2013-11-07 UTC
KevinMarks joined the channel
# 00:17 KevinMarks The hangouts on nexus 5 shows a phone icon by people in sms, not sure if it uses voice or native phone though
# 00:18 aaronpk the ios app has a full dialer, not sure about android
# 00:21 benwerd my guess would be no, because the android kitkat dialer is part of the google ecosystem anyway
ozten, skinny and npdoty joined the channel
earplugs, benwerd_, benprew, paulcp, npdoty, bnvk and caseorganic joined the channel
# 01:36 XgF aaronpk: The Android app doesn't really need a dialer because Android has great inter-app integration capabilities
# 01:38 aaronpk but how do you make a call from your GV number in android then?
# 01:39 aaronpk I mean I would assume it would work better/easier in android since there are fewer restrictions
# 01:40 XgF Not sure. Google Voice doesn't exist here in rightpondia
# 01:42 XgF I'm not sure how Google Voice integrates (if it does), but Android has native SIP calling support
# 02:12 benwerd_ Confirming that Google Voice does handle the outgoing dialing, if you want it to
# 02:12 benwerd_ I have it set to always ask me because I like control
# 02:12 benwerd_ but I could just say, ok, whenever I dial a number, do it with Voice
brianloveswords, tantek, bnvk and earplugs joined the channel
# 02:35 bret tantek: i'm not sure if you the discussion above or the news/ the hangouts app from google gives you a free incomning outgoing voip phone number... seems like that might interest you
# 02:37 Loqi bret meant to say: tantek: i'm not sure if you sawthe discussion above or the news/ the hangouts app from google gives you sawa free incomning outgoing voip phone number... seems like that might interest you
# 02:39 tantek tried signing up for Google Voice a while ago and gave up - too many steps - didn't finish
# 02:40 bret its basically the phone you get in gmail, but now they have an iOS app
skinny, earplugs, crossdiver and bnvk joined the channel
# 03:45 Loqi [@t] @kevinmarks @nrrrdcore it is! Thanks for that breath of fresh air at #Roadmap2013
crossdiver, pfenwick, bnvk, snarfed, hober, tantek, npdoty and skinny joined the channel
# 06:30 KevinMarks On android you can have Google voice take over completely, or just if there's no carrier
cweiske, pfenwick1 and andreypopp joined the channel
# 06:47 neuro` cweiske: I guess it's OK as you can export the content and just have your domain point to your gh page. It's just like having a shared hosting.
# 06:47 tantek cweiske - github as post backing store is no different than S3
# 06:47 tantek as long as the content is wrapped in URLs with your domain, the rest of just a backend hosting implementation detail (swappable)
# 06:49 tommorris this is why I use code.tommorris.org - it points to bitbucket, but at some point I can rewrite the URLs and host my own
crossdiver joined the channel
# 06:52 neuro` KevinMarks: I know there's a tech conference in the US when I wake up and my timeline is full of you.
# 06:52 cweiske does github and bitbucket support custom domain names?
# 06:52 Loqi cweiske meant to say: do github and bitbucket support custom domain names?
# 06:52 aaronpk github does for html hosting, not for your profile or code
# 06:59 KevinMarks I'm mainly amused by how 100 lines of go and the github api basically replaces Ghost
# 06:59 cweiske has anyone thoughted about having a status API for webmention?
# 07:00 cweiske it's advertised to do asynchronous processing on webmention requests, so the sender never knows what actually happened to his request
# 07:01 aaronpk 1) it's not always desirable to report why something was rejected. 2) you could always check the URL to see if your comment was posted
# 07:01 cweiske webmention is not for comments only, at least I see it that way
# 07:01 aaronpk 3) even if you wanted to report the status of why a webmention was rejected, there is very little you can do with that programmatically
# 07:02 cweiske aaronpk, the webmention sending side could poll after an hour
# 07:02 aaronpk my question would be what is the purpose of the sender finding out the status?
# 07:03 cweiske finding bugs or errors. when my mention gets rejected because it's seen as spam, I won't notice at all. I can't investigate why it's that way
# 07:03 cweiske when sending an email that gets flagged as spam, I get an immediate answer by the mail server
# 07:04 aaronpk I can see why it would be useful for debugging. but I don't want to tell someone their mention was marked as spam.
# 07:04 cweiske aaronpk, it could also be a reason "you're not within 2 levels of my friend's friends"
# 07:04 aaronpk my email server usually silently eats spam. if it replies that a message is spam that's a new attack vector my server can be used to send spam messages with
# 07:04 cweiske the mail server rejects mails with a clear status
# 07:05 aaronpk it'll say it's delivered, it just doesn't deliver it
# 07:05 aaronpk also my favorite anti-spam email technique is greylisting. the first time you send an email to me I reply "sorry i'm busy, come back later", and 98% of spam doesn't try again
# 07:06 aaronpk anyway, for webmention, until there's a very clear use case for *why* you want more information back and *what you will do with that information* it doesn't make sense to add anything to the protocol
# 07:07 cweiske I want to be able to programmatically check if my webmention was accepted
# 07:09 cweiske so that the software that is acting on behalf of me can tell me "hey look something is wrong"
LauraJ joined the channel
# 07:15 aaronpk 2 ways off the top of my head to accomplish this:
# 07:16 cweiske a third parameter, kind of "notifywhenready" in the original webmention request
# 07:16 aaronpk a) in addition to source= and target= you also send callback=http://example.com/webmention-status , my server can then post a reply to that URL with the status after it's done processing. (this is probably makes my server vulnerable to being used as an attack vector though)
# 07:17 aaronpk b) when I crawl your page looking for your link back to my site, I also look for rel=webmention-status (or just use the rel=webmention URL) and post a message to that URL
pfenwick joined the channel
# 07:19 tantek aaronpk re: vulnerable to vector - I believe there are webhook API callback design patterns to avoid that
# 07:19 aaronpk most of them involve pre-registration of callback URLs
# 07:20 aaronpk as in, I can't think of any service that lets you set the webhook URL on the fly
# 07:23 aaronpk but yeah, if I just post a reply to your webmention URL after I finish processing, you'd be able to do whatever with that
# 07:25 aaronpk the post request would have to have some additional property to indicate as such, and would also include the status property
bnvk, poppy and Jihaisse joined the channel
pfenwick joined the channel
# 07:44 aaronpk cweiske: anyway, I'd be up for experimenting with this if you want to receive the webhook notifications and do something with them
npdoty joined the channel
# 08:20 cweiske I fallback to pingback if webmention isn't supported
LauraJ joined the channel
# 08:28 cweiske I actually have a standalone server that's not tied to my blog
# 08:28 Loqi cweiske meant to say: I actually have a standalone server software that's not tied to my blog
# 08:29 cweiske sending works by parsing my atom feed, extracting links from there
bnvk, crossdiver and skinny joined the channel
ozten, friedcell, pfefferle, bnvk and crossdiver joined the channel
abrereton and bnvk joined the channel
barnabywalters, adactio, skinny, andreypopp and crossdiver joined the channel
# 11:34 tantek.com edited /checkin (+369) "/* People actively publishing check-in data */ subsections for implementations, add benwerd, tools, date placeholders" (
view diff )
bnvk joined the channel
bnvk, benprew_, Jihaisse1, michel_v_, barnabywalters_, RexWhitefish, spinnerin_ and Jeena_ joined the channel
# 12:22 hadleybeeman Goodness, tantek, you're up late!
dietrich, otterdam, cweiske, crossdiver, pfefferle, brianloveswords, bnvk, snarfed, andreypopp, josephboyle, _6a68, tilgovi, ryana, paulcp, indiewebcamp and ericlambrecht joined the channel
benprew, josephboyle, jschweinsberg, carlo_au, tantek, bnvk and texburgher joined the channel
andreypopp and benwerd joined the channel
# 19:07 barnabywalters snarfed: loving facebook-activitystreams HTML+mf2 output! I just piped it through intertubes and it worked perfectly
# 19:07 barnabywalters Iām now browsing through a feed of content from facebook via fb-unofficial, benwerdās stream from werd.io and brian sudaās twitter stream
# 19:08 snarfed i know how you feel. :P i've read my fb and twitter streams in my feed reader for years, so much happier that way
cweiske joined the channel
# 19:18 aaronpk then I think you have to use the weird value class pattern thing
# 19:20 aaronpk well if you update the markup to match the value-class-pattern it should work
LauraJ joined the channel
_6a68 joined the channel
# 19:23 tantek cweiske no you cannot merge class="dt-published value-title"
# 19:23 tantek and abbr is good only for dates - not for datetimes
# 19:24 cweiske <abbr class="foo" title="YYYY-MM-DDTHH:MM:SS+ZZ:ZZ">Date Time</abbr>
# 19:29 tommorris cweiske: do jump into html5. the time and date elements are worth the it. ;)
# 19:29 tommorris and there's a defined XML parsing model for HTML5 if you want to write stuff that smells like xml
# 19:31 tommorris on a completely unrelated note, unlimited 3G/4G data plans ftw.
# 19:34 tantek cweiske - all my Falcon data files and generated post permalinks are well formed XML using HTML5
# 19:35 tantek unless you're appreciating it for some historical reason
# 19:38 tommorris also, HTML4/XHTML1 validation - there's no point. the browsers don't care. you go to lots of effort to validate in order to deny yourself the right to use the new stuff like data, date, time, section, footer etc.
# 19:39 tantek barnabywalters - alex has never understood "semantic HTML" so his frustration is his fault for not understanding, not the technologies.
# 19:40 tantek deeply flawed statements in that post, like "PowerPoint slides or PDF forms can transmit nearly all the same semantics as HTML from the perspective of users " - totally false - doesn't work across devices, open source implementations etc.
_6a68 joined the channel
# 19:41 tantek he also doesn't get that semantics is about content, not UI
# 19:41 tantek he comes from a very UI-centric JS-framework background
# 19:42 tantek sigh, it's another tl;dr long-winded article from Alex that misses so many points it's not worth responding to. It's not clear it makes any actionable points to anyone.
# 19:42 tantek he has a few gems in there (problems with article vs. section) but they're buried in mountains of misconceptions. oh well.
# 19:43 tantek barnabywalters - I wouldn't say "lots of good thoughts about HTML"
# 19:43 tantek if you want to iterate/extract/filter what you think those good thoughts are into minimal summaries, perhaps we can assess them - though "semantic HTML" is more of a subject for #microformats
benprew joined the channel
# 20:06 tommorris "Thereās some weird strain of the ālinked dataāĀ and āsemantic HTMLāĀ zeitgeist that forgets that the reason people build things on the web is almost always because they want some other human to find and have their lives enriched by that thing."
# 20:07 tommorris I'm not sure which life-enriching experience I value more: the times when I copy-and-paste stuff out of PDF documents and have to virtually retype it to have it format as valid text/plain
# 20:08 tommorris or the even more life-enriching experience of PowerPoint version incompatiblities
barnabywalters joined the channel
# 20:38 KevinMarks The one that bugs me now is that Google presentations, which started out from s5 and were thus clear html, now export as SVG, with text declared as paths. Whut?
KevinMarks_, josephboyle and barnabywalters joined the channel
npdoty joined the channel
LauraJ, abrereton and bnvk joined the channel
# 22:14 sparverius KevinMarks_: short sale twitter and send the report in scrawled on top of the receipt
LauraJ, lukebrooker and benwerd joined the channel
# 22:57 _6a68 I want to leave a message for tantek, but I don't recall the syntax
tantek joined the channel
# 22:58 _6a68 tantek: i was just going to !tell Loqi to tell you, I chatted a bit with rob about indieweb + moz
# 22:58 _6a68 yeah! he's still really high-level, pitching the idea internally
# 22:59 _6a68 lol, I was quite surprised to see the new director of product had the IRC nick b0bg0d on moz IRC
# 22:59 tantek meanwhile us builders can just build stuff and make it happen bottom-up
# 22:59 _6a68 the first day or two he was around, I figured he was some random dude from the tubes
# 22:59 _6a68 tantek: yeah! totally. i'm going to poke at indie web stuff in the off hours, and hopefully it turns into a day job sometime next year
# 23:01 _6a68 so the thing i've found difficult, thus far, is that there are microformats and concepts like POSSE, but not many toolkits/building blocks to stitch things together more quickly
caseorganic joined the channel
# 23:01 _6a68 i know there are a couple biggish/monolithish projects, but composable parts are so helpful in hacking stuff
# 23:02 _6a68 glad to hear my perception lines up with everybody else's
# 23:02 aaronpk there are a few more components now too, like the webmention client in ruby and php
# 23:02 snarfed heh, barnabywalters took the words out of my mouth
# 23:03 _6a68 well, speaking selfishly to my personal use case, I've wanted to "POSSE service X" where X might be twitter or github or tumblr
# 23:03 _6a68 and it didn't seem there was a smooth path to do that
# 23:03 tantek _6a68: missed you at IndieWebCamp Hollywood this past weekend (you're based in LA right?)
# 23:03 _6a68 tantek: yeah! i'm near venice. we're moving next week, everything's in boxes. was bummed to miss it
# 23:04 snarfed integrating indieweb stuff w/silos is near and dear to my heart
# 23:04 _6a68 moving in-town, but still ,it's been 5 moves in 5 years, you'd think I would be used to it
# 23:04 tantek the closest we have to POSSE components is the POSSE-per-service extension network that benwerd has built into idno
# 23:04 tantek it's really easy for him to add new POSSE destinations because of the POSSE-service-provider plugin architecture of idno
# 23:04 benwerd As it happens, I'm working some more on that this evening.
# 23:04 tantek see if you can re-use his POSSE components and wrap your own code around them
# 23:05 snarfed agreed posseing original posts is usually pretty CMS-dependent
# 23:05 tantek benwerd - you can start with putting permashortlinks/citations in the POSSE copies
# 23:05 tantek that's the biggest hole right now in idno POSSE support
# 23:05 _6a68 yeah, it's funny, i started thinking about this stuff a couple years ago, when i looked at the lockerproject code
# 23:05 _6a68 "why aren't all these silos hackable? they are built by hackers"
# 23:05 tantek none of are able to do original-post-discovery on idno POSSE posts to FB/Twitter/Foursquare/Flickr etc.
# 23:06 benwerd I'll also be adding some user-side POSSE UX stuff, and then looking again at URLs.
# 23:06 _6a68 benwerd: so here's a question, idno is PHP, right?
# 23:06 _6a68 suppose you had hackers who favored some other random language
# 23:06 _6a68 is it possible to deploy little pieces of idno as services?
# 23:07 _6a68 so that you could have a mix of languages speaking over http?
# 23:07 benwerd Huh. It's funny, I had another conversation about something similar earlier today.
# 23:07 _6a68 I suppose you could think of it as a bunch of queues, really, since you're pushing and pulling stuff from spot to spot
# 23:08 _6a68 heh, i suppose you could stretch the analogy
# 23:08 aaronpk take the work i've been doing on micropub and make an API that conforms to that and can posse to all these other services!
# 23:08 snarfed my main indieweb project right now is pretty similar: webmentions into and out of the silos, as a service
# 23:08 benwerd I think, and this speaks to a really good point barnabywalters made the other day about complete protocols vs small components of functionality, it's something I would *like* to do - but I need to unpick some of idno's platform-ness
# 23:08 snarfed similar to webmention.io, but talks each api on the back end
# 23:08 _6a68 and think of your domain as a flow of messages, and all the individual silo syndicators could just filter on the messages, transform them, and ship them off
# 23:08 _6a68 i've hacked PHP in the past, i can try to help untangle idno if that's useful
# 23:09 benwerd what it's certainly about 3mm away from is being usable as a back-end API for a completely different front end
# 23:09 barnabywalters another way to make this functionality small would be a service which subscribed to your feed using PuSH
# 23:09 _6a68 oh yeah, PuSH would be a nice way to coordinate the queues/services/buckets
# 23:09 barnabywalters then whenever the feed is updated, it POSSEs the latest content to a bunch of silos
# 23:10 _6a68 yep. i was working on logging infrastructure not too long ago, i can't help but think of this as filters/transforms on messages :-P
# 23:10 snarfed agreed, PuSH is awesome for all the parts on our end
# 23:10 snarfed we still need to poll silos for data that originates inside them, since they don't implement PuSH
# 23:11 snarfed aaronpk: i assume you're not talking about firehose?
# 23:11 _6a68 so what i think i'm hearing is that lots of people are working on slight variations of some basic push-pull-syndicate ideas
# 23:11 aaronpk snarfed: no, when you authenticate with your own token you get your personal stream
# 23:11 _6a68 and if i understand correctly, indieweb is about dodging monoculture (except for web standards)
# 23:11 aaronpk you get your home timeline, but also things like retweets of your tweets, favorites of your tweets, etc
# 23:11 _6a68 so what i'm wondering is, what is the best way to make everybody's little projects interoperable?
# 23:12 snarfed i expect to start with the polling, since it's already done, but i should switch to streaming, at least for twitter, soon
# 23:12 _6a68 are there some APIs that might be generic enough to let people hack freely and avoid monoculture, but still be able to use somebody else's (say) twitter module or (say) couch storage backend?
# 23:13 tantek _6a68 - why worry about making "everybody's little projects interoperable"? start with the pieces that you want to deploy on 6a68.net and iterate
# 23:13 snarfed right. and we already have mf2, webmentions, etc.
# 23:13 tantek _6a68 - what's the next piece of indieweb functionality you want to add to 6a68.net?
# 23:14 tantek start there, build it. figure out the next most important piece to you, iterate.
# 23:14 _6a68 tantek: well, this is just intuition, it seems like more fun if we build a bunch of things that can talk to each other?
# 23:14 tantek and the most empowered you are to do that is on your own site
# 23:15 tantek and doing that each individually is also the most effective path to collectively making "a bunch of things that can talk to each other"
# 23:16 _6a68 ah, so i'm going back through and reading the threads i missed. i hadn't thought of webmention, that's a great example of interoperability via web standards
# 23:16 snarfed a POSSE service could just accept webmentions that point to mf2-formatted pages
# 23:17 _6a68 thanks so much for the input, all. this kind of experience is why i love hacking on the internet <3
# 23:17 aaronpk snarfed: cool yeah, how would you handle alrternate content at the POSSE destination tho?
# 23:17 aaronpk for example I often automatically or manually shorten posts that go to twitter
# 23:19 snarfed sure, good question. i haven't thought it through.
# 23:19 snarfed totally doable though, straightforward design problem
# 23:19 tantek _6a68 - it looks like you have a feed of recent posts on 6a68.net but I don't see permalinks for them - perhaps consider adding permalink URLs for each of the "posts" that you have on your home page?
# 23:19 tantek that way any one of us could reply to them, cite them etc.
# 23:20 snarfed aaronpk: i've been mainly thinking about (and starting to build) the other side: POSSEing replies into silos and generating webmentions for replies that originate inside them
# 23:20 tantek it's a bit premature to talk about POSSEing things when the things themselves lack indieweb permalinks.
# 23:20 _6a68 tantek: i appreciate the suggestion. "ship something simple now."
# 23:21 tantek right, simple things first. with talking about POSSE, interoperable components, etc. I think frankly you're getting a bit ahead of yourself.
# 23:21 _6a68 i'm more anxious to go the other way, really: get my data back from all these services
# 23:21 tantek getting your data back is a one-time batch job. not something to setup a "system" for
# 23:21 tantek setup a system for your future, and eventually create a one-time batch-import solution for your past.
# 23:22 tantek if you focus on chasing getting your data back without having a live place on your own site, you'll forever be chasing getting your data back and never make progress on your own things.
# 23:24 snarfed and if you go the one-time export route, i'd strongly encourage starting from the downloadable archives that the services all provide
# 23:26 barnabywalters facebook and g+ both provide HTML files with microformats, so you could just dump those onto your web server to start with :)
# 23:27 _6a68 ok, i'm getting pulled into other stuff, but I'm going to reread this scrollback a couple of times and follow some links. thanks again y'all, such great thought-provoking conversation
# 23:32 tantek keep in mind that doing that one-time batch export from a silo only makes sense (like you won't have to do it again) when you've successfully setup and gotten into the habit of posting on your site first, POSSEing to that silo, and never again posting directly to that silo.
# 23:33 tantek otherwise you'll just find yourself wasting time repeatedly doing exports as long as you post directly to the silo
npdoty joined the channel
# 23:35 KevinMarks_ benwerd, I was trying to get idno up on heroku, but got lost in a Mongo Driver rathole
# 23:37 benwerd KevinMarks_: sorry to hear that. Not the first time I've heard it, unfortunately
# 23:38 benwerd MongoDB: pluses and minuses there, for sure. Really like using JSON for this kind of thing though. (In cases where I'm using a db at all. eg for nanowrimo, I haven't bothered.)
# 23:40 KevinMarks_ I have a mongoDB plugin installed, getting an error in DataConceierge not finding Class Mongo
caseorganic joined the channel
# 23:40 XgF My favorite (most productive) database solution is ZODB, but that's obviously slightly eccentric and very much limited to Python
paulcp joined the channel
# 23:41 benwerd Could be a PHP version thing - sounds like it's not lazy loading.
# 23:43 benwerd ... I probably shouldn't make PHP jokes when it's been my primary language for like ten years now.
# 23:43 aaronpk benwerd: have you considered making the DB backend swappable? like using postgres as a key/val store for JSON documents?
# 23:43 aaronpk benwerd: it's ok I make PHP jkokes all the time and I've been doing it since 99
# 23:44 aaronpk I also make node.js and ruby jokes all the time so I spread it around
# 23:45 benwerd aaronpk: Yes. I actually looked for a PDO-like NoSQL abstraction library, but no such thing exists. DataConcierge is more or less there for that reason, as KevinMarks_ says, but I haven't really thought about multiple db back-ends too heavily yet.
# 23:47 KevinMarks_ all the indieweb code I've looked at has some kind of db abstraction thingy in it
# 23:50 KevinMarks_ it's making me consider camlistore again, even though that is huge
# 23:51 KevinMarks_ (this is partly 'cos of using heroku, and thus being trained to think of filesystems as volatile)
# 23:51 benwerd They've both made some pretty solid cases for text file backends
# 23:52 KevinMarks_ tantek pointed out that using git(hub) as a filesystem backing store makes some sense
# 23:56 _6a68 if you're filesystem based, have you looked into S3 for backups?
paulcp joined the channel