aaronpkindiewebify seemed to resolve the twitter URL fine, but it's having trouble "fetching" the URL which is probably due to google appengine's weird URL handling
aaronpkdo you have some other plugin that's changing all your URLs to protocol relative URLs? it looks like almost every link on the page is protocol relative
gRegorLoveben_thatmustbeme has something (maybe it's that rel=contact), which I thought his Vouch code consumed. Might just be rel=me that points to a page on the same domain though.
LoqiXFN is an abbreviation for XHTML Friends Network, the network of visible links across blogs that claim various XFN relationships with/to each other https://indiewebcamp.com/XFN
tantekgRegorLove: no, it's just a phase, we brainstorm things, then prototype / experiment, then if things get traction (publishing & consuming), then we start specing/formalizing
LoqigRegorLove meant to say: Think I'll keep it simple and just allow the user to paste in a URL that has h-cards and use those as the whitelist basis.
gRegorLoveDoes it make sense to say "http://gregorlove.com/following/ is following Tantek", or shouldn't it be "http://gregorlove.com/ is following Tantek"?
tantekpart of the challenge with a lot of the newer proposed rel values is that they were proposed without specific consuming code use-cases (that anyone had any intent of actually building - beyond just "someone could use it to do x")
Loqifollow is a common button in silo UIs (like Twitter) that adds updates from that profile (typically a person) to the stream shown in an integrated reader, and sometimes creates a follow post either in the follower's stream ("… followed …" or "… is following …") thus visible to their followers, and/or in the notifications of the user being followed ("… followed you") http://indiewebcamp.com/following
tantekand the challenge is that since Google Social Graph API shutdown, no more popular (or even minor?) consuming applications (beyond rel=me / RelMeAuth / Web Sign-in / IndieAuth)
tantekI'm willing to bet there is *some* sort of improvement(s) we can make to IndieWeb plugin, and IndieWebify.me, and even IndieAuth to better detect specific errors and provide more specific actionable steps to fix problems
tanteksecond, since that's the current state, IndieWeb plugin could check for the cases that cause those errors, and help the blog author work around them
tantekUsually more feedback is a good sign, it means more people are 1) trying it, and 2) getting far enough to think it's worth the time to ask for help (rather than just give up)
GWGIn the pull request I sent pfefferle, I rewrote the plugin to allow for drop in replacement of the current synchronous handler with an asynchronous one.
tantek.comedited /Planning (+802) "note IWC LA has a page and appears to have fairly firm dates, New York City 2 planning! Dates and some initial +1/-1s" (view diff)
[shaners]Pivotal is providing lots of space. A handful of people to be around to help with logistics, I expect several Pivots will attend as well. The office has very fast internet. AV gear already set up in the common area (intros / demos) and video chat in all of the conference rooms. They will provide at least one meal, maybe more. I’ll prolly line up another sponsor or two just to help out. But it’s a pretty full package here.
[shaners]I’m giving a lunch time 10 minute talk about #indieweb, IWC and (briefly) Dark Matter later this month to the office. And then we’ll be doing weekly hack nights to build up to IWC in November.
tantekhmm, I think I found an interesting (quirky? rare?) person-tagging use-case, specifically, for when I *only* want to person-tag IF I can be assured of the person-tag having a specific point (or rect) on the photo
tanteknormally when I person-tag photos, it's "ok" for it to just be a list of people, because typically the people's faces are visible, and I can order the tags left-to-right, top to bottom.
tantekhowever sometimes a photo only has parts of people (hands, arms, feet), and in that case it only really makes sense (to me at least) to person-tag if I can pin the person-tags on the parts in the photo
[shaners]That’s interesting. I never got any traction from them on better export formatting. Or putting real datetime’s on their website. But maybe you’ll have better luck with them.
lmorchardAlas, no photo, and an HWC meeting didn’t quite come together. It had been a JS meetup that we tried to repurpose, and it kind of reverted back to a JS meetup
tantekgRegorLove: your suggestion of a template for HWC meetups might be a particularly good idea for HWC Edinburgh, since their details are the same for every week, and they meet on Tuesdays (every week)
tantektbrb: do you have particular recent(-ish) HWC Edinburgh wiki page that you think is worthy of semi-automatically cloning to every remaining Tuesday in the year?
kylewm"Folks, I’ve been telling you for going on 3 years now to turn off XML-RPC and that’s what these things [pingbacks and webmentions] use." -- is that true?
tantekas Mike's tweets made it clear, users care about what *user* functionality they are getting with a plugin, they don't really care about which does what protocol and why they are separate plugins for architectural reasons
voxpelliit's too bad that WordPress doesn't have a dependency mechanism that makes it possibly to automatically have all dependencies resolved without any custom code :/
tantekright, it makes sense to have a *simple* IndieWeb plugin that gives people rel-me and IndieAuth support so they can play with that and see *something* working
miklbvoxpelli I'm getting much closer. If there was a way to have a "theme" separate from jekyll, I'd already be dogfooding it on my own site to finish fleshing out
voxpellimiklb: the possibly tricky thing is to get both layouts + includes from the same module, which one would probably need, but maybe one could symlink stuff
voxpelliGWG: one could be a superset/subset of the other perhaps, especially if one could share some libraries between them through eg. submodules or something
miklbvoxpelli hmm. With some tweaking of config, can define custom paths for includes & layouts, might be a way to use that to have a "theme" in a submodule.
gRegorLoveHeh. On Twitter advanced search I tabbed to "Written in (language)" drop-down select and typed "en" to get English. The "n" popped up the Compose New Tweet UI
snarfedacegiak: if so, congrats!...and consider removing the "Project actively developed on Github at pfefferle/wordpress-semantic-linkbacks." changelog header :P
kylewmtantek: kevinmarks: the Known plugin Kevin is using to repost is more suited for bookmarks. that's why it syndicates to twitter like a bookmark rather than a retweet
aaronpksince essentially it's providing a "fake" representation of twitter.com and facebook.com URLs, and I'm choosing to trust bridgy over the canonical URLs for those
kylewmso, are you saying that if Alice POSSEd to Twitter and Bob POSSEd to Twitter, then Bridgy would backfeed Carl's reply all the way back to Alice without having to rely on Salmentions, so this is not a use case we need to worry about?
tantekI'd ask ben_thatmustbeme how he has approached / solved this problem since I think he does the most with showing threaded replies (including backfeed from Bridgy)
kylewmben_thatmustbeme: when you're crawling a reply chain to build a recursive reply context, and you encounter a url that doesn't have microformats like twitter.com, how do you get the mf2 representation of it?
tantekanyway, in general this conversation will make the most sense with more Salmention implementing sites/people involved to point out the actual problems (if any) that they've run into due to salmentions supposedly from silo backfeeds
snarfedah. so the idea is that salmention consumers always consume the embedded mf2 on the containing page, so they don't need to also fetch the original u-url?
Loqi[kylewm] so, are you saying that if Alice POSSEd to Twitter and Bob POSSEd to Twitter, then Bridgy would backfeed Carl's reply all the way back to Alice without having to rely on Salmentions, so this is not a use case we need to worry about?...
ben_thatmustbemewonders what would happen if the u-url of the reply-to-reply was set to a bridgy URL. Depending on the consumer code, if it fetches the reply from that url, it will find the real u-url of the twitter post
kylewmben_thatmustbeme: I think that is the solution voxpelli was playing with in the OP; where bridgy would do some user agent sniffing to decide whether to forward you on to the silo url or not
ben_thatmustbemeAlice posts on her site, POSSE's to twitter, Bob on twitter, reply's (bridgy hosts reply back to alice at /1) carl reply's to bob's reply. bridgy hosts it at /2 and updates /1 to list ONLY THE URL as the reply
ben_thatmustbemeits icky, yes, but the thing is, we already have a special case for bridgy. we are trusting bridgy to us the MF2 of an external site, which you normally shouldn't do. So we are already assuming the URL we fetched the MF2 from is not the canonical URL
snarfedso ben_thatmustbeme just to confirm, your proposal is that when sites render wms they've received, they should displa* the wm source's u-url, but they should put the actual original source url in their rendered wm's u-url?
aaronpkif that was the case then i'd have to special-case bridgy anyway still, since i want people to go to twitter.com when they click the permalink of a reply
snarfedaaronpk: seems like that's still doable. you just don't put u-url on that permalink. you instead put it on a separate link to the original source, eg your 'via' link
snarfedaaronpk: i don't think so, right? if anything, it's just different logic for when u-url is the same as source url (or missing) vs when they're different
LoqiSalmentions are a protocol extension to Webmention to propagate comments and other interactions upstream by sending a webmention from a response to the original post when the response itself receives a response (comment, like, etc.) https://indiewebcamp.com/salmention
ben_thatmustbemesnarfed, the use case i would want is, I post on my site (and posse to twitter), somone on twitter replies. someone else on twitter replies to that tweet. I should get a salmention to pull an update from the first reply
snarfedben_thatmustbeme: bridgy already does that for you now, no salmentions needed. discussed a bit earlier. :P we're now looking for an example that bridgy doesn't already do.
kylewmbut it serves a pretty different purpose, and has domain-knowledge of the different properties, e.g. properties are always multi-valued or always single-valued
aaronpki am kind of curious if i could use an SSH tunnel to reverse proxy port 80 and 443 to my local 80 and 443 and get SSL set up on it and everything
bear(your nginx forwarding of port # question above) -- it can't, you have to match them up with different server blocks (or pass the port # as a special header)
aaronpknot sure what you mean. basically I had nginx listening on 80 with ngrok as the proxy_pass server, and when I went to http://example.tunnlr.xyz/ it would successfully hit ngrok but said the tunnel was not found. but if I went to http://example.tunnlr.xyz:6080/ (where ngrok was actually listening on) then it found the tunnel
tantekwonders how long until aaronparecki.com is just a cache of what's on aaronpk's laptop, and he uses bluetooth to post from all his mobile devices to his laptop without having to touch the internet directly
tantek.comedited /principles (+57) "link to positives in dfn para rather than negs, put code of conduct at top as well for additional discoverabilty" (view diff)