# 00:12 seyz joined #indiewebcamp
# 00:42 andreypopp joined #indiewebcamp
# 00:52 hmans joined #indiewebcamp
# 00:52 b0bg0d joined #indiewebcamp
# 01:32 Loqi specific solutions will.
# 01:48 friedcell joined #indiewebcamp
# 02:12 peck_lx joined #indiewebcamp
# 02:35 andreypopp joined #indiewebcamp
# 02:54 andreypopp joined #indiewebcamp
# 03:04 hmans joined #indiewebcamp
# 04:08 andreypopp joined #indiewebcamp
# 04:43 barnabywalters joined #indiewebcamp
# 05:20 brbcoding joined #indiewebcamp
# 05:40 mxuribe joined #indiewebcamp
# 05:40 pbeaulieu joined #indiewebcamp
# 05:47 scor joined #indiewebcamp
# 05:58 barnabywalters joined #indiewebcamp
# 07:23 scor joined #indiewebcamp
# 07:23 scor joined #indiewebcamp
# 07:51 b0bg0d joined #indiewebcamp
# 08:00 josephboyle joined #indiewebcamp
# 08:06 xtof joined #indiewebcamp
# 08:10 josephboyle joined #indiewebcamp
# 08:17 aaronpk bret: yes, will make it public
# 08:18 aaronpk try it now with no key
# 08:28 brianloveswords joined #indiewebcamp
# 08:56 erinjo joined #indiewebcamp
# 09:09 tantek joined #indiewebcamp
# 09:19 b0bg0d joined #indiewebcamp
# 09:46 andreypopp joined #indiewebcamp
# 09:51 tantek joined #indiewebcamp
# 09:59 tantek !tell sandeepshetty PuSH was trivial to implement as a publisher. I think it's a good spec, except for the Atom dependency, which yes we should fix to work with HTML (hAtom/h-entry) in addition.
# 10:00 Loqi Ok, I'll tell them that when I see them next
# 10:09 tantek aaronpk - can we confirm a venue sufficiently to put it on the home page and /2013 page? Instead of the bold TBD for Where?
# 10:10 seyz joined #indiewebcamp
# 10:18 aaronpk tantek: yea, almost got it settled
# 10:31 morrocco_mole joined #indiewebcamp
# 10:37 scor joined #indiewebcamp
# 10:37 scor joined #indiewebcamp
# 11:07 eschnou joined #indiewebcamp
# 11:15 spinnerin joined #indiewebcamp
# 11:46 eschnou joined #indiewebcamp
# 12:27 sandeepshetty joined #indiewebcamp
# 12:27 Loqi sandeepshetty: tantek left you a message 2 hours, 27 minutes ago: PuSH was trivial to implement as a publisher. I think it's a good spec, except for the Atom dependency, which yes we should fix to work with HTML (hAtom/h-entry) in addition.
# 12:27 andreypopp joined #indiewebcamp
# 12:28 josephboyle joined #indiewebcamp
# 12:29 aaronpk yes exactly
# 12:30 b0bg0d joined #indiewebcamp
# 12:31 sandeepshetty aaronpk: There's only so much context I can add in 140 chars... I was specifically talking about feeds...
# 12:31 tantek sandeepshetty - then you should have blogged it on your own site
# 12:31 tantek no one here should EVER be using the "only so much context I can add in 140 chars" excuse.
# 12:32 aaronpk lol yea that's not an excuse here :) I go over 140 chars all the time!
# 12:32 sandeepshetty I know :) Was on the move all day so just quickly jotted down stuff... will be converted to blog posts soon..
# 12:33 aaronpk long-form notes are ok too!
# 12:33 barnabywalters I write long form notes quite often
# 12:34 sandeepshetty It was more about the convenience of a mobile client...
# 12:34 barnabywalters sandeepshetty: sounds like an itch which needs scratching ;)
# 12:34 barnabywalters I haven’t timed it but I’m pretty sure I can post notes to my site faster than I can to twitter
# 12:34 tantek hmm, perhaps we need to start a /excuses page
# 12:35 barnabywalters due to iOS app startup times
# 12:35 tantek Loqi, Y U T.CO? :(
# 12:35 barnabywalters and the fact I have a dedicated “new note” home screen icon
# 12:35 tantek barnabywalters - nice.
# 12:35 sandeepshetty No 3G where I was and edge is very slow... so firing a mobile client and typing is faster than browsing...
# 12:35 barnabywalters with web-action style URL param autofilling you can also make the home screen icons prefill tags and such things
# 12:36 tantek sandeepshetty - sounds like you need to investigate the "email to post" stuff that barnabywalters has working.
# 12:36 barnabywalters tantek: that thread’s interesting — anil seems to be talking about the granny antipattern
# 12:37 tantek in terms of authoring latency etc.
# 12:37 tantek barnabywalters - yeah
# 12:37 tantek note that Yoz already tapped me
# 12:38 tantek before picking a fight with Anil
# 12:38 tantek who sharecrops his notes
# 12:38 barnabywalters btw tantek — just working on new version of Taproot, using 6 char SXG epoch days + SXG seconds for note IDs
# 12:38 tantek and that's what I'm going to reply to first
# 12:38 barnabywalters s/talking about/promoting
# 12:39 tantek exactly, 500 years should be good enough for anyone (reasonably human-like)
# 12:39 tantek (insert obv 640k reference)
# 12:39 sandeepshetty anyone syndicating content on their site (webmention, feed, etc)? How do you deal with security?
# 12:39 barnabywalters as someone who considers himself reasonably human-like, I concur
# 12:39 tantek what are you talking about specifically?
# 12:40 barnabywalters sandeepshetty: you mean inbound content?
# 12:40 sandeepshetty I mean into the site...
# 12:40 tantek sandeepshetty - we're all syndicating out plenty via POSSE
# 12:40 barnabywalters aaronpk and I are doing reply contexts, talk to eschnou about comments ;)
# 12:40 tantek did you not see that blog post?
# 12:40 aaronpk barnabywalters: i'm doing comments now too!
# 12:40 barnabywalters tries to leave aaronpk a comment
# 12:40 sandeepshetty If you add webmention content on your site and it's html then how do you deal with security
# 12:41 tantek sandeepshetty - eschnou obv - indieweb comments
# 12:41 aaronpk sandeepshetty: I only use plaintext
# 12:41 barnabywalters which so far has been a fair balance
# 12:41 sandeepshetty aaronpk: so you strip out html?
# 12:41 barnabywalters sandeepshetty: from my reply contexts, I strip HTML then auto-link
# 12:41 aaronpk I do, but only because the mf2 parser does it for me
# 12:41 aaronpk and then yea I re-link. which has the interesting side effect that other people's hashtags end up linking to my own tag pages
# 12:41 barnabywalters aaronpk: I should mention: be VERY careful about that
# 12:42 sandeepshetty this brings me to stuff I've been thinking about feeds... they have the same problem no?
# 12:42 barnabywalters all it takes is someone to mark up their content as e-summary instead of p-summary and you’ve got injection
# 12:42 aaronpk barnabywalters: you should document that on the parser :)
# 12:42 aaronpk i'm using the p-name only
# 12:42 barnabywalters aaronpk: I do mention that the parser does no filtering and all data should be suitably filtered, but I should add that particular case as an example
# 12:42 barnabywalters aaronpk: e-name ;)
# 12:43 aaronpk ah yes plz document :)
# 12:43 sandeepshetty so my question is: Are html feeds broken fundamentally?
# 12:43 sandeepshetty all the big silos only do text
# 12:43 sandeepshetty (twitter, facebook, g+)
# 12:44 barnabywalters only do text? how’d you mean?
# 12:44 tantek sandeepshetty - "all the big silos only do…" is also a poor justification for anything
# 12:44 tantek really bad reasoning/methodology
# 12:44 sandeepshetty it wasn't meant as justification..
# 12:44 tantek mobile email clients are MUCH better behaved on slow mobile networks (Edge etc.) compared to any kind of mobile native posting client.
# 12:44 tantek all the big silos kill permalinks
# 12:44 tantek sandeepshetty - doesn't your mobile device have a web browser?
# 12:44 tantek why bias research towards bad actors?
# 12:45 barnabywalters gives aaronpk a high five
# 12:45 aaronpk high fives barnabywalters
# 12:46 sandeepshetty tantek: I prefer looking across the board... there's always stuff to learn...
# 12:47 eschnou joined #indiewebcamp
# 12:48 sandeepshetty I'm leaning towards plain text feeds...
# 12:48 aaronpk I was assuming plaintext feeds without realizing it
# 12:49 tantek sandeepshetty - sure, there's plenty to learn, even from bad examples, but it's always better to look at good examples first
# 12:49 aaronpk although there are well-established practices for sanitizing html at this point. just look at wordpress, mediawiki, forums, etc etc
# 12:53 tantek aaronpk - sounds like something you could add to the wiki
# 12:53 tantek since you're saying it's "well-established" ;)
# 12:54 bretolius joined #indiewebcamp
# 12:54 sandeepshetty sanitizing html seems to have edge cases..
# 12:54 sandeepshetty and stripping it means you loose context
# 12:54 aaronpk that's why i'm doing plaintext right now
# 12:55 tantek aaronpk - I think that's reasonable, get something simple built quickly
# 12:55 sandeepshetty will you ever be implementing a feed reader on your site? I want my site to be both a medium for publishing and consuming..
# 12:55 tantek aaronpk even start /plaintext and dump these thoughts there
# 12:56 tantek and we can link to them from /comments
# 12:56 sandeepshetty aaronpk: sorry I might have missed this but how are you ensuring you're getting plain text when the content is external?
# 12:57 tantek sandeepshetty - I think he's busy documenting that at /plaintext
# 12:57 tantek though that makes a good FAQ for that page
# 12:57 aaronpk sandeepshetty: apparently i'm not, but I thought I was. my assumption was that the "name" property from the mf2 parser was always plaintext
# 12:57 aaronpk so i will add a strip_tags() around the name
# 12:57 tantek aaronpk - anything can be e-*
# 12:58 tantek or someone can also embed literal < > & in the plaintext as well
# 12:58 eschnou aaronpk, did I just read you display indie comments?
# 12:58 tantek so you don't actually want to strip names
# 12:58 tantek strip tags
# 12:58 aaronpk none of the examples I was looking at had e-name
# 12:58 sandeepshetty and with what eschnou there seems to be a big security hole.. which is why I left out the bit about what to do with the webmention in the spec
# 12:58 tantek if someone puts "&" in the text, you should keep it, and escape when publishing if necessary
# 12:58 aaronpk tantek: yea my goal is plaintext, whatever that takes
# 12:58 tantek aaronpk - use of < > is plain text
# 12:59 tantek everything I'm typing into IRC is plain text
# 12:59 tantek and if I name my blog post: About the <time> element
# 12:59 tantek it should remain that way in whatever system
# 12:59 tantek if it turns into
# 12:59 barnabywalters if you htmlencode the entities from php-mf2 the worst that will happen is that you’ll get html showing up in the output
# 12:59 tantek About the element
# 12:59 tantek then you've got a bug
# 12:59 tantek you're losing plaintext
# 12:59 tantek so just "strip tags" is wrong
# 12:59 barnabywalters which is okay as it’s a fault of the published
# 13:00 aaronpk tantek: yes that is what I want to do. if you publish html on your site that renders as <time> you have to have escaped those characters
# 13:00 aaronpk which means by the time I get it, i don't get <, I get <
# 13:00 sandeepshetty tantek: those should use htmlspecialchars()
# 13:00 tantek sandeepshetty why?
# 13:00 tantek you can't assume the data will go into an HTML context
# 13:00 sandeepshetty tantek: re: and if I name my blog post: About the <time> element
# 13:00 tantek only the publisher can assume that
# 13:01 tantek the (re)publisher
# 13:01 tantek if I put it into an email subject line
# 13:01 tantek I don't want to see &
# 13:01 tantek that's just dumb
# 13:01 tantek premature html escaping causes plaintext corruption
# 13:01 barnabywalters aaronpk: “i don't get <, I get <” is not the case
# 13:02 tantek.com edited /2013/Guest_List (+405) "/* Creators */ putting in placeholder images for folks that forgot to - based on what they have at their site, github, twitter, or otherwise linkto from the contact/about pages. feel free to change, but something is better than nothing, public URLs" (
view diff )
# 13:02 tantek ok folks I just added a bunch of images for the slackers
# 13:02 eschnou aaronpk, hmm.. just commented and it didn't show up...
# 13:03 tantek so if you want a different image (ahem, caseorganic, termie), you might want to edit your RSVP
# 13:03 aaronpk eschnou: did you send a pingback or webmention?
# 13:03 eschnou aaronpk, pingback
# 13:03 sandeepshetty tantek: Whats the solution to the '&' problem?
# 13:03 barnabywalters (output from php-mf2, I have not tested with any other parsers)
# 13:03 aaronpk eschnou: did you get a successful reply from the pingback?
# 13:04 eschnou is digging in the logs
# 13:04 eschnou aaronpk, mm... no_link_found
# 13:05 sandeepshetty so the reason I brought this up is that if everyone plans to do plain text feeds then how do we add the lost context (microforats related to citation, etc.)
# 13:05 tantek is going to wait for aaronpk to document this on /plaintext
# 13:05 aaronpk can't right now, mostly distracted!
# 13:05 barnabywalters sandeepshetty: it’s safe to parse microformats and display the data in whatever way you see fit — provided the output is all htmlencoded
# 13:05 tantek sandeeshetty - feel free to add your questions to an FAQ section on the /plaintext page
# 13:06 tantek heck, CREATE the page if you have to
# 13:06 sandeepshetty barnabywalters: that has the '& problem that tantek pointed out..
# 13:06 aaronpk please do heh I will add stuff to it later
# 13:06 tantek sandeepshetty - type it into the new page - there's already too many hanging questions / answers here in IRC to track
# 13:08 barnabywalters if you’re displaying the content as HTML you can assume it’s going in an HTML context and html-encode it
# 13:08 barnabywalters the problem is solved outside the parser
# 13:09 tantek barnabywalters - the parser should not be escaping plain text
# 13:09 tantek except for JSON - since that's where it puts it
# 13:10 eschnou I did not follow you plaintext discussion, but is it not just a matter of strip_tags, and allowing some safe tags (e.g. a anchors, span, etc?)
# 13:10 sandeepshetty eschnou: there are always edge cases with that approach..
# 13:10 tantek eschnou - unfortunately no - nothing is "just a matter of…" right now
# 13:10 tantek WE can't follow the plaintext discussion because we exceeded IRC thread readability
# 13:11 tantek hence pointing people to write up their questions/answers on the wiki
# 13:11 aaronpk we broke irc
# 13:11 tantek which forces a focusing function
# 13:11 tantek aaronpk - IRC just has limitations
# 13:11 tantek like Twitter or anything else
# 13:11 tantek just need to recognize them and know when to move things elsewhere
# 13:11 aaronpk yep good call
# 13:11 eschnou tantek, as long as we don't move to a mailing list, I'm fine :)
# 13:11 tantek eschnou - you said it brother
# 13:12 tantek mailing lists are worse for all the above
# 13:12 sandeep.shetty.in created /plaintext (+285) "Created page with "* Displaying HTML content from external sites (feeds, webmention, etc.) posses a security risk. * Prefer plaintext * Whats the right way to sanitize external content. * However c..."" (
view diff )
# 13:12 tantek so I start writing an @-reply, and then realize I can improve the UI-flow of my @-reply authoring, so I jump into writing code...
# 13:13 tantek itch -> scratch
# 13:13 eschnou I like IRC, and the fact we log the channel, I actually like to get my morning coffee reading your discussions of the night :-)
# 13:13 tantek (certainly not itch ->
# 13:13 tantek scratch ) :P
# 13:13 aaronpk eschnou: sometimes i wish i was in your timezone
# 13:15 sandeepshetty Jsut added some points /plaintext.. feel free to add stuff... I'm new to this
# 13:16 aaronpk eschnou: what's the post you tried to reply to me with?
# 13:16 eschnou aaronpk, xkcd
# 13:16 barnabywalters just caught up — tantek: RE parser shoudn’t be escaping plaintext: it isn’t
# 13:16 barnabywalters unless I am missing something
# 13:17 eschnou but the issue may be there: that link actually 301 to my domain at eschnou.com, maybe that your code does not follow these redirects
# 13:17 sandeepshetty barnabywalters: how do decide if the external content is plaintext or html?
# 13:17 eschnou aaronpk, I should use the cname domain, but my current implementation makes it tricky and I didn't spend the time on this yet.
# 13:18 barnabywalters sandeepshetty: go by the microformats spec (assume plaintext from name, summary and HTML from content) and if you overencode, it’s the publisher’s problem which needs to be fixed on their end
# 13:18 barnabywalters if in doubt (dealing with unknown properties), encode
# 13:18 aaronpk barnabywalters: wait what! now you say name has to be plaintext?
# 13:19 sandeepshetty This isn't a publishing problem.. it's a security one... assume malicious parties...
# 13:19 aaronpk eschnou: yea probably I am not following the redirect...
# 13:19 barnabywalters sandeepshetty: I am
# 13:20 sandeepshetty barnabywalters: you'll have to explain.. are you saying you won't encode the name and summary?
# 13:20 barnabywalters with the logic I described, the worst that happens assuming a malicious party is you get some escaped HTML showing up on your site
# 13:20 barnabywalters okay, I’m writing this up with code samples :)
# 13:23 aaronpk so barnaby just replied to eschnou's reply
# 13:23 aaronpk I want to be notified of barnaby's reply
# 13:23 termie tantek: looking now
# 13:24 tantek aaronpk - sounds like a use-case you should document on /webmention
# 13:24 termie tantek: that's a good image, it will make people not recognize me
# 13:24 tantek termie LOL
# 13:24 termie tantek: so i can sneak up on them and get all indieweb
# 13:24 eschnou aaronpk, well, I think this is more the case of you 'subscribing' to the comments feed
# 13:24 termie and yay caseorganic, she was fun to talk to at xoxo
# 13:24 eschnou aaronpk, how to subscribe and to which feed is what to discuss
# 13:25 aaronpk ooh I do like the subscribe idea
# 13:25 tantek eschnou - subscription eventually comes down to PuSH notifications
# 13:25 tantek so if we're talking about a notification based protocol, we should consider multiple possibilities
# 13:25 eschnou tantek, absolutely and the discussion we previously had on PuSH and microformats support
# 13:25 tantek both PuSH based and webmention based
# 13:26 tantek they're both just notification protocols
# 13:26 tantek no reason to dismiss either one out of hand
# 13:26 tantek we don't know what combination of each would actually work best in this use-case
# 13:26 eschnou tantek, well, as a (lazy) developer, I would mind if we just settle on one for now :-)
# 13:26 tantek hence once again
# 13:26 sandeepshetty People that have implemented PuSH: do you also subscribe to feeds with regular rss/atom?
# 13:26 tantek eschnou - sometimes an ounce of thinking can prevent a pound of wasted coding
# 13:26 tantek that's the problem with the "just settle on one for now"
# 13:26 aaronpk sandeepshetty: I publish to a hub, but don't subscribe to anything because it's too hard
# 13:26 tantek methodology
# 13:27 tantek a-ha - see now that's interesting
# 13:27 tantek subscribing too hard
# 13:27 tantek aaronpk - but you have no problems receiving webmentions
# 13:27 aaronpk yea, subscribing in PuSH is too hard because of the extra round-trip the request has to make
# 13:27 tantek so you ARE subscribing
# 13:27 sandeepshetty I would just prefer building in a feed reader into my site.. why go PuSH?
# 13:27 tantek sandeepshetty - depends on how "real time" you want it
# 13:28 tantek if you want your feed updates to be as fresh as txt messages on your phone, you want PuSH
# 13:28 barnabywalters and how many resources you want to waste
# 13:28 aaronpk no, webmention isn't subscribing, because anybody can send me a webmention unsolicited. subscribing implies intent on my behalf
# 13:28 sandeepshetty I think real time is overrated in this case..
# 13:28 sandeepshetty a 5 min dealy won't make a difference..
# 13:28 barnabywalters webmention and PuSH are solving different problems
# 13:28 aaronpk realtime is not overrated. I get IRC notiications about all sorts of stuff in realtime and it's great
# 13:28 tantek aaronpk - I think you meant was, subscribing implies an open connection / relationship of sorts
# 13:28 tantek whereas webmentions are purely transactional
# 13:28 tantek however, you asked for a relationship
# 13:29 tantek you said you wanted to receive webmentions
# 13:29 tantek for a specific post's comments
# 13:29 aaronpk anybody can send me a webmention whether I want it or not
# 13:29 sandeepshetty tantek: webmentions are unsolicited
# 13:29 aaronpk that's very different from subscribing
# 13:29 tantek but that's not what you asked for
# 13:29 tantek you specifically asked for future webmentions for a specific post's comments
# 13:29 tantek which *is* a request for a relationship
# 13:29 aaronpk no, I said "I want to be notified of barnaby's reply", nothing about protocol
# 13:30 tantek and presumably every other reply
# 13:30 sandeepshetty tantek: one is with a specific person the other with everyone..
# 13:30 tantek not just barnaby's
# 13:30 aaronpk yes in this case I really do want to subscribe to that thread, and the thread on barnaby's site too. which is going to quickly turn into a de-duplication problem :)
# 13:31 aaronpk this needs a new page lol
# 13:31 tantek aaronpk - so, you're going to create that right? ;)
# 13:31 aaronpk yes this time
# 13:32 tantek gives up on incremental coding improvement for now, but has noted what functions he needs to change to get the incremental UI improvement he wanted.
# 13:35 eschnou aaronpk, I cheated and pinged manually with the good link and it is there, yeah!
# 13:36 sandeepshetty I like hackable protocols :)
# 13:37 barnabywalters I am up to speed on the e-* p-* issue now
# 13:37 barnabywalters there is indeed a potential overescaping problem when dealing with incoming e-* content
# 13:38 barnabywalters my initial thoughts on how to fix it are to see if the content which is expected to be e-* parses as HTML
# 13:39 barnabywalters the problem mainly is that if you expect a property to be parsed as e-* but it was marked up and parsed as p-*, you’re either going to get incorrect HTML or mistake encoded HTML for literal HTML
# 13:40 tantek barnabywalters - worth bringing that up in #microformats
# 13:40 barnabywalters my brain is frazzled from a day of statistical stuff and an evening of whatever it was we were just discussing ;) off to bed now
# 13:40 Loqi buenas noches
# 13:41 tantek poor kid, we wore him out.
# 13:42 tantek this channel is getting intense!
# 13:42 tantek can I just say how much you guys all rock?
# 13:43 tantek is still doing lots of catch-up coding on the whole comments / reply-contexts thing.
# 13:43 sandeepshetty aaronpk: re getting notifications about new comments.. old school blogs used to have a "email me when a new comment is posted" type feature
# 13:43 aaronpk yea! i think that's what eschnou was alluding to
# 13:43 tantek sandeepshetty - good point!
# 13:43 tantek aaronpk - can we use any of those email providers you mentioned/listed on the wiki for that use-case?
# 13:43 eschnou sandeepshetty, I actually do have this feature :-)
# 13:43 aaronpk hah yes, but I don't *really* want to get them via email
# 13:43 tantek eshnou - did you implement it? or do you use an email provider to send?
# 13:44 tantek aaronpk - it's for your readers!
# 13:44 sandeepshetty it's better to solve is orthogonally like that...
# 13:44 tantek and it's something they can get from your site, that they can't get from Twitter!!@
# 13:44 eschnou tantek, well, I just use the php send email stuff
# 13:44 tantek eschnou - good to know
# 13:44 eschnou tantek, I guess you can configure the smtp gateway of your choice etc.
# 13:44 aaronpk heh yea I gave up on sending email from my own servers a long time ago. really not fun dealing with delivery issues
# 13:45 sandeepshetty aaronpk: it isn't meant to be super realiable..
# 13:45 tantek worth documenting both approaches
# 13:45 eschnou one of the challenge is that I do all this synchronously, but when the comment thread gets long, it means someone posting a new comment sees the waiter rolling for a while...
# 13:46 aaronpk eschnou: I implemented a simple task queuing system into p3k this weekend
# 13:46 eschnou the good thing is I never really have much comments on my site anyway :-)
# 13:46 sandeepshetty eschnou: You can jsut fork off a process...
# 13:46 eschnou aaronpk, that's what I need indeed...
# 13:46 aaronpk check out beanstalkd. works with any/multiple language
# 13:46 aaronpk super simple and lightweight
# 13:46 eschnou thanks, will have a look
# 13:47 sandeepshetty you could also do a lo-fi solution... add an entry and have a corn send out the email..
# 13:47 sandeepshetty If you can't tell I'm all for lo-fi solutions :(
# 13:48 Loqi aww, cheer up
# 13:48 eschnou well, I'm off as well, cheers all!
# 13:48 sandeepshetty argh I should just go to bed..
# 13:48 aaronpk yea that would be cool too. simpler/slower queuing, fewer moving parts
# 13:49 aaronpk sounds like we need a /queueing page now too
# 13:49 b0bg0d joined #indiewebcamp
# 13:49 texburgher joined #indiewebcamp
# 13:49 sandeepshetty Time for me to get some shut eye... later.
# 13:52 sandeep.shetty.in created /queueing (+158) "Created page with "Approaches to handle work asynchronously * A proper queuing system like beanstalkd, gearman * Make an entry somewhere for the task and then have a cron do it."" (
view diff )
# 13:52 tantek I was using an h-as-reply class for my replies, but now I'm finding I don't have much use for it
# 13:53 tantek for display purposes (e.g. reply contexts) I can simply ask, does this post have an in-reply-to URL or not?
# 13:53 tantek and if does, display it as a reply with reply context
# 13:55 tantek can I just say I love having #ownyourdata @-reply support?
# 13:56 tantek like I can finally have threaded conversations with Twitter sharecroppers again.
# 13:56 tantek speaking of which, I have to respond to Anil.
# 14:05 andreypopp joined #indiewebcamp
# 14:20 Loqi let's discuss further (ttk.me t4QE5)
# 14:21 tantek ok, done replying. Feel free to add more if you think I missed anything.
# 14:21 tantek (still holding the but-why-are-you-still-sharecropping-on-twitter-then ace for later use)
# 14:22 tantek and keep an eye out for genmon in case he actually shows up!
# 14:23 tantek he's in the UK so he may be off for the night.
# 14:23 aaronpk you can !tell him something so you get pinged when he joins!
# 14:23 tantek aaronpk - I'm not sure what his irc nick will be
# 14:23 tantek though I can guess genmon
# 14:23 tantek what's the harm I suppose
# 14:24 tantek !tell genmon welcome to #indiewebcamp! how may we service your #indieweb needs? :)
# 14:24 Loqi Ok, I'll tell them that when I see them next
# 14:29 tantek man, posting all those @-replies really makes it clear I need reply-context on my replies on my own site
# 14:29 tantek not very useful without the context :/
# 14:29 tantek or at least "View conversation on Twitter" rel syndication links
# 14:33 andreypopp joined #indiewebcamp
# 14:36 aaronpk +1 for syndication links at the very least
# 14:46 tantek aaronpk, thanks, I do appreciate that input
# 14:46 tantek in figuring out what to prioritize implementing, sometimes it's hard for me to tell which itch is hurting more
# 14:46 aaronpk that would at least let me auto-discover your syndicated posts when replying. also when reading your site it's not a big deal to click through to a tweet most of the time
# 14:46 tantek so outside perspective can help a lot there :)
# 14:54 tantek so e.g. "View on Twitter" and "View conversation on Twitter" links then?
# 14:54 tantek would that be click-through discoverable enough?
# 14:54 aaronpk tantek: yes those titles sound completely reasonable
# 14:55 aaronpk or instead of "View on Twitter" it could be "Reply on Twitter" if you want something more actionable
# 14:55 aaronpk that was my reasoning for using "reply on twitter"
# 14:55 tantek I don't want to send a message that I'm advocating using Twitter for replies
# 14:55 aaronpk that's the other way of looking at it
# 14:56 tantek since that doesn't seem like a good idea to propagate vis-a-vis supporting an #indieweb way
# 14:56 aaronpk i'm curious what you think of #3 on http://indiewebcamp.com/comment-notifications (conversation threading)
# 14:59 tantek 1 sec, updating Falcon working on
# 15:03 tantek.com edited /Falcon (+489) "/* Working On */ add more specifics/reasons for rel syndication link use-cases" (
view diff )
# 15:04 tantek I'm really glad I'm documenting this itching intuition before/while implementing this stuff, as then it feels much more real than the hypothetical "if only …" or "I wish…" type declarations people usually make
# 15:04 tantek as I implement each one, I plan to move the reasoning bits to the generic pages on each subject
# 15:04 tantek seems like each is useful as a case study
# 15:05 tantek so re Option 3
# 15:05 tantek a few things
# 15:05 tantek I don't think we can avoid "threading" at least semantically whatever we do. Threading happens naturally due to the freedom of hypertext structures and our use of in-reply-to links
# 15:06 tantek this is independent of any protocol decisions/design
# 15:06 aaronpk i'm not talking about threading in terms of nesting and indented displays
# 15:06 tantek I'm saying it's useful to acknowledge that semantic threading is inevitable
# 15:06 tantek as that helps inform the design of anything related
# 15:07 tantek we can choose to flatten or not in protocol or in display for other reasons
# 15:07 tantek I think Option 2 is something we're going to have to solve any way for multi-replies
# 15:08 tantek which IMO is the right way to approach solving it - solve multireplies
# 15:08 tantek in terms of "much more careful markup to make work properly", perhaps we can combine in-reply-to and breadcrumbs
# 15:09 spinnerin joined #indiewebcamp
# 15:11 aaronpk is not a fan of breadcrumbs
# 15:11 tantek yeah - the presentation is kinda ugly IMO
# 15:11 aaronpk tends to lend itself to overengineered sites
# 15:11 aaronpk I used to implement strict breadcrumbs both on my personal site and also commercial sites I was making, eventually just dropped the whole thing.
# 15:12 aaronpk in like 2003-2004
# 15:12 tantek they're only useful if they help the reader with a sense of where they are
# 15:12 tantek and if the crumbs are actually useful places to go
# 15:12 tantek re: Option 1 btw - I think subscribing to comments is fascinating and part of a larger "subscribing to an indieweb site" discussion
# 15:12 aaronpk I find the "Where you are" only really applies when you're reading really big sites that have a full TOC like reference manuals
# 15:12 sandeepshetty joined #indiewebcamp
# 15:13 tantek aaronpk - yeah, deep TOCs
# 15:13 sandeepshetty Ok I couldn't sleep...
# 15:13 tantek welcome back!
# 15:13 sandeepshetty aaronpk: re Comment-Notifications I think what you are looking for is doen by salmon..
# 15:13 sandeepshetty sending comments upstream (hence the name salmon)
# 15:14 aaronpk possibly. I would have to dig back into xml land to figure that out, and I remember it being really painful before
# 15:14 sandeepshetty Might be interesting to think about how it could be done with webmention
# 15:15 sandeepshetty tantek: re the '&' bug: if you encode "tantek & aaronpk" only during displaying it as html, then it still "appears" as "tantek & aaronpk"
# 15:15 tantek sandeepshetty - that's right
# 15:15 tantek that's the way to do it
# 15:15 sandeepshetty So always encoding plaintext is the way to go.l..
# 15:15 tantek push the encoding responsibility to the last moment
# 15:16 tantek sandeepshetty - that's what I do with my own content!
# 15:16 sandeepshetty Thats how I would do it anyway.. its the responsibility of the templating engine in my case..
# 15:18 sandeepshetty also re composite streams (discussion from 2013-05-28): I guess I'm the only on that has only a single post type...
# 15:19 Loqi sandeepshetty meant to say: also re composite streams (discussione from 2013-05-28): I guess I'm the onely one that has onely a single post type...
# 15:19 tantek sandeepshetty - no, the default is that people only have a single post type. a blog post.
# 15:19 tantek what's different is having multiple post types
# 15:19 tantek er, less common
# 15:19 sandeepshetty why do you need multiple post types?
# 15:20 tantek sandeepshetty - because I like to post notes and articles.
# 15:21 tantek and now replies
# 15:21 tantek though for me replies aren't a new "type" as much as an aspect of any post type
# 15:21 sandeepshetty to me the only diff between notes and articles is the title..
# 15:21 sandeepshetty its more a last mile display thing than anything else..
# 15:21 aaronpk sandeepshetty: my notes are plaintext, articles have formatting. also see: http://indiewebcamp.com/Semantics_Of_Article-Note_Distinction
# 15:22 sandeepshetty I just do markdown for both
# 15:22 aaronpk I did, but recently converted by notes to all plaintext for a number of reasons
# 15:22 sandeepshetty be interesting to know the reasons...
# 15:23 tantek yeah - that page (distinction)
# 15:23 tantek sandeepshetty - for me, the display is quite different
# 15:23 tantek and really this is all UX/design based
# 15:23 sandeepshetty doesn't it kinda suck to have the ux dictate the backend?
# 15:23 tantek because I want to make things look different, it makes sense to do so based on an explicit semantic difference.
# 15:23 aaronpk mostly due to syndicating content elsewhere. pushing to twitter, if I had <a href="http://tantek.com">Tantek</a> somewhere, twitter would only display Tantek. Same for when barnaby's site would pull out my h-entry.
# 15:24 tantek sandeepshetty no! UX should always dictate the design of EVERYTHING
# 15:24 tantek user-centered design
# 15:24 sandeepshetty aaronpk: I just push markdown... works well for me..
# 15:24 aaronpk sandeepshetty: you push markdown out to twitter?
# 15:24 b0bg0d joined #indiewebcamp
# 15:24 sandeepshetty yeah let me dig up some tweets..
# 15:27 sandeepshetty this for example: https://twitter.com/sandeepshetty/status/268135012493324289 it's just a bookmark (which is also just a regular post on converspace)
# 15:28 tantek wow that's fascinating
# 15:28 sandeepshetty tantek: From a stream perspective it's all just content I'm publishing.... why the need for them to look different?
# 15:29 tantek sandeepshetty - we should document this as another POSSE to Twitter alternative approach
# 15:29 tantek sandeepshetty - aesthetics
# 15:29 tantek personal preference
# 15:30 sandeepshetty I should mention I don't auto-post to twitter... I think each medium (twitter/facebook) has it nuances.. so I hand edit before posting..
# 15:31 sandeepshetty plus I work on making it enticing to click through to go to the site..
# 15:31 tantek sandeepshetty - that's a totally fine approach to POSSE
# 15:32 aaronparecki.com edited /Semantics_Of_Article-Note_Distinction (+312) "/* Personal, subjective distinctions */ adding my thoughts" (
view diff )
# 15:32 sandeepshetty aaronpk: what was the problem with posting markdown to twtiter?
# 15:32 tantek as long as you're *only* posting to Twitter stuff that is at least somewhat on your own site
# 15:32 aaronpk i wasn't posting markdown to twitter, I was posting the text to twitter converted from my HTML.
# 15:32 aaronpk so <a href="http://tantek.com">Tantek</a> would end up on twitter as just "Tantek"
# 15:33 aaronpk and if I linked to an article, it would just get the name of the link
# 15:33 aaronpk so I switched to plaintext just like twitter does it
# 15:33 tantek aaronpk - that's too bad - because you could auto-convert that to an @-reference
# 15:33 tantek (like what Barnaby does)
# 15:33 sandeepshetty ah.. I just have everything in markdown hence the need for me to move to https://github.com/converspace/markaround eventually so that I can do microformts in plain text
# 15:33 aaronpk for people yes, but not for other links
# 15:36 sandeepshetty tantek: When you have sometime take a look at https://github.com/converspace/markaround its a very early draft.. I'm hoping to start implementation in a months time... I know there are something there that you don't like (based on your markdown wiki page)
# 15:37 andreypopp joined #indiewebcamp
# 15:37 tantek oh you've seen my markdown rant page huh? ;)
# 15:38 tantek I see we agree on some things "only one way to do something"
# 15:38 aaronpk heh I also have my own extended markdown
# 15:38 tantek oh look, References!!!
# 15:39 sandeepshetty aaronpk: I also do extended markdown on sandeep.io (stuff like @names , hastags, etc..)
# 15:39 tantek sandeepshetty - yeah, some of it I like, others not
# 15:39 sandeepshetty but I don't think of them as part of content. Just a presentation thing.
# 15:39 tantek I'm pretty big on the normalish readable text principle
# 15:39 tantek rather than cribbing from MediaWiki syntax
# 15:40 sandeepshetty The big things for me there are the consistent link syntax... check img linking and regular linking..
# 15:41 sandeepshetty and I was going for faster typing.. because it almost always gets converted when viewing...
# 15:42 sandeepshetty without giving up too much on raw readability though...
# 15:42 tantek all the indented list stuff you have gives up on readability IMO
# 15:42 tantek (it's one of my pet peeves with media wiki syntax)
# 15:46 sandeepshetty It's a tradeoff for being able to type in textareas... tabs are big pain in the *** inside textareas
# 16:03 tantek I guess not quite out ;)
# 16:36 josephboyle joined #indiewebcamp
# 16:38 tantek um I just came up with I think another post type, that's kind of particular, maybe an edge case
# 16:38 tantek subset of drawing
# 16:39 tantek I guess technically a photograph thereof
# 16:39 aaronpk why is that not just a photo?
# 16:39 tantek but I'd display it differently than just a photo
# 16:39 tantek because it has a name and caption built into the image
# 16:39 tantek but I'd want to reflect those semantics in alternate markup
# 16:39 aaronpk huh interesting
# 16:40 aaronpk some of my photos have a name and caption though
# 16:41 aaronpk i think it's just a photo post that also has a name and text content
# 16:44 tantek one difference - a comic would likely have multiple panels
# 16:44 tantek and dialog
# 16:44 tantek and I'd want to mark that up some way so that it was accessible
# 16:48 tantek I guess I'd need to post a real world example to play with
# 16:54 tantek how sad - it uses longdesc - which 404s
# 17:09 tantek the path sandeepshetty is taking is fine. minimal implementation based on polling, then he can assess later whether he needs/wants PuSH or not
# 17:09 tantek whether any one of us agrees or not, doesn't matter, it's part of the strength of the indieweb, each implementer decides for themselves what approach to prioritize
# 17:09 tantek in contrast to monocultures
# 17:33 tantek ok so for comics posts
# 17:33 tantek I'm thinking they resemble articles since they typically have a name
# 17:33 tantek at least XKCD style
# 17:34 tantek and I'm thinking of putting the whole thing in an <object>
# 17:34 tantek and having the fallback heading etc. in the <object>
# 17:35 tantek so it'll be a special kind of article - we'll see how much I have to change my CSS to make this work
# 17:39 tantek but I'm reusing the "a" shortcode for it
# 17:40 tantek er, I mean "b". ahem.
# 19:02 scor joined #indiewebcamp
# 19:02 scor joined #indiewebcamp
# 19:23 brianloveswords joined #indiewebcamp
# 19:47 tilgovi joined #indiewebcamp
# 19:55 bret although, going through the comments makes me frown. Lots of "twitter is still good, lets fail to realize the bigger issue here"
# 19:55 bret as if twitter is going to a) ever going to have as big of an impact as FB, and b) not be susceptible to all of the same problems
# 19:56 pdurbin twitter used to be great. facebook never was
# 19:56 bret facebook used to be okay when it was highschool and college
# 19:57 pdurbin I was out of college already when facebook was just college
# 19:58 bret in retrospect, given what it turned into though, i see that this was a drastic illusion and mistake us kids made in promoting it
# 19:59 bret oh well oops
# 20:38 tantek joined #indiewebcamp
# 21:04 tantek joined #indiewebcamp
# 21:33 andreypopp joined #indiewebcamp
# 22:18 andreypopp joined #indiewebcamp
# 22:22 cweiske joined #indiewebcamp
# 22:41 tantek well one distinction I've found between a comic and an article is that I don't want to display an explicit <h1> heading before a comic - since the comic has the heading built into the image
# 22:41 tantek so I'm going to use h-as-article h-as-comic
# 22:43 cweiske what about a hidden topic?
# 22:44 cweiske because the image is not easily machine-readable
# 22:44 tantek cweiske - hence I'm using <object> for the image with text fallback inside
# 22:45 tantek but in this case I may be able to simply query the entry-title node to see if it has an entry-content ancestor
# 22:45 b0bg0d joined #indiewebcamp
# 22:53 tantek and if so - to not duplicate it in its own h1 outside the entry-content
# 22:53 tantek e.g. in a stream context
# 23:38 tantek my first comic post:
# 23:56 indiewebcamp-vis joined #indiewebcamp