#indiewebcamp 2012-10-09
2012-10-09 UTC
xtof_fr joined the channel
# 
xtof_fr transclusion/inclusion could help to be sync'd on multilingualwiki. Guys : really happy to experiment first tests on http://xtof.me/wiki/POSSE#POSSE - insights welcome
# 
tantek anyone try out Julien51's #indieauth node implementation? https://github.com/superfeedr/node-relmeauth
# tantek e.g. http://aaronparecki.com/2012/281/article/1/providing-apis-for-content-driven-websites and what it linked to
# tantek interview where I explained my URL design: http://monkinetic.com/2010/05/26/tantek-celik-diso-20-brass-tacks.html (which I suppose I should write-up more on my own site at some point :) )
spinnerin, tilgovi, tantek, danbri, xtof_fr, barnabywalters, friedcell, friedcell1 and catsup joined the channel
dascher, spinnerin and barnabywalters joined the channel
# 
barnabywalters Afternoon tantek
# barnabywalters I’m using phpunit to test php-mf2, so the best format is just strings of HTML
# barnabywalters I’ve been writing minimal tests to do sanity checks on the property parsing methods, and I’ve implemented some of the value-class tests on microformats
# barnabywalters .org
# barnabywalters I’m also putting together test suites based on current real world usage of mf2 as per current implementation list
# barnabywalters Ah — I actually have no idea what I’m doing when it comes to unit tests — I only learnt how to do them last week
# barnabywalters So at the moment I’m just storing the small strings within the test functions
# barnabywalters And full html pages as .html fixtures
dascher joined the channel
# barnabywalters tantek: Not yet, I'll set that up this evening
# barnabywalters Btw tantek did you have a look over my interpretation of web actions post?
texburgher, barnabywalters, sandeepshetty and barnabywalters_ joined the channel
# barnabywalters Http://waterpigs.co.uk/articles/web-actions
# barnabywalters Please excuse half-finished redesign :)
# barnabywalters goes and reads those articles
# barnabywalters I saw the mailchimp one (and noted some irony: http://photos.waterpigs.co.uk/p/oc/tags-screenshot)
# barnabywalters I didn’t see the others, the iA article is particularly good
# barnabywalters For me, the problem I’m wanting to solve with my interpretation of web actions is that no content producer can predict the actions I might want to carry out on a piece of content better than me
# barnabywalters And with what services, etc
# 
tantek.com edited /webactions (+974) "depersonalize a bit, add Drop Social Buttons and Drop Delegated Logins sections, note Mozilla wiki page on webactions" (view diff)
# barnabywalters The other, more indieweb relevant point is the decoupling between feed readers and publishing clients
# barnabywalters tantek: What, on a technical level, are you actually proposing?
# barnabywalters My interpretation does not sacrifice silos
# barnabywalters E.g. Twitter already implements it, readability already implements it
# barnabywalters tantek: Seeing as the silos implement simple, URL based action APIs, I don't see why they should be dropped
# barnabywalters tantek: RE Indieauth login to other people's sites, I actually have a big issue with that
# barnabywalters tantek: I am not thinking purely of social silos. Readability is not a silo, it is a service. I might want hook into a native app through a custom uri scheme
# barnabywalters tantek: Or with my approach, you have a browser extension which automatically adds whatever buttons you see fit to ALL content on the web
# barnabywalters No need to trust the other site and login
# barnabywalters Solution: provide elements to put the buttons in, or put them somewhere in the browser chrome
# barnabywalters E.g. Operator
# barnabywalters tantek: Browser extension is also a more consistent UI across different sites
# barnabywalters And ensures consistent action availability
# barnabywalters Potentially
# 
singpolyma barnabywalters: why add buttons to content? add buttons to your browser
# singpolyma context-sensitive to buttons
# barnabywalters singpolyma: Exactly
# barnabywalters Yep
# singpolyma Like operator, but more useful buttons ;)
# barnabywalters singpolyma: More customisable buttons
# singpolyma tantek: all the web actions stuff. Bookmark, share, send, etc
# barnabywalters tantek: Also, consider that viewing articles as they were intended is by no means the only way content is viewed
# singpolyma web actions being in the page design is normally an anti-pattren
# singpolyma because the web browsers don't have the features
# barnabywalters Readability, safari reader — both could add in these buttons without interfering with design
# barnabywalters tantek: But it is arguable how much they actually do help, as per all those articles
# singpolyma tantek: Right, publishers need them beacuse browsers don't have them, but in general it results in giant clutter on the page for services I don't use, etc
# barnabywalters tantek: An extremely annoying part IMO.
# barnabywalters tantek: Also, not all buttons should go at the bottom of a post, e.g. Readability and so on
# barnabywalters How can arbitrary services automatically tell where the best place is for any given service CTA?
# barnabywalters tantek: It’s very r
# barnabywalters Rare I see a single one
# singpolyma I don't mind a single button as much, but that's much less useful (since it's less likely to be the thing I actually use)
# barnabywalters Do I?
# barnabywalters goes and checks
# singpolyma aaronpk: sort of like the old "subscribe on every feedreader" moved to an in-browser "subscribe" button for awhile (though Chrome never had one of those and I think the assumption of publishers now is that people have gone back to polling instead of use feed readers)
# singpolyma tantek: well, I agree than inline buttons are required for now, because of the browsers not having them
# barnabywalters So every site would have to "design" the locations of all possible buttons they want to support
# barnabywalters That’s what they do at the moment, and some people still get it wrong (I am guilty of this)
# barnabywalters tantek: Exactly! Providing buttons in the content is an assumption
# barnabywalters A dangerous one
# singpolyma tantek: yes. it's off by default in Firefox now as well. But the in-page ones aren't coming back, people just weren't subscribing (or so the browser/UX people tell me)
dascher joined the channel
# barnabywalters Why is it not an assumption that the user a) uses one of those services and b) that they will want to carry out the action on the content?
# singpolyma tantek: no, I mean the action wasn't being use. Like I said, the buttons aren't back in the content, they're just gone
# singpolyma death of feeds and all that
# barnabywalters Okay, equally by *not* providing buttons you assume people are not using services, and you get no data from that
# singpolyma Maybe. AFAICT we're asking the question "Should browsers support behaviour their users want to perform on content (ala bookmark button)?" and additionally "If browsers should support such actions, is that better than webpages trying to list every action one might want?"
# singpolyma The latter is harder to answer than the former, though I personally find such buttons in pages useless
# singpolyma tantek: how is "users want to" an assumption? The assumption is that users ever want to do things?
# singpolyma Well, if the users never want to do things then the answer to the questions would be "no"
# singpolyma thus the question is fine
# singpolyma I'm not sure what you're getting at
# barnabywalters re-reads, been a while since I read through that
# singpolyma That's sort of abstract, but sure
# singpolyma On the "(actually I'm fairly sure they're not)": FTR, I have never used a "web action" button, but I did use my "save on delicious" bookmarklets, etc, back in the day
# singpolyma but I am not a useful case study, of course :)
# singpolyma agreed
sandeepshetty joined the channel
# barnabywalters tantek: Bookmarklets are the only way of doing cross browser extensions and extensions for mobile safari, too
# barnabywalters tantek: Heh :) So RE focus enabling design, the jist of it is that users will want to do different things with the same content in different contexts
# barnabywalters Evening sandeepshetty — we've been having a long, web actions/activity dialog—relevant conversation
# barnabywalters That’s a good summary
# sandeepshetty :)
# sandeepshetty is reading the logs
# barnabywalters And another argument against having on-page buttons: there’s no way of judging what is important to the user
# barnabywalters tantek: Have you ever come across writing kit for the iPad?
# barnabywalters It has some UIs which I think are very relevant to this discussion
dascher joined the channel
# barnabywalters It’s a markdown writing app for the iPad, which is kinda an oxymoron, but it has a built in research browser with actions specific to the current task
# barnabywalters I.e. getting links/quote material for your writing
# sandeepshetty That was some conversation.. For me it is about owning my data.. so I'd rather have a bookmarklet (that I own) that picks up microformats from the current page or I feed in stuff manually and save..
# sandeepshetty on the backend I want to write handlers for those action.. for example if I like a link maybe bookmark it or tweet it..
# sandeepshetty this way I get my lifestream that I can also search instead of having it locked up at different places
dascher joined the channel
# barnabywalters is off to eat
barnabywalters_ joined the channel
# barnabywalters_ tantek: just before I go have tea, here’s a UI for php-mf2: http://waterpigs.co.uk/php-mf2/
tilgovi joined the channel
tantek joined the channel
# tantek.com edited /Why_web_sign-in (+30) "/* Why not 3rd party provider email */ clarify 3rd party specific vulnerability applicability" (view diff)
dascher joined the channel
# tantek aaronpk - interesting critique of OpenID / phishing that is worth analyzing for its applicability to RelMeAuth / Web Sign-in / IndieAuth: http://www.links.org/?p=187
# tantek and follow-up: http://www.links.org/?p=188
# singpolyma tantek: the first article is just the age-old "redirection make phishing easier" argument. Not specific to OpenID
# singpolyma oh, second article is more of the same, ok
# singpolyma So, I don't disagree per-se (though building phising sites on the fly is harder than he makes it seem)
# singpolyma Though, as I say, there's nothing OpenID-specific about this problem (RelMeAuth, for example, has the same issue)
# singpolyma I *do* think that providing phishing protection at the IdP level mitigates this, and the lack of common phishing protection at the end-website level means that if your IdP has phishing protection you're actually *safer* with a redirection-based auth
# singpolyma aaronpk: yes
# singpolyma aaronpk: well, sort of. Any redirection-based protocol (including but not limited to, OpenID) makes it so that you *expect* to see a particular page when logging in to any website, so if any website can spoof your normal login page, they've got you
# singpolyma they don't have to make you think you're going to, say, google, only that you're trying to log in to something
# singpolyma BUT they have to actually have a phishing site built for your particular IdP *and* your IdP must not have any useful phishing protection that works
# singpolyma Right
# singpolyma So, I don't disagree, but I think that if you choose a good IdP you are *better* protected than if you use passwords on random websites
# tantek whoa, someone *griefed* Twitter's registrar with a *phishing* complaint about t.co causing it to go offline last night! http://news.cnet.com/8301-1023_3-57528165-93/twitter-outage-caused-by-human-error-domain-briefly-yanked/
# singpolyma Similar to how if you use a TLS-login-form IdP with non-TLS'd websites you get more security than if you use passwords on non-TLS'd websites
# singpolyma By "good" in this case I mean "offering phishing protection" (either by having a set-by-user-in-cookie image thing, like MyopenID, or a phone-verification option, like MyOpenID, or a Yubikey option, like clavid, etc)
# singpolyma (or the vidoop image grid)
# singpolyma aaronpk: they would need to determine what sort of page/auth mechanism you use, and mimic it pretty closely, and then you would have to not notice the address bar
# singpolyma tantek: yes. MyOpenId and ClavId both have 2-factor options
# singpolyma as does google :)
# singpolyma tommorris: yeah. One of Clavid's options (which I don't use, because I use TLS certs and Yubikey) is called "mysaferlock" and is a similar tech to the vidoop image grid
# 
tommorris oh yeah, tantek, did I tell you, I have managed to rescue ~600,000 shortened URLs by emailing a friend of mine who used to run a now dead URL shortener. tomorrow, planning to pump out a CSV file with the whole lot and put them up on the Internet Archive as part of 301works. I used it for about a year on Twitter.
dascher_ joined the channel
# 
@sandeepshetty @sandeepshetty It's time to move onto http://www.converspace.org/ and http://activityweb.org #fsw #indiewebcamp
dascher joined the channel
# tantek.com edited /Principles (+459) "re-order principles a bit, expand on a few, bolden some shorthands" (view diff)
# tantek tommorris, how does OSM deal with government requests for blurring? e.g. http://news.cnet.com/8301-17938_105-57528736-1/apple-maps-outs-secret-military-site-irks-taiwan/
# tantek hey look an actual press article that mentions the NASCAR problem (although by calling it a "Nascar screen") http://www.independent.co.uk/news/uk/politics/national-virtual-id-card-scheme-set-for-launch-is-there-anything-that-could-possibly-go-wrong-8196543.html
# tantek !tell briansuda press article that mentions NASCAR in the context of identity providers: http://www.independent.co.uk/news/uk/politics/national-virtual-id-card-scheme-set-for-launch-is-there-anything-that-could-possibly-go-wrong-8196543.html - perhaps that's sufficient to create a WP article that won't get deleted?
danbri, josephboyle, lmorchard and dascher joined the channel