2015-12-08 UTC
shiflett, funwhilelost, yakker and snarfed joined the channel
# 00:36 aaronpk 201 is an HTTP status code used to indicate that a resource was created in response to the HTTP request.
tommorris joined the channel
# 00:37 aaronpk 202 is an HTTP status code used to indicate that the request was received but processing is not yet completed.
# 00:40 gRegorLove I thought in the last conversation I participated in, we were discussing returning a Status: header
# 00:57 aaronparecki.com edited /Webmention (+156) "/* Verification */ verification is a MUST if the receiver is going to use the webmention, allowing receivers to avoid verification if not needed/wanted (#12)" (
view diff )
DePhraser joined the channel
mxuribe and snarfed joined the channel
davidmead and funwhilelost joined the channel
# 02:16 bear does anyone have a set of curl commands that I can use to test if my indieauth flow is correct ?
[snarfed] joined the channel
# 02:22 [snarfed] yeah that's not easy right now. kylewm would be interested too.
# 02:23 bear i'm trying to verify if the auth flow for indie-stats is working
# 02:31 davidmead hi snarfed. did you mention an issue with bridgy & instagram mentions least week?
# 02:37 [snarfed] hey davidmead! not that i know of. are you seeing something wrong?
# 02:38 davidmead likes and comments from my instagram photos are showing up on my blog…
# 02:39 davidmead now i noticed the syndication link wasn’t getting added to the post, so i added them manually and got bridgy to crawl again, but nothing
# 02:41 [snarfed] does your bridgy user page say it's found the new syndication links?
# 02:45 davidmead not sure if some of the indieweb plugins may be conflicting
# 02:48 [snarfed] does it say it's found a syndication link recently in the "Last polled" section?
# 02:50 davidmead it have by main url above that, next to the instagram url
# 02:51 davidmead i did include the webmention urls in the <head> of my site
# 02:52 GWG snarfed: Saw your issue. Easy fix.
# 02:52 [snarfed] ah, that's it. add your WP URL to your Instagram profile and sign up for bridgy again
# 02:54 GWG snarfed, there is another issue that will affect the request for webmentions
# 02:54 GWG The plugin honors the Allow Pings flag in WordPress
# 02:55 GWG As of 4.3, it is off by default for pages
# 02:55 GWG But you can manually enable by page
# 02:57 davidmead gwg so i have to manually allow pings page by page in WP too?
# 02:57 GWG davidmead: Pages as opposed to posts, yes. But I think there is a filter to enable them by default for new pages.
# 02:58 davidmead oh. i don’t have any pages set up in WP. my homepage is a basic HTML one, not the WP one.
# 02:59 bear [snarfed] - I just put a comment in your issue that shows how you can get to the raw data right now
# 02:59 bear until I make this thing have a proper REST api
# 03:01 bear indie-stats.com and it's large collection of historical html and mf2 data
# 03:04 davidmead snarfed - i changed the URL on instagram and tried again, but nothing…
# 03:05 davidmead just disabled the ‘Social plugin’ option to pull in comments and everything on bridgy lit up and it polling again
# 03:11 [snarfed] davidmead: now that bridgy has your wp url, try clicking crawl now
# 03:12 davidmead snarfed - i think its an issue with MailChimp’s ‘Social’ WP plugin
# 03:14 davidmead i think the only comments i got back from twitter was when i broadcast with their ‘reply’ field filled in
# 03:15 davidmead my bridgy page for twitter is also pointing to my homepage and not the WP page
# 03:15 [snarfed] right, then bridgy wouldn't send it twitter responses either
# 03:18 bear kylewm - let me look at that - getting an access token does require auth, so that is at least 90% of what I need to test
shiflett joined the channel
# 03:29 [snarfed] davidmead: looks like you're getting Instagram responses!
# 03:30 davidmead i am now snarfed. i disabled the Social comment feature and they started coming in :)
# 03:31 davidmead but twitter responses still not coming in snarfed. even after changing url on twitter
# 03:35 [snarfed] did you sign up for bridgy again with twitter? and click crawl now?
wholewheattoast joined the channel
# 03:36 davidmead i’m going to go through and make sure there’s a syndication link on those posts too
yakker joined the channel
arlen joined the channel
# 03:44 [snarfed] huh, the crawl now button isn't crawling on your twitter account. I'll investigate
Gold, KartikPrabhu, zero-gravitas, glennjones, sparverius, snarfed and funwhilelost joined the channel
# 05:07 snarfed the goal is to know explicitly which parser it's using, since they have different behavior...?
cweiske joined the channel
zero-gravitas joined the channel
tantek joined the channel
# 06:12 tantek 8 of ~16 person-tags made it through to the FB POSSE copy
# 06:28 snarfed it'd take some work, and no guarantee they'd grant us the permission, but it's doable. feel free to file a feature request!
# 06:29 tantek snarfed - I don't understand how it's different
# 06:31 snarfed we only know FB (app scoped) user ids for bridgy users
# 06:31 snarfed we could get the rest from taggable friends (we think)
# 06:31 tantek ok - from a user-perspective, I don't know how to file this as a different bug / feature request
# 06:32 tantek I'm person-tagging with explicit FB profile URLs - assumed that would "just work"
# 06:32 aaronpk wow seeing all this discussion really makes me not want to go anywhere near the facebook api
# 06:32 snarfed tantek: yes. the UI wouldn't change, the work would be entirely under the hood
# 06:33 snarfed not entirely their fault; they probably have the single broadest and deepest feature set of any silo
# 06:35 aaronpk and largest user base, which means they are the largest attack surface
lukebrooker joined the channel
# 06:38 kylewm snarfed: not to beat a dead horse, but any idea how you'd match up the FB URL with the firstname/lastname we'd get back from taggable_friends?
# 06:43 snarfed kylewm: yeah, good point, no idea. that may be a showstopper
# 06:44 snarfed tantek: sure! the API part would have been useful too, since we know that's how we'd do it, but no matter, that definitely describes the user feature. thanks!
# 06:45 tantek I figured you could add the API details for the requested user feature :)
# 06:47 bear the beginnings of an api - looking forward to people filing issues for new features
# 06:47 bear cweiske - I do, working on what that is now
# 06:48 bear I wasn't seeing it in my local tests so that I fixed it
# 06:48 bear my hunch is that a dependency is out of date on the server
# 06:49 tantek snarfed, kylewm, next-up for me & person-tagging is to figure out a way to immitate person-tagging on the Twitter POSSE copy of photos
# 06:50 tantek I'm hoping that by experimenting and prototyping that I can come up with something straightforward for Bridgy Publish to do, or at least a recipe / functions for others to use when Bridgy Publishing a photo to Twitter
# 06:50 aaronpk i thought twitter supports proper person tagging now
# 06:50 Loqi A person tag (AKA people tag) is a special kind of tag that refers to a specific person by URL rather than just a word or phrase, and is done as an explicit tag by the user, beyond just mentioning a person via hyperlink / h-card https://indiewebcamp.com/person-tag
# 06:52 KartikPrabhu has it been answered somewhere except in the context of Twitter, which also seem arbitrary
# 06:53 tantek KartikPrabhu: if I say "photo" that's a mention of the word photo, whereas if I explicitly say "#photo" that's a tag of the term "photo".
# 06:54 tantek it has nothing to do with markup at the user level
# 06:54 aaronpk for me, I tag someone in a post if they are "in" the post somehow.
# 06:54 tantek KartikPrabhu: if you're using a loose / liberal meaning of "markup" - it's a meaningless argument
# 06:54 aaronpk for photos, it's when it's a photo of them. for notes, it's if the note is about them.
eschnou joined the channel
# 06:55 aaronpk I might even person-tag a note and not mention the person's name in it at all.
# 06:55 KartikPrabhu tantek: no. I am just asking when should I write a note "hi tantek" vs "hi #tantek"
# 06:57 KartikPrabhu "I might even person-tag a note and not mention the person's name in it at all." i don't see how you write a post about someone and then not mention them in the post but then tag them
# 06:57 tantek KartikPrabhu: if you don't use person-tags yourself in existing behavior, then don't bother
# 06:58 tantek when you start using it (e.g. on silos), I think you'll understand it
# 06:58 tantek e.g. difference between "someone took the photo" and "someone is in the photo"
# 06:59 aaronpk KartikPrabhu: "I had a great time at IndieWebCamp last week" (person tagged with .....)
# 06:59 tantek snarfed, do multiple "p-bridgy-twitter-content" elements get concatenated?
# 07:00 snarfed tantek: huh, no, but that's a great feature request, i see the use
# 07:00 tantek thinking: <p class="e-content p-bridgy-twitter-contentt">...</p> ... <p class="p-bridgy-twitter-content">...</p>
# 07:00 tantek excellent use of prefix scoped mf2 class names btw
# 07:00 snarfed oh that's a different use case than i imagined...but also good
# 07:01 tantek snarfed, in my particular use-case I'm going to provide tweet text and tweet virtual person-tags
# 07:01 tantek where the virtual person-tags are NOT part of the e-content that goes to FB
# 07:02 KartikPrabhu aaronpk: oh thanks for the example. So you wanted to tag anomalily because she is related to the post/you wanted her to see it, but the content is not about her
# 07:15 tantek snarfed, do we have a top 10 API wishlist (fixes etc.) for FB's API?
# 07:16 tantek snarfed, I'd like to collect such things for both FB API, and Twitter API
# 07:17 tantek so that if/when any one of us has a chance to ask 1:1 to representatives of those silos, we can ask for specifics
# 07:17 tantek feel free to braindump (you too kylewm) and I'll wikify somehow
# 07:17 snarfed twitter's is basically, please add API support for all features launched after 2009, kthxbye
e-lima joined the channel
# 07:18 tantek if you could reconstruct that list I'd appreciate it!
# 07:18 tantek tweet permalinks of each request would be even better!
# 07:18 snarfed yup! my two were getting a tweet's favorites and replies
# 07:21 kylewm for FB, * return the post permalink URL in the response to publishing a post
KevinMarks__ and tantek joined the channel
# 07:22 kylewm * find the ID of a post from its permalink URL (so I can like/reply to it)
# 07:23 kylewm * set a custom original URL for each post the way PESETAS from Instagram to Facebook does
snarfed1 joined the channel
Jihaisse joined the channel
KartikPrabhu1 joined the channel
acegiak joined the channel
# 07:32 snarfed * standardize on a single (or few) id formats, or at least document the existing ones, or at bare minimum stop adding more X
tantek joined the channel
# 07:42 matrrix snarfed: Does https://brid.gy/about#facebook-publish imply that if someone comments on a FB post that I have posted using Brid.gy won't get webmentioned back to my site? or is it the inverse that is affected?
# 07:45 KartikPrabhu matrrix: brigy-publish is for publishing while (normal) bridgy is for backfeeds. If you are signed up for bridgy you can do both
# 07:46 matrrix hmm, the backfeed isn't working for one of my bridgy-published posts...
# 07:47 matrrix Actually, the things listed in Responses are from May
# 08:00 bear yea, it says that the code returned from the auth providor is invalid
# 08:00 matrrix I just had to make the post public
# 08:00 matrrix is that supposed to be so?
# 08:02 bear cweiske - oh! ignore my pm ping - it's because you run your own indieauth service
# 08:03 matrrix Even when it is the one that posts it?
# 08:05 KartikPrabhu yeah. the FB API might not give access to read private posts but can give access to post private posts. But i must defer to snarfed or kylewm on this one
friedcell, fkooman, KevinMarks and eschnou joined the channel
# 08:50 bear cweiske++ doing one early so I can do another in a couple minutes
# 08:54 bear cweiske++ for being a very patient person while I debugged my indieauth code against his implementation
# 09:00 bear cweiske++ for being a very patient person while I debugged my indieauth code against his implementation
# 09:02 bear k, way past my bedtime - thanks again for helping me test cweiske
anm joined the channel
modem_ and mapkyca joined the channel
# 10:10 Loqi Ok, I'll tell him that when I see him next
adactio, KartikPrabhu1, frzn, ttepasse, glennjones and loic_m_ joined the channel
LauraJ, johnstorey, lewisnyman, Pierre-O and tantek joined the channel
# 13:16 tantek petermolnar: worth investigating how they could confirm Google indexed anything, and yet "JSON-LD content fails to show up in the search results."
zero-gravitas joined the channel
# 13:21 GWG !tell tantek Interested in attending. I have family there I can stay with.
# 13:21 Loqi Ok, I'll tell him that when I see him next
glennjones, minsky, loic_m__, loic_m_, mxuribe, zero-gravitas, ttepasse, danlyke and [snarfed] joined the channel
# 14:48 [snarfed] hey aaronpk just FYI indieauth choked on G+l ogin for me just now
lewisnyman and ttepasse joined the channel
# 15:06 GWG Working on designing my new contexts and kind presentation.
Pierre-O and snarfed joined the channel
loic_m_, mlncn and tantek joined the channel
# 15:46 tantek !tell GWG FYI I will be in NYC the weekend of the 19th-20th - might still be able to pull-off a one-day IWC NYC if we get a venue and rally a critical mass.
# 15:46 Loqi Ok, I'll tell them that when I see them next
shiflett and johnstorey joined the channel
loic_m_ joined the channel
# 15:55 GWG tantek, Saturday is better for me. But I'd come out to say hi either way
elima_ joined the channel
jedahan joined the channel
# 16:27 aaronpk how about once you get a second response identical to an existing one then you group them
# 16:28 aaronpk so if someone replied "cool" it would show up as a comment, but if someone else replied "cool" after then it would group them and show a counter
# 16:32 ben_thatmustbeme petermolnar: thanks also from that "This doesn’t just affect the ðŒ† character, but more important symbols like U+01F4A9 PILE OF POO (💩) ""
# 16:35 aaronpk alternate option, use an emoji library like twemoji to turn emoji into img tags, then group responses by img tag
# 16:41 aaronpk well this is odd, for some reason my mysql won't let me enter 💩even though it says the column is utf8mb4
# 16:42 aaronpk oh you have to set the whole table to utf8mb4 too
loic_m and GWG joined the channel
# 16:51 aaronpk did you do the optimize table thing so that it rebuilds it?
gRegorLove joined the channel
# 16:57 aaronpk it's a mysql command, not sure if it requires mysql root access
# 16:59 aaronpk i didn't need to change my.cnf for mine. i think that just sets up defaults for new tables.
# 17:03 ben_thatmustbeme SHOW VARIABLES WHERE Variable_name LIKE 'character\_set\_%' OR Variable_name LIKE 'collation%';
tantek joined the channel
# 17:12 GWG tantek: Think we could talk any of the Boston contingent into stopping by NYC?
# 17:13 ben_thatmustbeme my hosting provider is normally pretty good about getting back to me quickly on tickets
# 17:15 tantek GWG maybe? especially if we did it as one day that they could train in and out for without an overnight stay
# 17:15 GWG The bus is only $10 if you buy in advance
# 17:15 tantek I'm going to be asking the Mozilla folks this week in Orlando
# 17:16 GWG If not, I'll stop by and see you just to say hi
# 17:16 tantek Let's both ping Zilar again to see if he can come up with something for just the 19th
snarfed joined the channel
# 17:20 GWG I am certain he isn't with the Times anymore.
wholewheattoast joined the channel
davidmead, wolftune and todrobbins joined the channel
# 17:38 tantek since those are the only two actually documented / specificed / and very well interoperably implemented extensions so far
# 17:39 tantek "property" is more of an experiment than an extension, it doesn't even have an extension spec yet
# 17:39 tantek and thus because it is not even close the level of quality / progress as Vouch and Salmention, "property" does not deserve any consideration to be in core
# 17:40 aaronpk so what exactly are people implementing from then?
# 17:44 aaronpk I think my part of swat0 didn't require supporting it
# 17:48 tantek.com edited /Vouch (+9) "move protocol summary to before indieweb examples and rename to protocol, already have summary at top, this is the place for details." (
view diff )
nitot joined the channel
# 17:49 KevinMarks_ Google indexing js injected content and not JSON-LD is actually logical from a UX pov as the former is sen by users
funwhilelost joined the channel
# 18:00 tantek now just to move the relevant spec text from comment-propagation to /Salmentions and reword it to handle all responses (e.g. tags, likes like SWAT0 does)
loic_m_, bengo and loic_m__ joined the channel
# 18:13 snarfed looking forward to seeing people add snapchat, tinder, and yo!
davidpeach joined the channel
# 18:18 gRegorLove If the table character set and collation are correct, that's usually the only other step I've needed to do before inserting Unicode
# 18:20 gRegorLove Heh "Yo is a social application for iOS, Android, and Windows Phone. Initially, the application's only function was to send the user's friends the word "yo" as a text and audio notification,[1] but it has since been updated to enable users to attach links and location to their Yo's.[2][3]" https://en.wikipedia.org/wiki/Yo_(app )
friedcell and wolftune joined the channel
cweiske joined the channel
# 19:04 petermolnar KevinMarks_ I remember converting data from WINDOWS 1250 to to utf8; man, those were the days! /s
# 19:05 KevinMarks_ I set up the db at technorati to use utf8 from the start, before mysql supported it explicitly
# 19:09 ben_thatmustbeme petermolnar: still nothing, i'm pretty certain there are setting in my.cnf that are old and need to be updated
# 19:09 aaronpk ben_thatmustbeme: what version of mysql out of curiosity?
# 19:10 aaronpk on my 5.5.23 I managed to get it to not complain about invalid characters when I enter, but they show up as ???? when I view it
# 19:11 aaronpk petermolnar: not yet. I'm still on 5.6, probably won't switch to 7 for a while
# 19:11 aaronpk ben_thatmustbeme: yeah I wasn't sure if that was an artifact of my client or what
snarfed, hs0ucy, loic_m__, myfreeweb, j12t, snarfed1, Pierre-O and tvn joined the channel
snarfed and tantek joined the channel
# 19:55 Zegnat All my own code just kept running, so not really. But that is with a relatively "young" codebase, no PHP 5.4 compatibility that might have gotten deprecated
# 20:03 Zegnat But obviously YMMV, so read the migration guide. It isn’t as long of a read as you might think. And it can save you a lof of headaches in the end. E.g. they changed JSON parsers which might lead to bugs if you are pulling external data
Pierre-O and mxuribe joined the channel
squeakytoy2 joined the channel
bengo joined the channel
# 20:20 KevinMarks_ deciding that utf8 only allowed 3 characters at a time was the daft part
# 20:23 KevinMarks_ you may need to add SET NAMES utf8mb4 on the client or it will mung back to the previous 3-byte only utf8
bengo joined the channel
# 20:24 KevinMarks_ wow, they actually store 3 bytes/char in utf8 columns? how fucked up is that
j12t joined the channel
# 20:31 ben_thatmustbeme KevinMarks_: tried the SET NAMES bit, no luck, i've tried a number of ways, the server itself is set without support for it
# 20:31 KevinMarks_ seems to imply that indexes do some kind fo brain-dead parsing of the fields
# 20:32 ben_thatmustbeme once thats done, i think i have a way to recognize <any set of 1-4 emoji> (as some are multi-character compositions) or any single html special characters in that range
# 20:39 KevinMarks_ m̶̤͙͕͉̣̘͎̰Ì̫̱Í̖͔ͣ̑ͯ̂̉̀̂ͩͪ̓ͦ̓͢ͅyÌ¡Í‡Í‡Ì–Ì Í™Í‡Ì±ÌªÌªÌ¤Í†Í¬Ì‡Ì„Í®ÌŠÌŽÍªÌˆÌͩ͂ͮͨ̀ÍÍ s̡̡͌͋ͦ̆͊ͨ̈͋͗͊ͩ̽͗̈ÌÌÌ҉͇̜̜͎Ì̺͓̹̥̜͎̯qÌµÌ¶Ì£Ì¦ÌŸÌ—Ì¦Ì Ì–Ì¯Ì©Ì²Ì²Í’Ì͌̎ͣ͌͆̽ͫͤ̈Ìͬͪͧͤ̑ÌÍl̨̤̤ÌÌ–Í“Ì͚̜Í͒ͣͪ̾̾̋̅̓̑ͦͯͬ̕̕
# 20:44 snarfed hÍÌ̲͕͓m̜̙̟͔̣̫m̛̪̤̻̯͔͎͙.̫͉̪̺ ҉͓s̛̬̪̮̲̘o̟͔̤̦̗̯͟m̯̱͔eÍŸw͚̟̹͇̻̜ͅh̺͎͖̺̯̰̹aÍ•Ì͎̫͉̲̜͞t͇̪.̤̞ÍÌ°ÌŸÌž ðŸ¸Í“͎͙͓̳💩̩̖̥̫̫ͅâ¤ÌœÍ™Ì¥Ì̲ï¸Ì
# 20:44 Zegnat KevinMarks_, it will not let me combine them here. Though I am not sure if that is Unicode or Apple blocking it
# 20:48 aaronpk pretty soon this spec is just gonna end up describing bitmaps
# 20:49 ben_thatmustbeme start with control sequence .... then encode each block of bits as a character until ...
# 20:55 ben_thatmustbeme wonders if he should actually check for ZWJ and VARIATION SELECTOR to make sure i'm not just getting multiple Emoji right next to each other
# 20:56 aaronpk i assume that's the one being used on twitter.com so it must be pretty good
tantek joined the channel
tvn_ joined the channel
tvn_ joined the channel
johnstorey joined the channel
bengo joined the channel
# 21:14 kylewm.com edited /Micropub-brainstorming (+437) "/* Discovering Supported Syndication Targets */ note aaronpk's suggestion to host an h-card for the syndication target instead of passing information to the micropub client directly" (
view diff )
# 21:15 kylewm aaronpk: can you make sure i characterized your suggestion correctly there?^
snarfed joined the channel
# 21:18 ben_thatmustbeme argh.... The character's code point is U+200D zero width joiner (HTML ‍ · ‍)
# 21:19 tantek KevinMarks - perhaps #indiechat? Unless you're specifically testing archives?
hs0ucy and [aaronpk] joined the channel
# 21:22 [aaronpk] kylewm: yes that's it. I did add the jf2 suggestion right below that too
shiflett, tantek, bengo, mlncn, lukebrooker, singpolyma and friedcell joined the channel
# 21:45 tantek !tell ben_thatmustbeme when you were implementing Salmention sending, what spec / page / instructions did you follow?
# 21:45 Loqi Ok, I'll tell them that when I see them next
# 21:46 tantek !tell aaronpk when you were implementing Salmention receiving, what spec / page / instructions did you follow?
# 21:46 Loqi Ok, I'll tell them that when I see them next
# 21:47 aaronpk hmm so I did the part where I had to recognize that a post was updated when I received a webmention that I had received before
# 21:48 tantek interesting - did that already make it into the core? I mean it makes sense obv as part of SWAT0
# 21:48 aaronpk well it's just webmention update handling. not sure it needs its own name
# 21:49 tantek that's specifically just for updates from the source itself
# 21:49 tantek intended to scope to just that h-entry, its content
# 21:49 tantek whereas extending that to the *childern* of an h-entry, e.g. comments on that h-entry, that's the Salmention piece
# 21:49 aaronpk "it SHOULD update any existing data it picked from source for the existing webmention" doesn't mention h-entry or anything
# 21:50 tantek aaronpk - indeed it's vague - yet the intent was "existing data" really only assumed the h-entry directly, not including its children
# 21:50 aaronpk well it just says "source" so it's pretty vague and doesn't really communicate that intent IMO
# 21:50 tantek Salmention makes that explicit, that you must check the h-entry AND its children for updates
# 21:51 tantek so the question is whether we should make that explicit in the core spec, or leave it to an extension
# 21:52 aaronpk I think the explicit language around posts and comments should be in the Salmention extension and that's a good justification for the extension existing
# 21:54 aaronpk fwiw adding the code to handle checking for updated children added quite a lot of complexity to my webmention processing, and it is not even a complete implementation of it
# 21:55 tantek should "Updating existing webmentions" explicitly mention properties of the h-entry as a minimum viable/required/MUST aspect then, and then MAY for checking for updated children?
# 21:55 tantek and then the Salmention extension can make checking for updated children a MUST
# 21:55 tantek just trying to find out a good incremental separation here
# 21:55 aaronpk it would have to mention it in vocab-agnostic terms such as "properties of the object on the page" or something, since the webmentioned document may not be HTML+microformats
# 21:57 aaronpk but, is there value to not requiring that behavior?
# 21:58 tantek only if an implementation has value without it
# 21:58 tantek in general, specs work better (get better interop) with fewer features, all being required / MUSTS
# 21:58 aaronpk so the question then is how many current webmention handlers check for updates?
# 21:59 aaronpk up until the SWAT0 demo I treated an update naively, just replacing the existing content from the page with the new content, but the notification was the same as the first time it came in
# 22:00 tantek aaronpk - I think that's a reasonable implementation of updates
# 22:00 tantek the key thing is that the MAY for checking children of the object could there provide the "for example, to support Salmentions" (spec link)
# 22:00 tantek that would help with incremental implementability
# 22:01 tantek being able to "complete" an implementation of Webmention without Salmentions, then later add support for it
# 22:10 tantek.com edited /Webmention (+466) "/* Updating existing webmentions */ expand upon webmention update implementation requirements, may, and if so, MUST support Salmentions" (
view diff )
# 22:11 tantek note the way the SHOULD / MUST + MAY / MUST interactions work
# 22:16 tantek aaronpk - it's a MUST conditionalized on a MAY
# 22:16 bear the MUST is nestled nicely in the part about the optional implementation in the MAY part
# 22:16 aaronpk yeah i just mean that if someone were to develop an extension that worked in a different way it woudln't be possible to reference it then
# 22:16 tantek that's the point, it says you MUST at a minimum support direct properties, but may support more, but if you support more, you MUST do so this way
# 22:17 tantek correct - by default for interop we assume there is no need to develop an extension that worked in a different way
# 22:17 tantek because the assumption of "an extension that worked in a different way" leads to *less* interop by default
snarfed joined the channel
# 22:17 aaronpk so it's extensions for the purpose of eventually coming to consensus on a single spec
# 22:18 aaronpk and also encourages more implementations since each piece is smaller
# 22:18 tantek not even "eventually", but rather, consensus by default, and raising the barrier for other options to require s a good use-case with good data
# 22:19 tantek yes re: encourages more implementations since each piece is smaller
# 22:19 tantek this is a key way of designing modularity into specs
# 22:21 bear setting the tone for small but purposefully composed modules is key
# 22:21 bear at least in the dozens of failed specs i've witnessed over time
# 22:21 tantek I too have witnessed far too many failed spaghetti specs
# 22:22 tantek where you think you're done, and then there's a *requirement* to implement a whole another spec
# 22:22 bear the xmpp community is digging out from that tech debt by forcing new specs to be tighter focused
# 22:22 tantek whereas this way, in the "core" spec, the "feature" is a MAY (which means, STOP HERE by default), and if/when you do decide to venture further, here is how to do so
# 22:22 tantek it also forces the core spec to be functional WITHOUT extensions
# 22:23 aaronpk sounds like the opposite approach to modularity than OAuth 2 took :P
# 22:23 tantek otherwise you end up with monolithic outcomes
# 22:23 bear that part is missed by a lot of folks - how important it is for the core spec to just-work
# 22:23 tantek also why it is good to reject by default experimental additions like "property"
# 22:24 bear later, if a extension ends up being always implemented and of such value, it can be folded into the core
# 22:24 bear but then that is a purposeful choice driven by implementors
# 22:25 Loqi ben_thatmustbeme meant to say: i think i have it http://pastebin.com/YDGJUwRA
# 22:26 ben_thatmustbeme and thats not even following all the possible options just a simplified one for all those currently used
jedahan joined the channel
# 22:32 tantek bear - if the modularity succeeds, it's less work to leave it modular!
# 22:32 tantek e.g. note that HTML does not include all of HTTP as a big thing
lukebrooker joined the channel
# 22:46 Loqi Ok, I'll tell them that when I see them next
shiflett joined the channel
# 22:49 bengo Have y'all tried Medium's 'import from URL' for POSSEing to there?
# 22:49 bengo Yeah I just had to edit some of the code blocks in lists and stuff
# 22:51 KevinMarks_ my new game with medium is to POSSE very old blogposts there when I am reminded of them
eschnou and tantek joined the channel
# 23:09 aaronpk lol! why does bridgy think my URL is micropub.net?
# 23:09 aaronpk it's in my bio, but i've set my URL to aaronparecki.com
# 23:09 snarfed we're promiscuous and take all URLs in your profile
# 23:11 aaronpk seems like a reasonable choice to prioritize the one in the actual URL field
# 23:12 snarfed and the visible html in those mf2 pages is low low priority
# 23:12 aaronpk feature request: visibly link the profile URL and put it first in the list, and treat other URLs in the bio as alternate URLs and make them show up after
# 23:13 aaronpk because now my comments on my posts look like they came from micropub.net
# 23:13 aaronpk because most webmention consumers only care about the first author URL
# 23:14 aaronpk and my notifications said "micropub.net commented on your post"
# 23:14 snarfed odd, you're right, it doesn't seem like this has been broken for long, but i don't know what would have changed recently
M-Kodo, M-RyanRix, M-kegan, snarfed1 and KartikPrabhu joined the channel