KartikPrabhuhmans: the current idea seems to be to use rel=following for people whose updates you are following and rel=follower for people who are following your updates. Not in wide use yet though afaik
hmansIf I have multiple <a> tags linking to the same friend (eg. if I have an avatar image, a name div and an extra div, and each of those have individual links), is there a best practice regarding the application of the rel attribute?
hmansKartikPrabhu, mostly because there's some stuff in the div that's not supposed to be linked, and some other stuff that is actually links to something else.
voxpellihmans: nice! I would probably add the rel-contact as well, for increased compatibility, + you should probably add an hcard to somehow show that these are your contacts and this connect them with your identity through some rel-me's
hmansKartikPrabhu, I started out without having #indieweb on my radar, so it's not 100% compatible yet, but the core protocol is now simply Webmention, and I'm going to attempt to switch the JSON that it's using to transmit posts across nodes to just microformats2. If that works, #pants can talk to any webmention/mf2 enabled site and vice versa.
snarfed1i think this would only work when you trigger the publish via webmention, not in the web ui, since wordpress has no way of knowing you did that. still useful though
voxpellihmans: as long as you have a rel-me from the profile page to this page (and back) then one can crawl it to find out who follows who – which is kind of crucial if you want to list the ones following you on your own site :)
snarfed1sure! my other thought…do you know if you can make a post type automatically start with some text when you start a new post, instead of just blank? if so, i could have that add the bridgy publish links
hmansI may change this at some point, but it was the easiest thing to implement given the time available. (Nothing you can see on #pants existed 2 weeks ago.)
hmansvoxpelli, in its current implementation, your site will send webmentions to your followers when you post something, leaving it up to them on how to deal with that data. I'll change this to send a webmention for their /following page so the receipient can just check for the rel=following.
GWGsnarfed1: I designed it all in my head as a possible extension of the Comments Plugin I did. But if you are thinking of a Bridgy Publish plugin, I'd split the metadata collection part of the Comments Plugin in favor of that
hmansKartikPrabhu, yeah, I know. It's a bit of a mess right now. I apologize. I'm in the middle of moving from my own protocol to #indieweb, so there's bound to be some noise.
snarfed1right now my plan is just to make pfefferle's webmention plugin add the rel-syndication links to the content when it does a new successful bridgy publish
voxpelliWebMention could actually be a good way to communicate follows and unfollows, perhaps its a bit of an abuse of the protocol, but not much more so than eg. likes and such
voxpelliKartikPrabhu: likes pointing to a post or follows that points to a profile – different kind of targets, but both activities rather than content
voxpelliyakker: a feed reader with a Twitter / Facebook like presentation would probably be ideal, but would still be a feed reader (many now uses Facebook/Twitter as their feed reader anyway)
voxpelliwhat I'm kind of is missing is the discovery part – isn't eg. mentioned at all on http://indiewebcamp.com/discovery – should one extend the blog roll page with that?
voxpelliwell, that would be one way to use it, another would be to confirm that you follow me (and have you ping me your follow list when you start or stop following me)
voxpelliAnd as before – hmans wants to have a list on his site with people following him (just like eg. Twitter has a list of people following me) – and he would need to get that data from somewhere
voxpelliIt would also make my follow data independent of any indie reader – they could easily consume my friends list from my page and subscribe me to all of the ones I'm following there
voxpelliGWG: Yes, we need to agree on the semantics used – is eg. XFN contact enough or do we need eg. rel-following as KartikPrabhu pointed at – and most of all: We need to document it
GWGKevinMarks: You have a longstanding reputation for thinking about blogging and identity online. What do you think is the best way for your own site to be discovered by others that doesn't involve a silo solution?
KevinMarksit links to homepage.mac.com (Which they killed) hosting a Quicktime reference movie (whcih they killed) to a streaming Quicktime movie (which they killed) of a steve jobs keynote (which is now offline)
gRegor`I find new people and interesting things usually through what my friends are sharing. As the web moves more towards indieweb, that will remain the same, they'll just be sharing on their own sites instead of in silos.
voxpelliBy finding who people mention (something eg. Twitter is pretty bad at nowadays, Jaiku did a good job with it though) and by having services that crawl the social graph and makes suggestions you won't have a problem with discovery on the indieweb
voxpelliif one could just discover "everyone" on the indieweb that wouldn't be a problem, but hard to do that without any central or semi-central points
voxpellitrue, more people should do like Adactio and post all of their links on their own pages rather than to eg. Pinboard :) http://adactio.com/links/ Would be a great start
GWGI remember a few years ago where someone did that deliberately as the opener for his keynote, then said, "Now that we've gotten that out of the way...let's have a real discussion."
gregorlove.comcreated /Nucleus (+279) "Created page with "{{stub}} '''<dfn>Nucleus</dfn>''' is an open source <abbr>CMS</abbr> that runs on [[PHP]] and [[MySQL]]. == Indieweb Examples == * {{gRegor}} has been running it on gregorlove...."" (view diff)
gRegor`It's hard to separate personal feelings from it, heh. Nucleus was my first open source contribution and it was a pretty stable CMS, especially for its day.
gRegor`I would love to see it continue and improved, but since I'm basically one of the only developers that's active anymore... I just can't take on the leadership of that, heh
KartikPrabhubut I still want to put a partial h-card on my homepage and link to the full one on my about-page so that anyone interested in parsing it... maybe to find my email can discover it
voxpellitantek: that section of discovery seems to be just about documenting how people are presenting data, so that one later can make an algorithm for discovering that data
GWGtantek: Okay...so I write an article on people-focused communications. How do I get it to people who might engage in a dialogue about my thoughts, as opposed to just leaving it there and hoping people find it?
tantekwhich is essentially what I did in my first blog post on the topic - I focused nearly completely on user-focus, user-needs, user-features, user-flow
gRegor`"A conference by and for our local community | July 16, 18 & 19, 2014 Hear a talk or give a talk on .... technology | design | entrepreneurship | art | culture | community"
gRegor`KartikPrabhu: So thinking about this more, I think my use-case (and yours?) is that we don't necessarily want to display our name and photo on every post. So I think it makes sense for us to just use rel-author instead of a minimal h-card, right?
tantekgRegor`: to make progress on a problem unconstrained by plumbing-specific-details, you must distill a use-case and user-flow that is 100% user presentation and features (rather than mentioning h-card or discovery or rel)
KartikPrabhui point my address book to someones homepage because that is what I know about them. But that person may have other info such as email/Twitter/G+ etcc
tantekKartikPrabhu: I *am* saying whether you store public/private info - only public is needed for a nicknames-cache, and a nicknames-cache is sufficient for those use cases
KartikPrabhu"use-case" is that I want to store my contact list somewhere so that if I change my phone or am on some friends' device, I cna just go to mydomain's contact list and do the "communication" thing
tantekexcept those icons don't launch websites - they open static duplicate copies of a "card" of information about a person that does not auto-update and goes out of date
gRegor`But I hardly ever use "People" on my phone. Just "Phone" which automatically shows your starred people and most frequently contacted, plus search.
voxpelliKartikPrabhu: Still going to be very implementation specific, don't you think? Just like every blog platform has their own way of posting blogs posts
gRegor`So my takeaway so far, if I'm following tantek correctly, is that we should focus on the simpler user-flow of manually add/edit/delete to the /nickname-cache (or whatever name one uses), then worry about the more complex flow of fetching the information by just entering a URL
voxpelliI think one would have to make some special case for the about page – like following all rel-me:s within the same domain 1 step away, otherwise you can easily end up on Twitter etc
KartikPrabhuvoxpelli: if hardly anyone is following rel-author from a post page then I doubt people will follow rel-me to about page to get person info
voxpellihardly anyone seems to consume you mean? well, all webmention endpoints out there are still pretty young so I wouldn't draw any conclusions from that really
KartikPrabhuthen address books are younger... and there is no point trying to come up with something whose couterpart in another more mature use-case is not yet adopted
KartikPrabhuI mean I can always set block elements to look inline with CSS... then that would get confusing to readers who want to hand-write fragmentions
KevinMarksso this is a fuzzy matching failure bug. the answer may be to whitespace strip text and innerText in the match algo, though that may overmatch
KevinMarkswell, it may be a made-up case, but if we start using fragmention to highlight sentences (which is reasonable, especially for annotations) it may crop up again
KartikPrabhuKevinMarks: phrase-level highlighting should be possible, but that would requite fragmention.js to play with the DOM. Add a <span fragmention> around the phrase