julien51So, yes, when I read your proposal, I found it very similar to what we had designed for PuSH 0.4. The bad news is that we're struggling to get adoption
waterpigs.co.ukcreated /subtome (+251) "Created page with "{{stub}} SubToMe (Subscribe To Me) is a button content publishers can put on their sites which enables users to subscribe to their feed in whatever reader they choose — sort o..."" (view diff)
waterpigs.co.ukedited /PubSubHubbub (+456) "/* Testing */ expanded testing section, added link to notifixlite and mention some XMPP providers" (view diff)
julien51cweiske : we heard a lot of good things about it too. We would love for it to work with http://subtome.com, do you know if there is a page where you can add new feed url that would accept a query string to pre-fill the feed url?
barnabywalterscweiske: sure, URLs can handle custom schemes — but no browsers I know of allow you to pick from a list of eligible apps when you go to one of those scheme URLs (as subtome or webactions toolbelt do)
julien51we want to introduce an 'API' to avoid one redirect so that the SubToMe button itself calls (jsonp!) the subscribing application directly to perform the subscription
tantek!tell julien51 welcome back! sorry to miss you in the channel. webactions better for publishers since no script is needed. also, have documented "webintents" problems since 2011, their death shouldn't have been a surprise.
tommorristantek: I'd question whether Wikipedia has a liberal TOS. it's got an honest one which basically boils down to "we give a shit about your privacy and have strict rules about what we do with your personal information, and if you are dick we might ban you"
tommorrisa while back, for work, I had to write a load of custom script crap to transfer from Redmine to Bitbucket. code-repository stuff isn't as portable as it ought to be.
tantek.comuploaded /File:twitter-home-in-stream-reply-context-thread.png "screenshot of Twitter home page stream fragment showing a tweet followed by an @-reply tweet that by default shows a reply-context of the person being replied to with smaller icon, and original text being replied to, same font, just slightly smaller text size for the original tweet, and above that, the precedeing thread, "View 3 more tweets", vertical ellipses up to the original tweet."
tantek.comedited /reply-context (+237) "/* Twitter home page */ twitter home page reply-context of a reply-thread, ellipsed/collapsed display" (view diff)
tantek.comuploaded /File:twitter-home-in-stream-reply-context-tweets.png "screenshot of Twitter home page stream fragment showing an @-reply tweet that by default shows a reply-context of the person being replied to with smaller icon, and original text being replied to, same font, just slightly smaller text size for the original tweet, and above that, the immediately previous tweet that was the start of the thread."
tantek.comedited /reply-context (+180) "show reply context with @-reply, what it's in response to, and the start tweet that that is in response to" (view diff)
waterpigs.co.ukedited /reply-context (+382) "/* Twitter home page */ added comment on the constant flipping of the directionality of time on twitter.com" (view diff)
Loqijulien51: tantek left you a message 1 hour, 30 minutes ago: welcome back! sorry to miss you in the channel. webactions better for publishers since no script is needed. also, have documented "webintents" problems since 2011, their death shouldn't have been a surprise.
julien51Absolutely no script is required by any publisher to use SubToMe. Also, SubToMe is browser agnostic (it's web based!), and has already been implemented by many feed readers. But I guess you could also *wait* for another couple years before something happens :) I decided to move forward to make something that works without compromising the future.
Loqisandeepshetty: barnabywalters left you a message 8 hours, 57 minutes ago: RE webmention on WPC — oops, looks like I forgot to implement that yet! coming soon :)
Loqisandeepshetty meant to say: julien51: while I'm caching up and since you are here would love see some comments from you here: http://indiewebcamp.com/push-vs-pull
julien51when you say pull is simple for publishers, it's actually not accurate. It "appears" simple, and yet that's pretty hard to handle. For example, at some point, almost 40% of Tumblr's traffic was consumed by search engine. That's not *easy* to handle
grawitythis is partly out of curiosity (I'm interested in the reasons behind decisions), and partly because I want to log in to the wiki with such an address :)
grawitytantek: my guess is lazy admins (they go through the backlog only several times a year) plus the requirement to have at least two existing nameservers
tantekgrawity - when you say "mine" - it's not clear how to trust the benevolent nature of eu.org moving forward. at least with domain name rental services you there's a well-understood profit motive / customer service to not screw people over.
julien51sandeepshetty: well you explicitly wrote that "(PuSH was designed by Google engineers to solve their problems with aggregation and not solve the problems of the subscriber - which is a central actor on the indieweb)."
grawitytantek: at first look it's a stable one (it was itself created as a protest against registrars trying to screw people over, and the owner later founded gandi.net)
julien51not really, in both cases you can use libraries to do that on your behalf. It's just that we all forgot how it's done for https and valid certs since we used these libraries, but my point is that you can't say "security is complex in push" and just plainly ignore that it is complex on "pull" too.
sandeepshettynot really, a publisher *has* to confirm subscriptions from subscribers to avoid the attack vector. you don't have to do that with a pull.
sandeepshettyand a subscriber just has to access the fee url over https with pull while with push there has to be some mechanism to share a secret and then use that to sign the payload and then the subscriber has to check the signature
julien51"just access the free url over https" is actually quite complex but hidden from you because you used libraries to do that. There are libraries who do the same for you with PubSubHubbub.
breti have an idea for subdomians... if i share a domaing of my last name with my family, having a subdomain of my first name would be a valid use case
julien51sandeepshetty: please do also come up with data when you make your statements like "simple publisher.", or "* better for frequent updates because they are grouped together." because the data have shows the exact opposite, I'd love to learn from you!
tantekjulien51 - the default in the past 5 years of "social/federated web" crap is "email list = community". everyone "says it" by doing it - creating a new google group for whatever.
julien51tantek: I don't think you can measure the usefulness of something with the number of comments on a mailing list… I'm worth you and I'm tired of people discussing things in mailing lists (or IRC channels for that matter) I look at implementation. PubSubHubbub is present for a huge number of feeds. That's a fact which is far more interesting to me than X people discuss it on an obscure IRC channel :)
julien51tantek I don't agree with that. Sometimes things are good. If nobody moves them forward it's probably because nobody knows (yet!) where to push them forward
tantekyou're welcome to bring the implementation-centric discussions around PubSubHubbub here to IndieWebCamp as a lot of us *are* supporting it and think it's a good thing - we have pages on it :)
julien51I love the indieweb community (and you know it), but I don't think it's anything in terms of success/adoption compared to what we achieved with PubSubHubbub.
tantekjulien51 - that may be (nobody knows (yet!) where to push them forward) - but then that means it's vulnerable to a bigco forking and pushing it forward proprietarily
julien51tantek: this is exactly why I cam here… but i have't heard anyone tell me how we could make PuSH better. I heard/read uninformed statements like "simple publisher." or " implies your site cannot be static.".
julien51ha, that, yes, I read it and it's indeed amazing (maybe the only reason why I'll rewrite ouvre-boite.com), but it also reminds me of SWAT0 a couple years back where Bounds and Prodroumou and Googler demoed the same thing with Identica, cliqset and google buzz.
Loqijulien51 meant to say: tantek: we'd need to ask evan, but I'm sure there was at least another couple impls that 'worked' but were not used. Was it rstatus?
julien51we also need to stop trying to be nobel prizes. The open web community is full of people who think that whatever everyone else did is crap and that *they* alone will make something better and change the world
julien51Honestly, there is not a week where somebody comes to me/superfeedr and says, we're doing this new federated social web thing and it's better than what you did, please support it and tell the hubs you host (Tumblr…) to do it.
tantekjulien51 - that's hilarious: "there is not a week where somebody comes to me/superfeedr and says, we're doing this new federated social web thing and it's better"
bretI was just happy to get rid of all the proprietary logos from my IM account list and use only xmpp for a good 5 years. i still had to maintaint 4 or 5 xmpp accounts, which was slightly better at the time, seemingly.
julien51as for POSSE I'm sure you were still logged in on Fcebook and twitter all the time. You had a point to make, but most people would not spend more than a day posting on their own site with POSSE if they figure out that they can just do it directly on Twitter/FB
tantekjulien51 what do you mean by "Dogfooding alone is self inflicted ban-hell" ? that doesn't make any sense nor does it match *anyone's* experience here.
julien51tantek: I mean that posting to your own site that no one reads quickly becomes really hard to use. People post to twitter or Facebook because other people follow them there.
sandeepshetty!tell julien51: I've updated the wiki page with my comments. BTW, except for the one instance you mentioned nothing about the push section was about PuSH.
bretIts super easy to add PuSH to a static site, but there are some technical issues with the way github handles their post commit hooks that cause some misalignments
bret!tell julien51: I would love to ask you a few questions about using superfeedr as a PuSH hub for github pages + post commit hooks next time you are on
tantekin general, any project that claims to be federated/social/indie web-like is probably something we should critically document on indiewebcamp.com since Wikipedia sucks so bad at that topic area in particular.