#tantekI actually really dislike talking about algorithms outside of use-cases / user-features
#tantekbecause that tends to go down the (dark) path of programmer-centric thinking, and loses sight of user-centric thinking
#aaronpkthat's fine, but all of those pages I linked specifically say they are algorithms
#aaronpk"How to determine authorship of a post on a page - AKA the Authorship Algorithm"
#tantekthat's probably my fault for falling into a particular writing pattern (I think I may have written those/that)
#KartikPrabhutantek, aaronpk: I think for new comers it is sort of hard to find the already-used algorithms because you sort of have to know where to look or what terms to search. It would be useful to have a list of implemented/suggested algorithms
#tantekright, it's part of default developer thinking
#tantekKartikPrabhu, the list of implemented/suggested user/site *features* is IndieMark
#aaronpkhow to: 1) determine the author of a post. 2) determine the canonical URL of a post. 3) determine the page name. 4) find post info to display a post as a comment
#KartikPrabhutantek: but unless I asked you I wouldn't know that I should look at IndieMark page
#tantekso we need to make that more discoverable then
#tantekKartikPrabhu - what is it you were looking for?
#aaronpki'm still not sold on the current iteration of indiemark. I feel like each incremental step needs to provide immediate benefit to the person doing it otherwise they have no incentive to do so other than to get points
#KartikPrabhuWhat kind of things people have done/want to do on the indieweb? This was before I stumbled upon IndieMark accidently.
#tantekaaronpk - those two are incongruous statements
#tantekeach incremental step *does* provide benefit to the person doing it!
#tantekthat's part of what each step has been based on
#KartikPrabhualso how have they done/implemeted something? Am I doing it correctly in some sense etc...
#tantekit's why I've created "Why" sections for each step
#tantekaaronpk, right at the top of the page: "Levels in general are ordered roughly by both immediate practicality to the indieweb site owner (somewhat by community interop), and somewhat by incremental difficulty. "
#tantekso yes, "immediate practicality to the indieweb site owner" is key
#tantekeverything in indiemark is provides immediate benefit to the person doing it
#aaronpkhm, there were some parts that didn't seem to reflect that
#tantekI'm totally ok with turning-off people who are anti-community as a first level focusing filter
#aaronpkmy point is that setting up indieauth, while a good thing to encourage interaction with the community, doesn't lead to any further benefit to one's own site until much later (like when using micropub, or for private posts)
#tantekaaronpk - well there's the obvious one. SEO from indiewebcamp :P
#tantek(from your User: page linking to your site)
#tantekbut I didn't think that we needed to state that explicitly :P
#tantekand what about linking to your site from the IRC logs automatically? (when you add yourself to irc-people)
#tanteksounds like I do need to add these to an FAQ
#tantekbecause honestly I thought these were all totally obvious but I guess not
#pdurbinaaronpk: what you're saying makes sense to me... about immediate benefit to the user. makes sense that owning your own domain is #1. not quite as much sense that setting up indieauth to be able to log into the wiki as #2
#pdurbinso the indiewebness of a site should be measured by the the popularity of the features that you've implemented? by the fact that you've implemented many of the same common features other indieweb site have implemented?
barnabywalters, Raymondo, CheckDavid, paulcp, scor, chloeweil_, chloeweil, eschnou, yaf, Sebastien-L, friedcell, paulcp_, glennjones and caseorganic joined the channel
#bretI just had an interesting realization the other day. tantek was talking about how when technarati indexed the html of blogs, quality of the data shot way up. I realized today that all the competent feed readers today HAVE to implement some kind of full text extraction (via readability or some other method) in order to combat low quality feeds
#barnabywaltersso that’s an example of RSS/ATOM’s hiddenness allowing them to be irrelevant to humans
#barnabywaltersit’s the same reason that parsing twitter.com’s HTML often gives you better quality results than using their API
#barnabywalterse.g. the API has no reliable way of getting all replies to a tweet, whereas they can be parsed easily out of the tweet permalink page
#cweiskeso this is an argument against APIs and for using microformats?
#Jeenabut I agree with cweiske, with a atom/rss feed you do one http request every hour or something and you get all the new stuff, with the content and with all the changes. Parsing the HTML index page + every single of the linked articles takes so much more time, just for the feed reader
#cweiskebarnabywalters, now replace microformats with html and api with microformats
#barnabywaltersJeena: but you *don’t* get all the new stuff, as you just pointed out
#Jeenaonly if the provider shortens the content, but most of them don't do that
#cweiskethe next problem will be that the html is missing microformat markup, instead of the api missing data
#barnabywalterscweiske: but if the data is within the HTML the barrier to making it parseable is tiny, and it’s much more likely that people will point out the problem because the feedback loops involved with using the data are smaller
#cweiskethe barrier to fixing the api is at the same level as fixing the html
#cweiskeand why should people be more likely to point out issues with microformat markup than issues with the API?
#barnabywalterscweiske: the information is more likely to be in the HTML in the first place because it needs to be there because humans want to see it there (cases in point: tweet replies, feed pagination links)
#bretiirc where you have to get the whole document in order to see if something was updated
#barnabywaltersyeah, in theory that is handled by something like PuSH but it hasn’t been a high enough priority (not enough visible value) for people to implement yet
#bretif we were to iterate on the rss/atom feed model, could we do push+html parsing only? fall back to rss if the subscription fails
#barnabywaltersthe situation could also be improved by people adding dt-updated to feed index pages
#barnabywaltersof course, in the case of notes, and often for articles, the whole content *is* shown in the index
#breta possible issue, rss/atom have this too, is that some publishers might intentionally botch the uF so that parsing only brings in some of the content
#barnabywaltersreaders should work TOWARD this goal instead of against it
#breti'm assuming an additional layer of agressive parsing would be needed in a practical reader application
#bretbarnabywalters: news blur will actually show you the page inside the reader. I guess I see what you are saying
#bretbut this ignores the needs of offline readers on small devices
#bretand ignores the fact that publishers that oh so need your hit on a mobile connection often send you 10Mb of javascript analytics.. nothanks
snarfed and LauraJ joined the channel
#barnabywaltersI’d say that’s a separate, but very valid, problem
#bretyeah, i agree, im getting kinda mixed up here
#JeenaI wrote a feedreader for Firefox OS. How would PuSH work with that?
#barnabywaltersthat is, creating formats and readers which are worth enough to publishers to implement support for is a separate issue to publishers forcing crappy experiences on their readers
#breti think any kind of general purpose reader that 'just works' is going to be a battle no matter what
#barnabywaltersJeena: PuSH is basically server-server, so you’d have to do something like set up a server which the app opens a web socket connection to, to which it forwards notifications received from a hub
#tantekboth Atom and RSS are just very crappy inefficient subsets of HTML
#tantekAtom did introduce some better defined vocabulary for feed/entry-like things. So that was good at least.
#barnabywalterstantek: I’d say they’re closer to being an unnecessary, inefficient wrapper for chunks of HTML rather than subsets
#JeenaI think we all agree that HTML + microformats is much easier then different XML formats. What we don't quite seem to agree is the number of HTTP requests per hour to get all updates.
#barnabywaltersaw geez I was just making a comic about the shadow DOM and now you’ve got me in full-on wiki editing mode ;)
#Jeenauhm not really, if you have 1000 people subscribing to your feed it is a http request every say 8 seconds with atom for everyone to check for updates. If you have a index site like mine with about 100 links and 1000 subscribers it is 27 HTTP requests per second just so everybody gets their feed reader to know that nothing has changed. This is kind of a huge change.
#Jeenadt-updated on the index page would probably be able to cut through that
#tantekbarnabywalters as if we needed *another* comic about the shadow DOM?!? or wait, I mean meme.
#barnabywalterstantek: well, it’s really about naming things. the shadow DOM is the punchline
#tantekJeena - there's plenty of data in the page to be more efficient than that
#Jeenawas the h-feed reference intended to support the theoretical call because of the many HTTP requests? It looks to me like it mostly helps with adding information about the feed author and some metadata about the feed like name, url and photo
#Jeenahehe barnabywalters the video was uhm hehe for non tech-people, but I think I start to understand. to implement PuSH I either can write my own Hub and handle all the stuff myself like subscription and push to everyone, or I could use http://pubsubhubbub.superfeedr.com/ as my Hub where I push only once and it takes care of the subscription and pushing to people?
#tantekJeena - or you can still use the Google test hub too
#aaronpkoh I remember why I didn't add h-feed, cause I was only showing 3 notes on my home page
#barnabywaltersthe only use cases of h-feed I’ve found so far have been: explicit naming of the feed, page-level author discovery and multiple feeds on a page
caseorganic joined the channel
#aaronpknow it makes sense that I show 10 notes in the primary content area of my hom epage
#tommorrisaaronpk: no, got a programmatic translation of XML to JSON.
#tommorrisbut the nice thing about XSLT is you can set up multiple rules across multiple files
#barnabywaltersbriansuda created a huge XSLT to turn hcard into VCARD
#aaronpkbarnabywalters: do you not have h-feed yet either?
#barnabywaltersaaronpk: I do have h-feed, mainly for author discovery
#barnabywaltersbut IMO that’s a separate thing to actually consuming feeds, i.e. at the moment there’s no reason for readers to demand h-feed markup over just having h-entries on the page
#tantekbarnabywalters, bret, Jeena please read http://indiewebcamp.com/API so you can refer cweiske to all those problems instead of having to restatement them next time he brings up APIs
#JeenaThere is one thing about APIs though, and why they are so popular. If you can say that your startup offers an (REST/JSON) API then many people start listening, it is kind of a prestige thing. Hm I wonder, is there a microformats parser for the iPhone?
#tommorristhe number of times I end up saying at work “well, let’s try implementing both approaches and then test them to see which works” is interesting. and the number of times people think that’s slightly mad rather than a pretty basic part of engineering is worrying.
#tanteklet me see if it works if I quote tommorris
#tantektommorris: as for cargo cult software engineering: the fact that people would follow sites like Twitter or Facebook in their engineering decisions
#tantektommorris: people stopped using Rails because Twitter had switched away from it
#tantektommorris: or Google had switched away from using RDBMSes, so they just decided that RDBMSes were useless and switched to NoSQL or map-reduce
#tantektommorris: even though relational databases are pretty useful and have been since the 60s.
#tantektommorris: the number of times I end up saying at work “well, let’s try implementing both approaches and then test them to see which works” is interesting. and the number of times people think that’s slightly mad rather than a pretty basic part of engineering is worrying.
#bretdid microformats not exist when atom/rss was getting started? how come two of basically the same thing came out of those efforts?
#willnorrisbret: but they have fundamentally different approaches anyway... an alternate machine readable format (rss/atom) vs embedding machine readable tags into the existing format (microformats). So even if the timeline were different, I would suspect both would have emerged... they just come from different schools of thought (my thoughts anyway)
#tommorrisplumbing in my geonamer code seems to be working swimmingly. will write it up shortly.
#tantekwillnorris - well, I can tell you that from my perspective, the drive to seriously pursue microformats came from *trying* to make the XML-stack work at W3C in the early 2000s, realizing that the whole thing was a massive enterprise house of cards, and said screw it, how simple can we make this
#tantekand as part of that I realized all the side URLs / side files problem
#brettantek: by grouping, do you mean enhanced recent changes? I dont see a grouping option
#willnorrisah, okay... so it really was a response to the complexity of having different formats? In that case, perhaps it was a necessary predecessor to prove that the model doesn't work as well
#tantekbret: "shocked by how many papers were pot-boilers that disguised trivial ideas with inflated language. These were papers that had absolutely no discernible academic value other than to pad a resume, and collect but a smattering of citations, mostly from similar papers"
#bretugghhh, the text in the quitelevel in that email thread are unreadable
#tommorristantek: what you say about the geo tags on my site?
#tommorrisI’m fairly sure that was allowed at one point.
#caseorganic.comedited /2014/SF () "(-3475) /* IndieWebCampSF 2014 */ Moved Guest List to new page, added categories: 2014 and Events. Added info about flights and hotel, additional metadata." (view diff)
#KartikPrabhuand articles like this seem to suggest that Web Components are a better way of doing XML-type things. Which is why I do not understand the former...
#KartikPrabhuall of this seems to be complicating HTML without any actual benefit
#tommorris"Screenshot of Wikipedia’s custom rich-text editor, with assistive tools for Wiki-specific markup"
#tommorrisThat’s not a screenshot of Wikipedia’s rich text editor. That’s a screenshot of Wikipedia’s markup editor with a bunch of JS. The one that Wikimedia are in the process of replacing.
#tommorrisPrecisely because having editors have to remember thousands of template names, arguments and then bolt-on hacks to support them wasn’t really working out
#tommorrisThere’s a legitimate point to that ALA article: there’s a mismatch between our authorship tools and the sort of things we can do on the web.
KevinMarks joined the channel
#KartikPrabhuBut on some level it seems to me that people are saying "I want to write a book in Japanese but I don't want to learn Japanese!"
#KartikPrabhuOur writing tools may need improvement, but certainly authours would have to learn something
#tommorrisThe web exists to serve humans. A reporter or blogger shouldn’t need to know what the box model is.
#KartikPrabhuof course not the box model. But they might have to learn some markup/template codes. Else they can just write in plain text and put it in a CMS.
#tommorrisWell, the problem is that CMSes usually suck because of no-dogfooding
#tommorrisThey are built and sold on feature lists rather than feature needs.
#tommorrisI now have a 13” RetinaBook - the thing has the same resolution as an HDTV! - but I have to grab a smartphone to do certain things.
caseorganic joined the channel
#tommorrisMy online banking is unusable. But the app on my phone isn’t. It does 80% of what I want to do on my online banking but without requiring me to sacrifice burnt offerings to the security gods. So I use the phone version and not the desktop version.
#tantektommorris, KartikPrabhu, I heard that Markdown solves all these problems. ;)
#tommorrisMarkdown solves many of the problems. The smartest thing Markdown did was allow you to still use HTML
#tantektommorris - install a simulator for your smartphone on your laptop? ;)
#KartikPrabhuyeah. I quite like Markdown too. but you still need to write raw html to do somethings
#Loqibarnabywalters: tantek left you a message 1 hour, 14 minutes ago: Indiewebifyme feature request: h-feed parsing/validating similar to current h-entry parsing/validating
#KevinMarksbut they still require an email to login with - or are they trying to get everyone to switch to phone numbers so the whatsapp merger makes sense?
#tantekWhatsApp scaled *before* selling. That's the key.
#tommorrisFriend of mine just got Adobe spam inviting him to "increase Wikipedia's return on marketing investment”. When was the last time you saw an ad for Wikipedia? Wikipedia’s marketing budget is $0. They get a pretty good ROI. ;)
#tommorristantek: yeah, it’s on my very slow to process list of next actions
#KartikPrabhuyeah .mobi is dumb. With all this responsive thing going on how the hell did .mobi pass?
caseorganic and caseorga_ joined the channel
#tommorrisKartikPrabhu: because the people who make the decisions are more interested in buying new designer suits and private jets than they are technical merit.
#KartikPrabhuso someone can pay ICANN to make more TLDs?
#tommorrisyou make a proposal to ICANN for a new gTLD.
#tommorristhere’s a proposal for .eco - for charities and organisations doing ecological stuff. and the money would go to some kind of environmental fund.
#tommorris(if .gay comes into existence, I’m going to register “not.gay” and point it to whichever right-wing fundamentalist just got caught in an airport bathroom with a wide stance.)
#tantekwhen moving sections like that to their own pages, leave in a placeholder so previous fragment links still work! e.g. any http://indiewebcamp.com/2014/SF#Guest_List that may have been shared with people