2013-10-25 UTC
# 00:01 tantek acegiak - moderation will get quickly overwhelmed and depressing
# 00:01 tantek that's one of the reasons Akismet is giving up on Pingback
# 00:01 tantek moderation is not a reasonable/scalable/sustainable part of the answer to defense against spam
caseorganic and bnvk joined the channel
bnvk joined the channel
earplugs and tpinto joined the channel
tpinto, paulcp, tantek, KevinMarks, mikeal, caseorganic and cweiske joined the channel
eschnou, scor, tpinto and tantek joined the channel
# 05:56 cweiske when receiving webmentions from and to fragmented URLs, shall the existance of elements with that name be checked, too?
# 05:56 cweiske example.org/foo#bar would need to contain a id="bar" html tag
eschnou, tpinto, npdoty and andreypopp joined the channel
glennjones and jsmansart joined the channel
abrereton, friedcell, andreypopp, caseorganic, tpinto, Jihaisse, glennjones and pfefferle joined the channel
tpinto and barnabywalters joined the channel
# 10:18 acegiak has anyone tested semantic linkbacks with pfefferle's fix?
# 10:32 pfefferle not yet… had a lot to do the last days… perhaps on weekend
# 10:32 acegiak hmm. now im not getting anything, which is interesting
bnvk joined the channel
bnvk and glennjones joined the channel
# 11:32 pfefferle perhaps we could extract the "parser" code into an external lib to use unit testing...
# 11:43 pfefferle that's true, but I think it's easier to test the parser separately from the pingback/webmention stuff to check where the error occurs
# 11:46 acegiak I'm just replying to myself en-masse and then I'll delete them all later
Acidnerd, Jihaisse and abrereton joined the channel
# 12:13 pfefferle barnabywalters is it against the spec to extend the h-* entries with a "value" attribute that holds the complete html-code of the element?
# 12:16 acegiak ok so it's definitiely the hentry walker that's giving me trouble
# 12:16 pfefferle that would make it way easier to check if the link is inside the entry… the rest should be only to check the type of the mention
# 12:19 pfefferle there are a lot of ways to markup "replies", "likes" or "rsvps" and it is a lot to check to get the right h-entry
# 12:19 pfefferle it would be easier to get the h-entry with pregmatching the link and THEN checking for the type (like, reply, rsvp)
# 12:26 barnabywalters you should never have to do that — there should only be one top-level h-entry on post permalink pages
# 12:27 acegiak ok I need to head to bed, I'll do some more tomorrow
# 12:28 pfefferle nevertheless do i have to check a lot of attributes to be sure the link is part of the h-entry
eschnou, scor and tantek joined the channel
# 13:48 tantek pfefferle: The link doesn't have to be part of the entry. It just typically is.
# 13:50 tantek Where are you getting the impression/requirement that you have "to be sure the link is part of the h-entry"?
# 13:51 pfefferle tantek but that is a good way to spam… if someone has a link to an article of my page in a global area (for example in a blogroll), then potentially every webmention from every site is a valid reply!?!
# 13:51 tantek pfefferle: Spam has to do with *who* it is coming from, not what page.
# 13:52 pfefferle not really, because everyone could send webmentions from the command line… if it is his page or not
# 13:52 tantek And in such and example, those would only be mentions not replies.
# 13:54 tantek I would also file that "problem" aa theoretical until we see a real world example.
bnvk joined the channel
# 13:57 pfefferle ok, let me summarize that… validating a request could be: check if site has mentioned the target-url anywhere on the site… get the first h-entry and use it as a for example reply?
# 14:00 pfefferle ah ok… Now I think I got it… Check if site has mentioned target-url… Check first h-entry if it has any kind of reply merkup -> it is a reply… if not it is a mention… ?!?
# 14:00 barnabywalters how to determine whether or not a page links to another, and if it does, of what type is the link
# 14:01 tantek Ok ok I'll get my laptop. Still dark outside here in SF.
# 14:02 pfefferle yes… for example a site has a rel-in-reply-to link outside the h-entry… is the h-entry a reply or not?
tantek_ joined the channel
# 14:02 Jihaisse why a site have an rel-in-reply-to outside the h-entry ?
# 14:03 tantek barnabywalters - it's also about keeping the specs simple
_6a68 joined the channel
# 14:03 pfefferle but that is a problem… there are a lot of different markups that people use...
# 14:04 tantek pfefferle, I think you are making it sound harder than it is. So we need to be more explicit.
# 14:04 tantek barnabywalters - there are three levels to this:
scor joined the channel
# 14:04 tantek 2. does the link back have any rel semantics? if so, use them
# 14:05 tantek 3. is there an h-entry on the page? if so, use it for comment/RSVP content
# 14:05 tantek (1) is a required part of webmention - does the source actually have a link to the target
# 14:06 barnabywalters why prioritising rel values over potentially more detailed h-entry properties?
# 14:06 tantek and is required for *anything* that implements webmention, e.g. a webmention proxy
# 14:06 tantek barnabywalters - because a simple rel value is MUCH easier to publish
# 14:06 barnabywalters Jihaisse: you shouldn’t have more than one top-level h-entry on post permalink pages
# 14:06 barnabywalters tantek: understood, but if people *have* published more detailed info, why not try to consume that first
# 14:07 pfefferle tantek yes, but you can't be sure if it has a relation to the first h-entry on the page
# 14:07 tantek Jihaisse: of course it is, e.g. an archive page, but typically it doesn't make sense for those to send webmentions
# 14:07 tantek barnabywalters - that's reasonable. we can treat rel-only markup as a fallback from h-entry markup
# 14:08 tantek pfefferle - it doesn't matter - if the first h-entry on the page *doesn't* relate to that link, then people will debug that quickly
# 14:08 tantek just as barnabywalters quickly debugged the bad double-marked up hentry/h-entry problem recently
# 14:10 tantek pfefferle - now if you're trying to write a webmention *validator* for people to check their posts with - that's a different story - then you can do all kinds of "lint" like warnings etc.
# 14:11 tantek in production it's better to just implement the algorithm we figure out
# 14:11 tantek rather than add bunches of edge cases and stuff
Acidnerd joined the channel
# 14:21 tantek.com edited /in-reply-to (+493) "/* How to consume in-reply-to */ provide more understandable summary, tweak algorithm details to make it simpler / more robust to code, note may use different algorithm if produces same results" (
view diff )
# 14:25 tantek great. I think in general we should provide brief prose summaries that help answer questions like in the discussion we just had here in IRC
# 14:26 tantek also - there's a bunch of potential FAQs up there
ozten joined the channel
# 14:45 pfefferle tantek we had a discussion about using mf2 to markup the wordpress plugins-list, to consume it on other places… do you have an ideas what to use or should i define a posh-format?
# 14:47 tantek also, might want to take a new microformats/poshformat discussion over to #microformats
# 14:48 pfefferle I want to build a wordpress plugin that lists all indieweb plugins around to make it as easy as possible to install them… and barnabywalters had the idea to use the wiki as "source"
# 14:49 tantek pfefferle - do you really want to make it that easy? I mean, won't that just help spammers spam the indieweb?
# 14:49 tantek (since spammers love to install wordpress spam blogs)
# 14:50 pfefferle not yet, the plugins are providing some helpfull stuff but you can't use them without indieweb and mf knowledge yet… and i am not sure if this will ever change...
# 14:51 barnabywalters pfefferle: start with what information do you need in order to generate the plugin installation UI and actually install the plugins?
# 14:52 barnabywalters then make a draft mf2 vocabulary out of that, reusing as appropriate from existing ones
# 14:52 tantek also - the plugins list right now is quite a hodgepodge
# 14:53 tantek plenty of things there you don't need for the indieweb - that list needs some better curation
# 14:53 tantek pfefferle - more importan than a format, is a higher quality list itself
# 14:58 tantek for example, Salmon, all the Syndication plugins
# 14:59 tantek and for Login, all you need is IndieAuth, you don't need separate OpenID or BrowserID/Persona
# 14:59 tantek when I've shown this to other folks who have Wordpress - their response is - wow I have to install all of that?!?
# 15:00 tantek and often it's not clear which plugin does what and why would you want to install it (what indieweb capabilities do you get when you install it)
# 15:00 pfefferle yes, but is that a problem? you said you want to scare spammers ;)
# 15:00 tantek maybe move some to a "Related Social Web" section below?
# 15:01 tantek pfefferle - the point about scaring spammers was a bit of a joke ;)
# 15:01 tantek the irony is that designing for ease of use for a "typical" wordpress user and a spammer is likely to be very similar
# 15:02 tantek a typical wordpress user will likely just want an "indieweb" plugin that "does all the right indieweb things"
# 15:02 pfefferle No, seriously… I like to build plugins to fit only a specific problem and not to solve all IndieWeb "problems"
# 15:02 tantek but at least we can minimize the set of plugins that are *essential* indiweb plugins
# 15:02 tantek and then maybe list optional indieweb ones below that
# 15:03 tantek and then list past/old technology ones (like Salmon, OpenID etc.) below that
# 15:03 tantek pfefferle - that's totally fine - modularity is good
# 15:03 tantek however, we should provide them in at least those priority clusterings for ease of discovery by new wordpress folks
# 15:04 pfefferle barnabywalters noone is using salmon… i built it for my ostatus plugin… long time ago...
andreypopp joined the channel
# 15:08 tantek thanks pfefferle - looking forward to seeing which plugins go in which set
npdoty joined the channel
tantek joined the channel
# 15:15 tantek we've been drawing a (minor) distinction between those who have actually made it to an event
# 15:16 barnabywalters e.g. barry frost isn’t a “community member” per se, but has been doing some awesome work and deserves to be mentioned in implementation listings
# 15:17 pfefferle am i part of the "attendies" if i attended two camps remotely?
# 15:18 tantek the realtime aspect of the participation is key - also it demonstrates a greater commitment to the community
# 15:18 barnabywalters I suppose listing them separately helps give people a sense of how easy they are to engage, how likely they are to respond to questions etc
# 15:18 tantek barnabywalters - people like to get a sense of how involved are people
# 15:18 tantek and indiewebcamp participant vs. community member vs. other independents is a good sorting for that
# 15:20 pfefferle I will add a simple description to every plugin any time soon
bnvk joined the channel
paulcp and _6a681 joined the channel
# 15:32 tantek also I don't know which of those are actually essential - you would know better
# 15:32 tantek if any there should be moved to "optional" please do so!
# 15:32 tantek thanks - that will help a lot - even just a few words summary (to fit on the same line)
# 15:33 tantek and then we can start seeing what other wordpress users think when they see the page
# 15:33 tantek I know there are WordPress users that want to be "more indieweb" but they're not sure how - if this page can help them - that would be great!
# 15:34 tantek pfefferle - openid had its chance as is now a bit of a disaster (from a standards/community perspective)
# 15:34 tantek plus as a *user* of openid, I can now use indieauth as my openid provider
# 15:35 tantek pfefferle - and it's not just me - aaronpk struggled with lots of problems trying to consume OpenIDs for the wiki when we set that up at first
# 15:36 tantek thanks so much pfefferle - this will help a lot
# 15:37 tantek I know a bunch of wordpress users (e.g. at work, Mozilla) and they've been asking what they can/should install - and I honestly didn't know!
# 15:37 tantek who else here is using WordPress on their primary domain?
# 15:39 pfefferle how can i link to someones profile page on the wiki, but using the name instead of the domain?
# 15:46 tantek once you've copied them all over, we can perhaps shorten the list on the "projects" page to just IndieWebCamp participants, since the WordPress page will have the full list
# 15:58 tantek makes it easier for people to one-click go see what their wordpress usage looks like
# 15:59 tantek (and verify that yes that person is using wordpress)
# 16:00 tantek !tell skinny Looks like you haven't signed into IndieWebCamp.com since we switched over to IndieAuth - could you sign-in and create your user page (even with just your name and link to your site) so we can link you up on other pages?
# 16:00 Loqi Ok, I'll tell them that when I see them next
# 16:03 tantek pfefferle - it helps to see a list of people in the broader context of projects
# 16:04 tantek to compare which projects have how much uptake/usage
# 16:04 tantek though I do like xtof's explicit listing of the plugins he uses
# 16:12 pfefferle tantek I think we should copy the research part of christophe ducamp to a "research" section on the wordpress section
# 16:13 tantek yeah I was trying to figure out what to do with that
# 16:13 tantek.com edited /projects (-463) "/* WordPress */ trim this section down to a summary and listing of IndieWebCamp participants using Wordpress since the main WordPress page is much better now!" (
view diff )
earplugs1 joined the channel
# 16:17 tantek ok pfefferle -I'm done touching those pages for a bit - have at it
# 16:22 tantek also - we should add acegiak as a community member in that list
paulcp joined the channel
# 16:30 tantek ok - *now* I'm stepping away from the keyboard ;)
# 16:31 tantek Thanks again pfefferle - I think we've made the /WordPress page *a lot* friendlier and more usable. Now to try it out with more WordPress folks and see if the indieweb plugins get more installation and feedback.
dybskiy, ozten, mikeal, jfranusic, bnvk_, paulcp, caseorganic and earplugs joined the channel
tpinto, tantek, paulcp, caseorganic, jfranusic, bnvk, jfranusi_, melvster, earplugs, friedcell, earplugs1, npdoty, barnabywalters and andreypopp joined the channel
# 21:34 mikeal brianloveswords: i really want a pirate radio station now
mikeal and paulcp joined the channel
mikeal and brianloveswords joined the channel
# 22:00 tantek.com edited /webmention (+154) "/* IndieWeb implementations */ link to sites that accept webmentions to make it easier to check how they're publishing discovery information" (
view diff )
# 22:02 brianloveswords mikeal: it's gonna happen! Or, at the very least, I will *try* to make it happen :D
# 22:11 tantek hmm - either tommorris turned off his webmention discovery, or never had support for webmention in the first place?
ShaneHudson2 joined the channel
# 22:15 Loqi Ok, I'll tell them that when I see them next
# 22:16 tantek.com edited /webmention (-89) "/* IndieWeb implementations */ update a few discovery implementations, can't find any sign of webmention on tommorris.org (home or permalinks) so commenting out" (
view diff )
# 22:19 tantek is there anyway on github to search a particular person's repos?
# 22:21 tantek and you have to specifically click Code to search code and not just repo names
# 22:21 tantek e.g. searching someone's repos for "curl" to see how they use it :)
bnvk joined the channel
# 22:23 tantek I'm thinking of writing a small function to do a head curl and turn any Link: headers into same JSON return format as the canonical microformats2 parsing output
# 22:23 tantek wondering if that would be useful to anyone else - it's how I'm going about building a building block for my webmention discovery code
# 22:24 tantek basically just treating the HEAD like another document with microformats
# 22:24 tantek I don't know if it is useful or not, so I'm going to try building it to find out
# 22:25 tantek feels like a good way to isolate away HEAD-specific logic
# 22:26 aaronpk I feel like my coding tends to be either way more overthought or way less thought through than yours :)
# 22:26 tantek I also like small contained bits of code I feel like I can write quickly, test, and then park.
# 22:27 aaronpk totally. i've been trying to modularize more stuff that way too, like the webmention client and such.
# 22:28 tantek it's also the only way I've found so far to incrementally open source my code in a way that won't cause me huge maintenance/support headaches
earplugs joined the channel
# 22:38 tantek and google is useless for searching for anything "LINK:" :P
# 22:39 pdurbin I feel like I try "link:" once every 3 years and it never works
andreypopp joined the channel
# 22:40 tantek pdurbin I'm actually searching for the literal "Link:" as in what most webmention implementations are returning in the HTTP HEAD
andreypopp, caseorganic, tantek, tilgovi and shaners joined the channel
# 23:31 peat ... the front page image is pretty dark considering the discontinuation message. "It's time to say goodbye."
mikeal joined the channel