether they linked back
#gRegor`Doesn't fall under a traditional site-death since it's more an app in conjunction with the site, but Readmill (ebook app) is shutting down: https://readmill.com/epilogue
#gRegor`You could post reviews and comments on the ebooks though, so some content creation.
#aaronpk"Temporary redirection. The request should be repeated verbatim to the URI specified in the Location header field but clients should continue to use the original URI for future requests."
friedcell and rtaibah joined the channel
#tantek.comedited /communication (+963) "/* Articles */ add challenge/problem subsection and link to Can We Talk article, add h-cite markup" (view diff)
#aaronpkgithub will often return a 302 and (relative) location header back to the same URL that you request, and on the second request it returns the actual HTML
#KartikPrabhuI think the improved shortening is great! But it still does not match Twitter's own shortening, which is what people might expect. any reason?
#tantekyeah, precomputing tweet lengths is a bit of a pain, but doable
pasevin joined the channel
#pdurbinKartikPrabhu: you make it sound like XMPP is going away
#snarfedtantek: yup. not as hard as the other way around though. the alg to linkify entities (mentions, hashtags, links) in tweet text takes some serious bookkeeping