2014-09-30 UTC
benwerd joined the channel
wolftune, snarfed and KartikPrabhu joined the channel
# 00:39 tantek alright - let's see how the twitters respond to "vouch"
yakker joined the channel
fmarier and gr0k joined the channel
brianloveswords, scor and Haxxa joined the channel
# 01:07 kylewm KartikPrabhu: twitter's API docs do seem to have changed, is that what you were refering to yesterday?
# 01:10 gRegor` takes back the karma he gave KartikPrabhu ;)
# 01:11 gRegor` GWG: wp_insert_post has been around since 1.0, so I think it's pretty stable and future updates will be made in a backwards-compatible way.
mlncn joined the channel
eburcat joined the channel
# 01:19 GWG gRegor`: I just hate to roll my own interface when there is one already, but considering the new API doesn't yet have taxonomy support, which I need, I may put into place my thoughts on endpoints...
# 01:19 gRegor` I wasn't suggesting your own interface
# 01:19 gRegor` I thought that was the point of micropub; you could use Quill or any other micropub writer
# 01:20 GWG gRegor`: Micropub would be implementing an API
npdoty joined the channel
brianloveswords joined the channel
npdoty joined the channel
# 01:34 kylewm "This site is not associated with Darth Vader"
KevinMarks_ and paulfitz joined the channel
scor, fmarier, j12t, aaronpk_, KevinMarks_ and paulcp joined the channel
erlehmann joined the channel
# 02:47 erlehmann '';!--"<XSS>=&{()}
gr0k joined the channel
# 02:55 gRegor` I had not thought of XSS in the chat logs
# 03:00 erlehmann gRegor` why not?
# 03:00 erlehmann '';!--"<XSS>=&{()}
erlehmann joined the channel
# 03:02 erlehmann gRegor` how do i get an icon in the chatlogs?
fabiok and snarfed joined the channel
shaners and fabiok joined the channel
# 03:51 shaners tantek_ Homesteading is the name of the software that runs my site.
# 03:51 Loqi shaners: tantek left you a message 10 hours, 9 minutes ago: what is the name of the software you use to run your site?
# 03:53 shaners tantek_ tantek___ Who is using just Apache with no cms / generator to serve their site?
paulcp joined the channel
# 03:53 shaners !tell tantek Homesteading is the name of the software that runs my site.
# 03:53 Loqi Ok, I'll tell him that when I see him next
# 03:53 shaners !tell tantek Who is using just Apache with no cms / generator to serve their site?
# 03:53 Loqi Ok, I'll tell him that when I see him next
michiell_ joined the channel
fmarier joined the channel
# 04:03 KartikPrabhu shaners: if Homesteading is running on your website, it counts as a current project I think
# 04:03 shaners Well... it's more of proto-homesteading. (I was still catching up on the logs.)
# 04:04 KartikPrabhu I see fair enough... just to note that my project /Bundle uns only on one website and its even open sourced yet :)
# 04:04 shaners it's the 2.0 rebuild/cleanup from my years of duct tape and bailing wire pile of code that's powering iamshane.com right now.
# 04:05 KartikPrabhu same with Bundle. It isn't organised well enough to even be alpha-open-source :P
GWG joined the channel
# 04:18 shaners KartikPrabhu there's no way that i could open source iamshane.com to anyone in any useful way, but it's important to me to open source software. thus, the rebuild / cleanup.
aaronpk_ joined the channel
joshwnj joined the channel
chrissaad, j12t and eburcat joined the channel
# 05:18 npdoty kylewm, KartikPrabhu - I think the attack is:
# 05:18 npdoty the attacker sends a pingback notification (or webmention notification) to your site
# 05:19 npdoty then your site, to verify the pingback/webmention and avoid spam, checks victim.com to make sure it has an actual link to your site
# 05:19 npdoty it doesn't, but in the process of checking it, you've loaded victim.com. and if enough people do that at once... ddoss
# 05:20 npdoty I think it does apply to the webmention case as well
aaronpk_ joined the channel
# 05:20 npdoty anyone who advertises support webmentions, and does the recommended check to prevent webmention spam, could be used to request/attack another site
# 05:23 npdoty I'm not sure the expiring endpoint makes a difference
# 05:23 kylewm "acting as a auth-free proxy server" sounds a little more nefarious than sending a GET on someone else's behalf
# 05:24 npdoty the attacker can in real time access your site, check the link header to get the webmention endpoint, and then use it
# 05:24 KartikPrabhu but to DDOS the victim the attacker has to do that to as many attacks they're setting up
# 05:25 npdoty right, but I think it only doubles the number of requests they have to make
# 05:26 npdoty the real advantage of the attack is that you're getting the requests to come from so many different IP addresses, not that each machine in the botnet is making lots of requests
# 05:27 npdoty I think in this attack they can use a single IP address. with that IP address, they can contact each of our indieweb servers with webmention endpoints
# 05:27 npdoty each of our servers sees only one request from that IP address
# 05:27 npdoty and then follows the webmention to the victim site
# 05:27 npdoty and the victim site sees the traffic coming from different IP addresses
# 05:27 kylewm what they could do with the expiring endpoint, is gather up 100 sites, and send them each 100 webmentions
snarfed joined the channel
# 05:28 npdoty I don't think the attacker is counting on an amplification of requests, it's just using our sites as a way to diversify the IP addresses
# 05:29 kylewm async processing of webmentions might actually help the attacker
# 05:29 npdoty slowing down the checking of the webmentions might help, but if the attack is large and sustained enough, then having the traffic spread out may not make a difference
# 05:30 KartikPrabhu npdoty: well then it seems there is no way to do real-time distributed notifications then
# 05:30 kylewm KartikPrabhu: if intermediaries == the unsuspecting sites that receive the webmentions, then that's what i mean
# 05:31 npdoty maybe we could treat it as explicit forwarding
# 05:31 npdoty tell the victim site, we're checking whether you made a webmention, and oh, by the way, this is the IP address that told us
# 05:31 npdoty and when the victim sees a gajillion requests all forwarded from the same IP address
# 05:31 npdoty it can start ignoring requests with that header
# 05:33 npdoty the server would still receive the same number of POST requests
# 05:33 npdoty but it would be able to separate out the good from the bad, when it saw that so many of them had the same originating IP address
# 05:33 npdoty such that the attacker would no longer be effectively distributing his single ip address into many IP addresses
# 05:33 npdoty ... at which point, he would use a different kind of attack
# 05:34 npdoty victim.com already has some abuse features in place, rate-limiting by IP address
# 05:34 npdoty and it can rate-limit by X-Forwarded-For IP address too
# 05:35 npdoty you could tell roughly as quickly as you can tell with any single IP address that seems to be DOSing, it just wouldn't be distributed any more
# 05:38 npdoty how did Loqi see that? does it just search for any mention of "webmention"?
# 05:38 kylewm !tell shaners with your Settings table in Homesteading, how do you switch between localhost and production?
# 05:38 Loqi Ok, I'll tell him that when I see him next
j12t and chrissaad joined the channel
glennjones, aaronpk_ and j12t joined the channel
# 05:55 npdoty I think that might really be the effective fix needed
# 05:55 npdoty but then again, maybe it's late and I'm not thinking super clearly and someone has already considered this and ruled it out for some reason
eburcat joined the channel
loic_m joined the channel
# 06:06 npdoty it looks like WordPress actually did implement (in response to a plugin that did it) a separate X-Pingback-Forwarded-For header
# 06:07 npdoty because why use a commonly-used non-standard header when you can use an uncommon variation of a non-standard header? ;)
# 06:08 KevinMarks_ Does Mozilla or chrome publish stats on typing urls versus search terms?
eburcat, loic_m_ and cweiske joined the channel
# 06:30 npdoty and filed an issue on the webmention spec repo
j12t, musigny, elima, aaronpk_, ShaneHudson, alanpearce and petermolnar joined the channel
friedcell and j12t joined the channel
alanpearce joined the channel
# 07:33 voxpelli And now I see it might have already been discussed here, sorry
# 07:40 KartikPrabhu voxpelli: yes it was mentioned by KevinMarks_ would be good to solve this to have better distributed comments
# 07:42 voxpelli My "solution" currently is to severely limit the number of source lookups per minute/hour I do on a host
# 07:43 voxpelli Such measures are probably the only possible ones, at least that I can think of
aaronpk_ and KevinMarks joined the channel
# 07:54 voxpelli KartikPrabhu: Yeah, thanks, should update my endpoint with some more such protection
# 07:57 KartikPrabhu voxpelli: fwiw: i don't think webmention is popular enough yet for these
# 07:59 voxpelli KartikPrabhu: Me neither, but Pingback probably at one time wasn't either ;)
# 08:00 voxpelli DDoS and such is a hard problem in that one would rather not dogfeed it but rather fix it beforehand :P
LauraJ, j12t and loic_m_ joined the channel
aaronpk_, chrissaad, loic_m__, KartikPrabhu, erlehmann, j12t, loic_m_, PierreO, eburcat, ShaneHudson, adonaldson, eschnou, glennjones and KevinMarks__ joined the channel
loic_m_, scor, JohnDuh, erlehmann, aaronpk_, friedcell, eschnou, j12t and PierreO joined the channel
eschnou joined the channel
loic_m__, aaronpk_, musigny, eburcat, loic_m_, Mark87, scor and alanpearce joined the channel
petermolnar, aaronpk_ and alanpearce_ joined the channel
# 12:52 annevk tantek_: tantek___: as well as OWFA something
# 12:52 annevk tantek_: tantek___: the letter per your suggestion
# 12:52 annevk tantek_: tantek___: however, the link for the latter is broken, how can I fix it?
petermolnar joined the channel
# 12:53 annevk tantek_: tantek___: also, I was thinking of making the license more visible, by having a "License:" field similar to "This version:"
j12t and petermolnar joined the channel
petermolnar joined the channel
# 13:03 annevk ben_thatmustbeme: your nick is one character too long for my client for some reason
eburcat and wolftune joined the channel
hmans and chrissaad joined the channel
# 13:18 GWG ben_thatmustbeme: Lots to do, as always.
# 13:20 ben_thatmustbeme yeah, same here. I am up really early now though thanks to my wife having to be up early, so I get to work like an hour early and just work on my site
# 13:21 GWG But, was looking at trying something different for my site when I have time to devote to it
# 13:22 annevk ben_thatmustbeme: your home page has mixed content
# 13:26 GWG ben_thatmustbeme: Since the API I wanted to use isn't finished yet, I was going to go with a simple page template for notes.
erlehmann joined the channel
# 13:32 Loqi ben_thatmustbeme meant to say: there we go
loic_m__, reedstrm, loic_m_, ShaneHudson and cuibonobo joined the channel
# 13:44 cuibonobo do any of y'all know why Google's date range search breaks for indie web camp irc logs?
paulfitz and ShaneHudson joined the channel
# 13:46 cuibonobo if i were to do a search for `webmentions site:indiewebcamp.com/irc` and then under Google's 'Search Tools' adjust the search from 'Any Time' to 'Past year', i get *much* fewer results than expected
musigny joined the channel
# 13:49 cuibonobo i care about this because this channel tends to be a much higher bandwidth than i can hope to keep up with, but i would still like to somehow keep track of topics that interest me
aaronpk_ joined the channel
# 13:50 cr what HTTP header should server add. Last-Modified?
# 13:50 cr donno why so many servers return Date: Tue, 30 Sep 2014 13:49:47 GMT - always
# 13:50 cr "now" time instead of date of the doc last-modification
eburcat, barnabywalters and mlncn joined the channel
# 13:59 ben_thatmustbeme could someone log in to ben.thatmustbe.me with indieauth and tell me if "Recent Drafts" appears on top of the right column?
brianloveswords joined the channel
# 14:01 cweiske ben_thatmustbeme, why does your indieauth client want me to do authorization instead of simple authentication?
# 14:02 cweiske you use response_type=code instead of response_type=id
# 14:08 cweiske you don't normalize the URL I put into the url box
# 14:08 cweiske Given identity URL "http://cweiske.de" and claimed OpenID "http://cweiske.de/" do not match
# 14:08 cweiske err. should I see the login form again after login?
# 14:11 ben_thatmustbeme i added some extra bits to try and see if login would work with open_id at some point
jgarber joined the channel
# 14:18 annevk tantek_: tantek___: the redirect to some Google-operated domain is not reassuring
# 14:18 annevk tantek_: tantek___: I'm inclined to simply drop that
# 14:20 kylewm ben_thatmustbeme: I don't see Recent Drafts when logged in
eburcat joined the channel
Pea1 and snarfed joined the channel
# 14:40 cweiske I can't continue because Given identity URL "http://cweiske.de" and claimed OpenID "http://cweiske.de/" do not match
j12t joined the channel
indie-visitor joined the channel
# 14:44 Loqi Welcome, indie-visitor! Set your nickname by typing /nick yourname
# 14:44 cweiske httpurl = "http://" hostport [ "/" hpath [ "?" search ]]
loic_m__ joined the channel
# 14:50 ben_thatmustbeme cweiske, indeed, but there needs to be some semblance of sanity there, you could trail with //////// and its valid, I'm willing to always have a trailing slash if its really a problem
# 14:52 cweiske indieauth-openid should do the normalization internally and decide that the urls are equivalient
# 14:52 cweiske In general, a URI that uses the generic syntax for authority with an
# 14:52 cweiske empty path should be normalized to a path of "/".
# 14:53 cweiske question is if you should normalize the url you get or not
# 14:53 annevk ben_thatmustbeme: for HTTP URLs, yes
# 14:53 annevk (or implement that to some approximation)
# 14:54 cweiske ok, then your normalization doesn't follow rfc 3986 :)
# 14:57 cweiske "An absolute-path-relative URL must be "/", followed by a path-relative URL that does not start with "/"."
gr0k joined the channel
# 14:59 barnabywalters cr: what tangible benefits does typing the document fragment give to users when they’re using their URL to log in somewhere?
# 15:04 ben_thatmustbeme An absolute URL must be a scheme, followed by ":", followed by either a scheme-relative URL, if scheme is a relative scheme, or scheme data otherwise, optionally followed by "?" and a query.
# 15:04 ben_thatmustbeme A scheme-relative URL must be "//", optionally followed by userinfo and "@", followed by a host, optionally followed by ":" and a port, optionally followed by an absolute-path-relative URL.
snarfed1 joined the channel
ShaneHudson, gRegor` and brianloveswords joined the channel
# 15:09 cr barnabywalters: beacuse you get nameclashes with the doc and things defined in the doc otherwise
# 15:10 cr doesnt matter. if you want to use webID tools, you'll give in and do proper web-architecture
# 15:11 ShaneHudson As anyone got a way of POSSEing G+? I know there are no APIs, but should be a way of doing it
shaners joined the channel
# 15:16 Loqi shaners: kylewm left you a message 9 hours, 37 minutes ago: with your Settings table in Homesteading, how do you switch between localhost and production?
# 15:16 barnabywalters cr: the representative h-card algorithm removes nameclashes, allowing domain -> profile mapping without the need for extra confusing URL syntax
# 15:17 ShaneHudson shaners: That's surprising, I thought Orkut at quite a large Asian userbase
# 15:17 shaners kylewm: how do I switch what between localhost and production?
# 15:21 cr barnabywalters: 'extra confusing URL syntax' = using a doc URI without a fragID for the thing it defines.
# 15:21 ShaneHudson Has anyone had experience with squarespace? I presume it can't be used for indieweb?
danlyke joined the channel
danlyke joined the channel
# 15:22 ben_thatmustbeme when i do that i get a code back, but the call to auth_endpoint returns an empty array
# 15:23 barnabywalters cr: look at the use of URLs in public spaces (e.g. advertising). It’s typically a domain, maybe with a single path segment. Never a fragment identifier
# 15:23 cr we're talking about identifying users now. not websites
# 15:24 barnabywalters actual, real people, who have no idea what a “fragment identifier” is or why they should care about pseudo-philosophical webarch reasons for it’s existance
# 15:25 cweiske ben_thatmust, now I get no errors, but am also not logged into your site
mlncn joined the channel
# 15:25 cr actual real people will get their URI minted for them and not have to know.
# 15:25 ben_thatmustbeme yeah, it does the same to me cweiske, same thing, call to auth endpoint returned nothing
# 15:25 cr you're making a lot of simple, trivial, indie tools harder if you decid you're going to have nameclashes everywhere
# 15:26 cr it's not pseudo-philosophical. once youve loaded your doc, the local identifiers start after the #
# 15:32 gRegor` !tell ShaneHudson I don't have experience with Squarespace, but my understanding is you have full control of the HTML, so you could use it for indieweb. Webmention would be tricky, of course. Unless we encouraged them to add support for it. :)
# 15:32 Loqi Ok, I'll tell them that when I see them next
# 15:32 gRegor` What is Squarespace?
# 15:32 barnabywalters at the very least squarespace gives you enough control to use a squarespace site as your indieauth domain
j12t joined the channel
# 15:38 gRegor` What is webID?
# 15:40 annevk ben_thatmustbeme: if your identity is a domain, you should not take a URL... but then the protocol requires a URL...
# 15:41 ben_thatmustbeme if i give with a trailing slash to indie_auth, when it hits the callback it does not have one
# 15:47 gRegor` cweiske had to run
# 15:47 cr <a class="u-url url fn" href="http://waterpigs.co.uk" rel="me">
# 15:47 cr apparently barnaby is a website?
vanderwal joined the channel
# 15:50 gRegor` The full name of the person is what's inside the link (Barnaby). And his URL is http://waterpigs.co.uk , is what that microformat indicates
# 15:51 gRegor` Yeah, but just guessing
# 15:54 ben_thatmustbeme first off, if i understand things correctly, rel= isn't microformats markup, the microformats there just says its a url (me being pedantic)
# 15:54 ben_thatmustbeme XFN 1.1 introduced the "me" rel value which is used to indicate profile equivalence and for identity-consolidation.
# 15:55 ben_thatmustbeme it says nothing about a person being that thing, just that that link contains and equivalent profile
# 15:56 gRegor` I was only parsing the mf, not mentioning the rel
# 16:03 ben_thatmustbeme !tell cweiske try logging in again to my site. I have some instrumentation now and have fixed an few small errors. might help
# 16:03 Loqi Ok, I'll tell them that when I see them next
tantek joined the channel
# 16:04 gRegor` Yeah, I'm not sure if cr was raising a problem with that h-card or?
aaronpk_, aaronpk__ and tantek_otp joined the channel
# 16:09 tantek_otp I read the logs and I really have no idea what cr was talking about in 90% of the statements.
aaronpk_ joined the channel
# 16:09 Loqi tantek_otp: shaners left you a message 12 hours, 15 minutes ago: Homesteading is the name of the software that runs my site.
# 16:09 Loqi tantek_otp: shaners left you a message 12 hours, 15 minutes ago: Who is using just Apache with no cms / generator to serve their site?
# 16:09 tantek_otp shaners, check out /file-storage - a few folks there "just" serving static files via apache
# 16:09 tantek_otp otp = on the phone
# 16:10 tantek_otp pre-socialwg prep call :/
fmarier joined the channel
loic_m_, mlncn and squeakytoy joined the channel
davidmead and reedstrm joined the channel
j12t joined the channel
# 16:53 cr as soon as one wants to cache, annotate, store i a graph-store, how do you know whether you mean https://waterpigs.co.uk/ the person, the web-document, or the website? by dismissing good URI-design as "complicated" you're inadvertently making things complicated
jet___ joined the channel
# 16:55 gRegor` cr: What URI design are you referring to?
# 16:55 cr the rel=me reusing a document/website URI for a person
# 16:55 cr how do i disambiguate? i'll rather just not consume microformats data until they stop this silly nameclash
# 16:56 cr the alternative is complicate implementations by using tons of named-graphs and annotating the named-graphs with contexts, so you can reuse the URI fore different things. but that's meh. too complicated..
tilgovi, gavinc and paulcp joined the channel
# 16:59 gRegor` cr: What is the use-case where you need to disambiguate something? I'm not understanding.
snarfed joined the channel
# 17:00 gRegor` My personal website represents me online. Why wouldn't I link to it with rel=me?
# 17:00 ben_thatmustbeme so instead of a user Ben and a link labelled "twitter profile" it becomes a link title "Ben" as well
npdoty joined the channel
# 17:01 gRegor` ben_thatmustbeme: I'm still not understanding. :)
# 17:02 Loqi Ok, I'll tell them that when I see them next
# 17:03 tantek cr, I think you're talking yourself into making it too complicated.
ShaneHudson, JohnDuh and JohnDuh_ joined the channel
# 17:20 davidmead when using POSSE for a reply, does it have to be a new post, or can it be a reply on the originating post?
# 17:21 davidmead i tried using "u-in-reply-to" but it didn’t post back to twitter
# 17:23 ben_thatmustbeme davidmead, its a new post, it has to be the h-entry itself with a u-in-reply-to (if i understand what you mean)
# 17:23 KevinMarks_ URIs are just notation conventions used for databases, but URLs locate resources universally, and are thus useful for humans and computers alike
# 17:25 ben_thatmustbeme it makes sense though, honestly, every reply could be replied to. if the only location was in the reply-thread of a post, I would have no way to reply to your comment specifically
# 17:26 davidmead ben_thatmustbeme As the originating post is sent to Twitter, brid.gy (or similar) pulls back the replying tweet and displays as a reply to that post, I’d like to reply to that reply in the reply thread
mlncn joined the channel
# 17:27 ben_thatmustbeme you could reply-to both, that way its processed as a webmention internally to yourself and bridgy would still pick it up as a reply to the other reply
# 17:27 KevinMarks_ I think cr should stop using English, which is semiotically impure, and communicate only with people who speak lojban
# 17:28 sparverius side benefits include: not being invited to meetings, social events, generally having to interact with people in real life
# 17:30 davidmead ben_thatmustbeme I guess. it seems to me that it breaks the conversational thread between the parties
# 17:30 ben_thatmustbeme i don't think so, you just are forking converstaions between parties into their own thread
Loqi joined the channel
# 17:31 gRegor` Each party can show the full context-thread and reply-thread, though
# 17:31 gRegor` What is context-thread?
# 17:32 gRegor` Hm. That's not what I think of when I say context-thread. :)
# 17:32 gRegor` I think of it as everything coming before the current note in the thread.
# 17:33 gRegor` And reply-chain I think of as everything coming after the current note, in the thread.
aaronpk_ joined the channel
# 17:33 gRegor` wonders about 'thread' vs 'chain'
# 17:34 davidmead ben_thatmustbeme grego` so maybe i’m still visualizing this “wrong” on my blog and need to think of a reply from Twitter like a new post on my site. my reply to it is a new post...
# 17:34 gRegor` Ah, I see 'reply-chain' got renamed 'reply-thread'. Sounds good.
# 17:34 davidmead the whole blog is not a post with replies attatched to it, but a series of linked replies to each other
mlncn joined the channel
# 17:38 gRegor` That page is causing confusion or?
# 17:39 ben_thatmustbeme my site is currently the only example of a context thread, but it got listed under reply-thread
mlncn-agaric joined the channel
# 17:40 gRegor` You're listed on both because you're displaying both
pauloppenheim joined the channel
# 17:41 gRegor` and the full contents of those reply-contexts, too.
# 17:41 gRegor` Since "in reply to [link]" is also a reply-context, many of us have a minimal reply-context
# 17:42 gRegor` So we need to get more people to list themselves on reply-thread, sounds like.
# 17:43 davidmead ben_thatmustbeme gregor` so the more i look at this the more it seems i have to look at something other than WP
benjamin-agaric joined the channel
# 17:43 gRegor` Gah. Ok, this *is* confusing to me:
# 17:43 gRegor` "A reply thread (AKA reply chain) is a threaded list of replies, and replies to those replies, displayed under the original post, sometimes as part of a reply-context."
# 17:43 ben_thatmustbeme davidmead, WordPress has a lot going for it, but for indieweb things its a lot of patchwork addons
# 17:43 gRegor` "sometimes as part of a reply-context" doesn't make sense to me there.
# 17:44 gRegor` But maybe it's just my not understanding the intent of those terms.
wolftune joined the channel
# 17:44 davidmead ben_thatmustbeme that’s what i’m finding - can’t use some of the tools like ownyorgram as WP doesn’t support micropub
# 17:45 davidmead ben_thatmustbeme i like Known but I don’t know if I can import my content into that as a direct replacement
# 17:46 gRegor` Yeah, I'm not aware of any WordPress -> Known import tools yet
# 17:47 gRegor` So is reply-context a generic name for a set of replies, that appears either earlier in the conversation, or later in the conversation?
# 17:47 gRegor` And we "need" more specific terms for the "earlier" and "later"?
# 17:48 gRegor` "context-thread" being the earlier, and "reply-thread" being the later?
# 17:49 gRegor` I'm just not sure more terms are necessary
# 17:49 davidmead "context-thread" should be the whole thing. "reply-thread" is after the post. you need another term for before the post - the thing(s) that prompted it
# 17:50 davidmead the term “reply” doesn’t make sense to me as descripbing something that came before
# 17:50 ben_thatmustbeme context-thread i would not put as the whole thing, so we really need any term for the whole thing?
# 17:51 gRegor` davidmead: It is a bit confusing, until you thiink of it as "someone else's reply"
# 17:51 ben_thatmustbeme it does, because my post is a reply, i want to my my reply in context, so i see the argument
# 17:52 davidmead exactly KevinMarks_ to me context is viewing the whole thing
# 17:53 ben_thatmustbeme the parts of a written or spoken statement that precede or follow a specific word or passage, usually influencing its meaning or effect: I agree with davidmead & KevinMarks_
# 17:53 KevinMarks_ The challenge is deciding which path through the chain of replies is the context
# 17:53 gRegor` I kinda like history-thread
# 17:54 KevinMarks_ We see this on twitter all the time, as each of us gets a different subset of the responses
# 17:54 gRegor` history-thread and reply-thread make a lot of sense to me; they're pretty self-explanatory
# 17:54 KevinMarks_ There isn't a globally visible thread, that's part of the value of twitter and the open Web alike
# 17:54 davidmead so a “context-thread” is made up of “historical-thread” “post” “reply-thread”?
# 17:55 ben_thatmustbeme i've been trying to work that bit out on my site, eventually i want to pull in all replys to my posts, even if they only reply to a reply of mine that has one of my posts in its thread-history
# 17:57 gRegor` davidmead: That sounds like the understanding of context-thread based on this discussion, yes.
# 17:57 gRegor` I'm not sure I would use the term context-thread myself, though. I don't see a lot of need for it.
# 17:57 gRegor` historical-thread instead of history-thread?
# 17:58 davidmead gregor` cool - so you can only have a “historical-thread” on your “post” if it is a reply to something else (webmention)
# 17:58 ben_thatmustbeme i think reply-context is best, but without the h-entry being a reply, it doesn't really make sense
# 17:59 gRegor` "A historical-thread is a reply-context that happened before the current post."
# 18:00 gRegor` "A reply-thread is a reply-context (eww, not liking this) that happened after the current post."
npdoty joined the channel
# 18:02 gRegor` "A history-thread is a type of reply-context that appears before the current post and displays a thread of what the post is in-reply-to"
# 18:02 gRegor` "A reply-thread is a type of reply-context that appears after the current post and displays a thread of replies to the post."
# 18:03 gRegor` That was not my thinking, no.
# 18:04 ben_thatmustbeme i sort of like that definition though, we know what people have been doing up til now, showing reply-context
# 18:06 teknotus I haven't forgotten. It's just tedius.
grantmacken joined the channel
# 18:06 gRegor` I guess /reply-context does currently imply a single item, "Popular silos (e.g. Twitter) display an entire thread of reply-contexts . . ."
# 18:07 teknotus Opps repeated previous message.
# 18:07 gRegor` Sure, makes sense
# 18:07 gRegor` In my attempts above, "A type of reply-context" could be "A series of reply-contexts" then
# 18:08 gRegor` Are we sticking with reply-context being a general term, regardless of where it appears in the chronology, though?
# 18:09 gRegor` reply-thread *could* be "a series of reply-contexts that appears after the current post..."
# 18:09 ben_thatmustbeme i think reply context only refers to backwards, thus reply context is the inverse of reply
# 18:09 gRegor` Hm. This is hurting my brain. :)
# 18:10 gRegor` I thought we kinda agreed earlier that reply-context could be before or after
# 18:10 gRegor` If it's only backwards, history-thread is kind of redundant, no?
# 18:10 gRegor` Other than being a term for multiple reply-contexts, I guess.
benjamin-agaric joined the channel
# 18:11 gRegor` Alright, work calls. I will check back later.
# 18:12 davidmead if you’re replying as a webmention about an post elswhere
ShaneHudson joined the channel
# 18:13 davidmead and how do we know how much of a “history-thread” to pull back and display?
# 18:14 davidmead just the post you’re replying to, or the post that that’s replying too to
# 18:14 ben_thatmustbeme reply-context is 1 level back, it could be multiple entries if you have multiple in-reply-to values
# 18:17 davidmead your post is the orginating object of that thread on your site
# 18:18 davidmead maybe if you adopt the ‘long now’ POV, it’s all a historical thread
indie-visitor joined the channel
# 18:28 anomalily ahhh nickser
eschnou joined the channel
# 18:36 ben_thatmustbeme well i stubbed them out as i understand them, there are probably more places where the confusion exists. but its a start
# 18:39 gRegor` For kicks, I might try setting up PESOS from Ello
# 18:40 gRegor` It would be fragile, though, just scraping the HTML
elima joined the channel
# 18:40 davidmead ben_thatmustbeme would I just add thoughts in that stub? or create a new one?
# 18:42 gRegor` There's a json file for each post, but it's only available if authenticated. Ah well, I probably won't bother. :)
gr0k joined the channel
# 18:43 kylewm gRegor`: Ello gives you access to json for each post?
# 18:44 gRegor` The URL indicates it is an API, actually, although private :)
# 18:44 gRegor` <a class='postbar__link postbar__edit' data-dispatch='editPost' href='/api/v1/posts/821143.json' title='Edit Post'>
# 18:45 kylewm interesting, they must be using that to do some ajaxy editor thing
Mark87 joined the channel
# 18:47 gRegor` I couludn't expose anything else poking around various /api/v1/ URLs
# 18:50 gRegor` Not sure what you mean, KevinMarks_
# 18:51 gRegor` 404, yobj. But I don't think I've seen it, no.
# 18:52 yobj you're getting a 404, gRegor`? That's odd.
# 18:52 gRegor` I don't know about demand. Mostly just curious how hard it would be.
# 18:52 gRegor` Also wondering, since Ello is new, if they could be more easily convinced to add some indieweb support
# 18:53 gRegor` Haha, yeah, ello is pretty wonky when clicking back.
# 18:53 gRegor` My homepage URL is inexplicably messed up now, too: "p class='profile__links'>gRegorLove.com"
# 18:54 gRegor` Ooh, nice, yobj.
# 18:55 gRegor` yobj: Did the gist URL get truncated?
caseorganic joined the channel
# 19:01 gRegor` Thanks, yobj. Added it to the /Ello page
cweiske and musigny joined the channel
# 19:06 cweiske ben_thatmustbeme, could you check why I don't get logged in?
# 19:06 Loqi cweiske: ben_thatmustbeme left you a message 3 hours, 3 minutes ago: try logging in again to my site. I have some instrumentation now and have fixed an few small errors. might help
eburcat, grantmacken, chrissaad, adactio and KevinMarks_ joined the channel
# 19:42 ben_thatmustbeme cweiske, everything looks good on my end, state matches up, your auth_endpoint is just returning "Validating token failed"
# 19:44 cweiske ben_thatmustbeme, do you request the second time with id or code?
musigny joined the channel
tantek, loic_m__ and erlehmann joined the channel
# 19:59 cweiske issue fixed. was a problem on my side with indieauth-openid
gr0k and petermolnar joined the channel
scor and wolftune joined the channel
KevinMarks_ joined the channel
# 20:19 cweiske is there someone else apart from barbabywalters and aaronpk who implements indieauth login on his website without relying on indieauth.com?
# 20:21 bear as long as the potential for others to do the same, I would say no
# 20:22 kylewm cweiske: tantek uses a differnet implementation of RelMeAuth
# 20:22 Loqi bear meant to say: as long as the potential for others to do the same exists, I would say no
# 20:24 cweiske "Kevin Beynon got IndieAuth login to his own site working!"
# 20:25 kylewm so do you think one of us needs to build a second implementation? to test the specification?
# 20:27 cweiske ben_thatmustbeme and barnabywalters being the exceptiosn
# 20:27 reedstrm so, we're talking the server auth side? Haven't gotten that far, sorry :-(
# 20:28 kylewm oh I see, ben_thatmustbeme uses your authorization_endpoint if you provide one
caseorganic, jet___ and adactio joined the channel
tantek joined the channel
krendil, tantek__ and eburcat joined the channel
# 20:37 ben_thatmustbeme btw, it is actually on my todo list to roll my own auth endpoint, just to break everyone's application who is assuming indieauth.com :)
# 20:38 tantek ben_thatmustbeme++ for making the cutting edge bleed :)
# 20:40 tantek !tell aaronpk all of tantek_ and tantek___ and tantek_otp are leftover connections from ?beta Join - server might need rebooting to let go of those connections.
# 20:40 Loqi Ok, I'll tell them that when I see them next
# 20:40 tantek I mean, let's not inflate our channel stats ;)
# 20:41 Loqi Ok, I'll tell them that when I see them next
# 20:41 tantek !tell annevk I'm ok with a "License(s):" header section at the top similar to "This version:" that just used the same language. If you want to change the language, make a specific request and I can follow-up with Mozlegal.
# 20:41 Loqi Ok, I'll tell them that when I see them next
# 20:48 cweiske aaronpk's plan is that indieauth.com will once redirect to your auth provider
# 20:49 tantek cweiske thanks for continuing to push on this (actual federation of indieauth).
# 20:49 ben_thatmustbeme yeah, still, it centralizes everything at indieauth. If the site goes down, no one can login
# 20:50 tantek cweiske++ for living on the indieauth bleeding edge.
# 20:51 ben_thatmustbeme i need somewhere new to push the bleeding edge of things, once i get through a few minor bits
# 20:55 Loqi Ok, I'll tell them that when I see them next
mlncn and loic_m_ joined the channel
# 21:19 ben_thatmustbeme !tell KevinMarks I have my MP client set to send both syndicate-to= A, B and syndicate-to[]=A&syndicate-to[]=B so it should work for either implementation now.
# 21:19 Loqi Ok, I'll tell them that when I see them next
# 21:19 cuibonobo ben_thatmustbeme: just tried logging in to your site but nothing happened
mlncn-agaric joined the channel
tantek joined the channel
# 21:22 tantek hey cuibonobo did anyone answer your question about google searching the irc logs in date ranges?
ShaneHudson joined the channel
# 21:22 tantek is still catching up on logs and various confusions
# 21:23 cuibonobo ben_thatmustbeme: i've tried just now with a bare domain and with https, but still nothing
# 21:23 tantek I'm confused about the reply-context confusion
# 21:23 cuibonobo tantek: nope. but then again, figuring out why google does anything is always a mystery
# 21:23 ben_thatmustbeme cuibonobo, what URL? your site listed in irc-people doesn't have an indie-auth endpoint
# 21:24 tantek cuibonobo: FWIW I've been having trouble with google date-range search of my own site - they stopped indexing "everything" :(
# 21:25 cuibonobo ben_thatmustbeme: ah! i've never had to list the endpoint. i assume everybody else just falls back to indieauth.com if none is listed
# 21:28 cuibonobo tantek: ugh. that sucks. have you considered doing search via javascript or something? seems like you can't depend on outsourcing to Google anymore
# 21:29 tantek yeah, it's quite annoying. I'm even using *their* PuSH hub which means they should have results for me within seconds
# 21:30 tantek almost caught up with logs - then just need to review all the *-context and *-thread hubbub
# 21:32 tantek just as I was trying to get to the bottom of the reply-* confusion!
# 21:37 cuibonobo regarding how Google even gets your page's date to begin with, i thought it would be related to Last-Modified headers but nay! the search bot just does a best guess at scraping the contents of your page. :s
# 21:38 cuibonobo well it's certainly a criticism if you choose to depend on them for searching your site
npdoty joined the channel
mlncn-agaric joined the channel
# 21:43 Loqi A reply context is the display of what a reply post is in reply to, including linking to that original post with in-reply-to markup, showing some amount of that original post like author name, icon, summary / ellipsed content, and datetime published http://indiewebcamp.com/reply-context
# 21:43 tantek FWIW I dislike "history-thread" because "history" implies more comprehensiveness than what is actually displayed.
# 21:44 tantek history implies *all* replies to the original post that were made *before* the current reply post are shown, which AFAIK no one does
# 21:44 tantek everyone that implements showing *previous* reply-contexts (both silos and indieweb) *only* recursively show the thread back to the original post - no other replies
# 21:47 tantek !tell ben_thatmustbeme I'm not going to edit your changes to reply-thread / history-thread without first discussing with you and seeing if we can figure out what was confusing and debate which terms better reflect what is going. I updated /reply-context dfn/summary a bit - please take a look.
# 21:47 Loqi Ok, I'll tell them that when I see them next
fmarier and gr0k joined the channel
# 21:50 emmak tantek: as a user, do you think a full reply history would be more useful than reply context?
# 21:50 Loqi KevinMarks__: ben_thatmustbeme left you a message 31 minutes ago: I have my MP client set to send both syndicate-to= A, B and syndicate-to[]=A&syndicate-to[]=B so it should work for either implementation now.
# 21:51 tantek emmak - in the example of Twitter, it's nice to be able to scroll to the top and see what tweet started the whole thing
# 21:52 tantek then if I want to see the whole reply *history*, I click on the permalink of the tweet that started the discussion, which then will display all the threads spawned
# 21:52 tantek I should say, it will display all the *comment* threads
squeakytoy2 joined the channel
# 21:54 gRegor` tantek: To make sure I'm understanding, "history" implies including replies tot he original post from third parties, that aren't in a direct chain with you?
# 21:55 tantek the stuff *above* a *reply* post is a reply-context
# 21:55 KevinMarks__ ah, so tantek is using "history" the way we were using "context" earlier
# 21:55 gRegor` I think some confusion (for me) is your use of "reply-contexts" and "reply-thread". "everyone that implements showing *previous* reply-contexts (both silos and indieweb) *only* recursively show the thread back to the original post - no other replies. hence reply-thread"
# 21:56 gRegor` reply-thread to me is all the replies below a reply post
# 21:56 tantek no need to call them something special reply-*
# 21:56 KevinMarks__ we were trying to distinguish previous posts ("history") from subsequent posts( ("replies")
# 21:57 tantek this is what I get for being behind on distinguishing (and separating) reply vs comment :/
# 21:57 tantek reply is a reply post on its own permalink - the original
# 21:57 tantek a *comment* is a reply syndicated into the context of the thing it is in reply to
# 21:57 gRegor` What is reply-thread?
# 21:58 gRegor` ^ thoroughly confused now :)
# 21:58 tantek so I can either fix that - or wait for ben_thatmustbe to debate it with him
# 21:58 annevk tantek: I moved the license to the bottom, it's per the recommendation for MediaWiki I think
# 21:58 gRegor` Calling them "comments" makes sense and I'm getting a better idea the terminology, though. Not sure how I forgot something so simple.
# 21:58 Loqi annevk: tantek left you a message 1 hour, 16 minutes ago: I'm ok with a "License(s):" header section at the top similar to "This version:" that just used the same language. If you want to change the language, make a specific request and I can follow-up with Mozlegal.
# 21:59 tantek annevk - ok. I don't know what you mean "per the recommendation for MediaWiki" but I'll take a look
# 21:59 tantek gRegor`: right, they've always been called comments
# 22:00 tantek the only *new* notion that's being introduced is a "reply" - which is a post on its own permalink
# 22:00 tantek that happens to be in reply to some other post
# 22:00 tantek and then from there, reply-context, and multiple reply-contexts chained up to the original make a reply-thread
# 22:00 tantek but I can see how "thread" is confusing as it can imply the entire chain(s) up and down
# 22:01 tantek goes to re-read the text he wrote on the Mozilla wiki.
# 22:01 tantek annevk - *nothing* in that text says "bottom" or specifies *location*
# 22:02 gRegor` But that redirect doesn't work now and I get a 404
# 22:02 annevk tantek: I meant "Per CC0" vs <img>
# 22:02 tantek annevk - the Per CC0 vs <img> is a default MediaWiki configuration limitation
# 22:03 annevk I didn't want <img> since a) the CC images are not licensed under CC0 and b) their SVG images are low quality
barnabywalters joined the channel
# 22:03 tantek that that's ok - whether for HTML or MediaWiki
# 22:04 gRegor` Ok, I think I've found an example to demonstrate some of my confusion, tantek.
# 22:04 tantek you got the <indie-action> web component stuff working in shrewdness right?
# 22:04 barnabywalters currently running a batch job over shrewdness. wow there are a lot of posts in there now
# 22:04 tantek because I couldn't reproduce how to make it work
# 22:05 barnabywalters tantek: actually no I didn’t get that working, mainly because that would require adding the indie-config web component code to shrewdness and it’s not very polished at the mo (if it’s even published)
# 22:05 gRegor` Including one's own replies-to-replies :)
# 22:07 tantek gRegor`: your confusion is over-applying what reply-thread means which maybe means reply-thread is a bad term
# 22:07 tantek they're not showing replies underneath, they're showing *comments* underneat
# 22:07 Loqi tantek meant to say: they're not showing replies underneathh, they're showing *comments* underneath
# 22:08 barnabywalters okay, shrewdness updated to show <indie-action> elements next to posts if you don’t have micropub enabled
# 22:08 gRegor` Ok, but it's more than just direct comments, right? I mean, it's Lilly's comment, then Shane's comment to her, then Kartik's comment to Shane, etc.
# 22:08 tantek barnabywalters: but to make <indie-action> elements work you need to do all the web components stuff that voxpelli figured out
# 22:08 barnabywalters also at least partially fixed the URL specificity which was causing multiple http+https versions of posts
# 22:08 gRegor` I don't know if we need a term for that concept or not, but it strikes me as different than typical comments on a blog.
# 22:08 tantek is voxpelli still around? because so far he may be the only who got <indie-action> web components *working* on his site
# 22:09 tantek barnabywalters: of course they do :) the point was to get more folks trying what voxpelli got working
# 22:09 tantek so we can see if a web component solution actually scales across different implementations etc.
# 22:09 tantek gRegor`: the stuff below a post is just comments. not "different than typical comments on a blog".
# 22:10 gRegor` I'm not explaining my thoughts well.
# 22:10 kylewm the indie-config stuff was pretty non-trivial
# 22:10 barnabywalters !tell voxpelli did you publish your indie-action webcomponent code anywhere? I’d love to add it to shrewdness for people who don’t have micropub enabled!
# 22:10 Loqi Ok, I'll tell them that when I see them next
# 22:11 barnabywalters kylewm: on your own site it’s actually pretty straightforward. Once I’ve got it enabled in shrewdness I’ll write a how-to
# 22:11 tantek kylewm: yes - we need to improve the documentation there, but that starts with getting a second implementation
# 22:12 tantek gRegor`: right, twitter does not show indentation for nested comments
# 22:13 kylewm it is commented out because right now it loads eagerly; if I remember correctly it’s because it did not use the web-components library
# 22:14 kylewm (registering the actual web+action handler on the other hand *was* straight forward)
rknLA joined the channel
# 22:16 tantek right, there's the <indie-action> half, and the registering the web+action: handler half
# 22:17 tantek the web+action: half others have gotten to work, e.g. aaronpk
# 22:17 tantek but the <indie-action> web component / iframe etc. half no one else has gotten working
# 22:17 cuibonobo i think i get what gRegor` is saying. calling the entire tweet thread "comments" implies that all these tweets are commenting on the tweet at the top, but they're not.
gr0k joined the channel
# 22:18 adactio There's actually one other little thing we should be doing on the JavaScript side for <indie-action> and that's registering it as a new element (different to registering an action/handler):
# 22:19 tantek cuibonobo - but that's what people on reddit and slashdot call the full set
# 22:19 kylewm adactio: *anybody* who has <indie-action> tags should be doing that?
# 22:19 tantek all of the (nested) comments are thought of as "comments" on the original post in some way
# 22:19 tantek adactio - do you understand how voxpelli got <indie-action> web components working?
# 22:19 adactio kylem: Technically, yes (for it to qualify as a custom element). In practice, I'm not sure how much difference it makes.
# 22:20 tantek because AFAIK no one else does (besides voxp )
# 22:21 gRegor` Yeah, that is what I meant, cuibonobo. And it may not need any term other than "comments", sure.
# 22:23 tantek !tell voxpelli could you provide the URL where you have (had?) <indie-action> markup plus web componenets support e.g. with reply and pay actions?
# 22:23 Loqi Ok, I'll tell them that when I see them next
# 22:23 kylewm tantek: it’s there on the post permalink pages
# 22:23 tantek gRegor`: from this discussion I'm convinced that the term "thread" is already overloaded and is thus the source of confusion
# 22:24 gRegor` Definitely talk to ben_thatmustbeme about this, though. I think he has a better handle on it since he's actually implemented it.
# 22:24 tantek kylewm: good find! looks like he is using <script type="text/javascript" src="/js/x-tag-components.js"></script>
# 22:25 gRegor` For me, I think the -thread suffix makes it clearer that it's one or more of that thing.
# 22:25 kylewm tantek: yeah that bit was the part I was not willing to dive into when I tried to implement
# 22:25 tantek kylewm - as the site *registering* the web+action: handler - you shouldn't have to
# 22:25 tantek that x-tag stuff is only needed for those that *publish* <indie-action> elements
# 22:26 tantek hmm - this might be enough to reverse engineer and get it working
# 22:26 tantek but I think it would work better side-by-side rather than over IRC
# 22:27 tantek hmm perhaps I can talk benwerd into adding registering of web+action handling to Known
# 22:29 tantek i.e. add an "IndieWeb Examples" section and document your implementation
# 22:30 tantek and we can document voxpelli's example page(s) with <indie-action> too
# 22:30 tantek gRegor`: let's drop *thread* it's too often confusing
# 22:31 tantek and gets even more confusing when combined with reply- (my fault)
# 22:32 tantek so for now I'm going to start referring to showing more than one reply-context as showing reply-contexts recursively
# 22:32 gRegor` So go back to calling it reply-chain?
# 22:32 tantek I think that's only slightly less confusing that reply-thread
# 22:32 tantek chain can also imply bidirectionality to some
# 22:33 tantek I'm saying none of those are good (all are confusing and/or inaccurate/imprecise). reply-thread reply-chain history-thread history-chain
# 22:33 tantek whereas "recursive reply-contexts" is quite specific
# 22:34 tantek assuming that reasoning makes sense to you, let's wait for ben_thatmustbe to get back and see if it makes sense with him
# 22:35 gRegor` recursive reply-contexts sounds good for reply-contexts going back beyond the in-reply-to, sure.
# 22:36 gRegor` What is recursion?
# 22:43 tantek in other news - no other comments / follow-ups re: "vouch" webmention protocol extension - does that mean everyone's convinced or just waiting for user-flow and protocol-flow diagrams? ;)
# 22:47 tantek bret - that would be great. also ok to lazy evaluate and await /vouch to be filled out with more readable prose description that separate statements in IRC
# 22:47 Loqi tantek meant to say: bret - that would be great. also ok to lazy evaluate and await /vouch to be filled out with more readable prose description than separate statements in IRC
squeakytoy2 joined the channel
# 22:52 gRegor` KevinMarks: Is feed.unmung.com open source?
# 22:52 gRegor` I just tried it with the RSS feed I had to wrestle with a couple weeks ago and it worked pretty well.
# 22:53 gRegor` It didn't capture multiple dc:creator elements, but otherwise good.
friedcell joined the channel
# 22:54 KevinMarks__ examples welcomed - there is a lot of variation in feeds and I need to add moe to the template to expose them
# 22:57 gRegor` Posted an issue