#tantekeprodrom cwebber2 if you /me in IRC you can chat without getting meta-chat recorded in IRC minutes
#cwebber2aaronpk: there were a few things to address, I've been working on a new draft that should be ready to go, I've summarized the pages in document
#cwebber2aaronpk: here's the latest micropub draft, I've added in a link to the test suite describing what the test suite will do, I've added conformance and exit criteria listing all features you'd expect, basically spec infrastructure we did with webmention
#cwebber2aaronpk: so the accessibility group looked over all 5 specs, only had comments about micropub, they said they'd like examples of posts including accessibility info, such as alt-text on an image
#cwebber2... I've opened this as an issue on micropub
#cwebber2... main issue is there isn't a syntax to provide alt-text of an image, because microformats doesn't either, and it's based on it
#cwebber2... so I've opened an issue on microformats to track that as well... so that's blocked until we can solve from MicroFormats side
#cwebber2tantek: from SocialWG perspective of studying proprietary APIs, have you done any background research as how twitter, instagram, etc have provided alt-text of an image, or if they do at all? if they don't that's also useful info
#cwebber2... from what I've seen of people who actually consume alt-text of images, it's a description of the image, rather than what someone says about the post
#cwebber2tantek: that's my understanding as well, and twitter allows you to enter alt-text as a separate field in the ui
#cwebber2... so from a ux perspective, certainly a deistinction
#wilkiealt-text is objective and a title is subjective. yeah, twitter is a good example of allowing one person to provide both types of description.
#cwebber2tantek: it may be good to discuss inline that there's an outstanding issue and summarize and link to it so that anyone else reviewing the draft can go "oh yeah someone took a look at this"
#cwebber2eprodrom: sure, so we had CR transition meeting which went well
#cwebber2... we had one major element from meeting, which was at-risk features
#cwebber2... we could have as exit criteria which was that any features not implemented in at least 2 implementations need to be marked for removal
#cwebber2... making it possible to have a bit more of a process in removing elements
#cwebber2... we got pushback on this for probably some pretty valid reasons
#cwebber2... first of all, we didn't actually use the term "at risk" which is the proper term for w3c
#cwebber2... so we were putting every feature "at risk" but not flagging(?) them "at risk"
#cwebber2... the second is the question of whether it makes sense that everything at risk, with the reductio ad absurdum (?) that nothing is implemented, or we have a small set of features and types that don't really hang together or have relationship together
#cwebber2... so ralph has suggested that we identify a small number of classes which will definitely be in a future version, where we wouldn't be able to go forward without those types and properties
#cwebber2... most of these were addressed by James, who did an omnibus pull request over the weekend, he can't be with us this time around, but what we have is a number of items about i18n in activitystreams
#cwebber2... the one normative recommendation was that if there was not a way to identify the default language for a document that we recommend that implementers use the named map / content map / ** map properties rather than their default version
#cwebber2... james noted that there are some implmeentations that will not be using the json-ld structure, and so he recommended that we use the longer mechanisms
#cwebber2... he actually had text that said which we should use the map version
#cwebber2... my feeling is that these map versions... that was a normative change, deprecated properties replaced them with these map ones
#cwebber2... I think that instead we should suggest we should use the default language mechanism
#rhiarocwebber2: I don't have a strong opinion. I think that it's not a bad thing or very hard even if you're not using a json-ld processor to look in that context
#rhiaro... Assuming that you're not using.... I don't have a strong opinion
#cwebber2eprodrom: another possibility is that if we use a language tag which means unknown, that could be somethign we use here to specify that if there is not a language tag, assume it's UND (undetermined)
#cwebber2sandro: then we wouldn't need to change the examples
#cwebber2eprodrom: we should probably change the examples anyway
#rhiarocwebber2: I think I'm all for the suggestion that we make it a SHOULD when you know the language and we also say the implementations must look at @context to look for it. It's not too tough to look there even if you're not using json-ld
#rhiaro... But I'm not sure that we should change every example in the activitystreams doc to include an @context with the language property, the reason being that there's a classic problem where people start wanting to provide a language with content they don't know, and suddenly you have a bunch of content in another language tagged as English because the programmer was lazy
#rhiaro... I think it's okay ot have some examples int here that might reflect that user submitted content is probably unknown in most cases
#sandro+1 cwebber2 important to avoid giving @language when system doesn't know it!
#rhiarosandro: It looks like in json-ld it can go in the ocntent if you're using the expanded form, but since you're using the compacted form there's no place for the language to go so it has to go in the context
#rhiarotantek: sounds like there's more non-trivial discussion, so we can move on to other points. Do we have a github issue for this?
#rhiaro... There is the question, I'm not exactly sure what happens next, but I think we push back publication. When do we make our decisions here and then go forward? Or do we fix after we publish?
#rhiarotantek: the first part is there is non-trivial discusison, we're not going to get an answer on this telecon. We allow people to continue iterating on that issue, specifically cwebber2, rhiaro and sandro to chip in more
#rhiaro... Second is about publishing. Either we decide to publish with this issue outstanding in which case we may want to call it out explicitly in the draft
#rhiaro... Alternatively, up to you evan, is to wait to publish to resolve this issue first, then incorporate
#rhiaroeprodrom: At this point we can publish a working draft that *evan is cutting out*
#eprodromI'd like to publish a WD and then look at CR next week
#rhiarosandro: I think I heard Evan say he wasn't comfortable overriding one of James's decisions without James, so sounds like we need to wait until next week until James is back online and we can get his attention
#rhiarotantek: sounds like you had a similar approach, just figuring out syntax
#rhiarosandro: No question what the syntax is. I don't believe there's a design decision
#rhiaroeprodrom: We've got two feet on the ground and it's which we want to lean on. I'd like to publish what we have as an ED as a WD, and that will make it easier for review
#rhiaro... With James's changes and the new at risk language
#rhiaro... Just won't have the normative change to recommend using the map properties
#rhiarotantek: Can we call that out as an issue inline
#rhiaro... sandro could I have an update of what are we publishing and when?
#rhiarosandro: So we could try to go ahead with jf2 by itself. I had been waiting for other things, but as you've heard the other things have been snagged on other issues
#rhiaro... I guess we might as well go ahead and do jf2 by itself
#rhiarocwebber2: I've been pushing pretty hard on getting an implementation out there and am pretty close to giving a usable draft
#KevinMarksthe specfic issue will be for parsing photo explicitly, and if we need to define a new object for those to hold richer metadata in parsed form
#ben_thatmustbemesandro, i'm assuming you want me to rewrite those staged version with the new date (7/19)
#ZakimAs of this point the attendees have been sandro, aaronpk, rhiaro, akuckartz, cwebber, csarven, wilkie, eprodrom, annbass, ben_thatmustbeme, KevinMarks, tantek
#Zakimleaving. As of this point the attendees have been sandro, aaronpk, rhiaro, akuckartz, cwebber, csarven, wilkie, eprodrom, annbass, ben_thatmustbeme, KevinMarks, tantek
#rhiaroaaronpk: oh because the annotation on there is currently stored on a solid server which doesn't know how to not bother trying to auth you for public stuff
#rhiaroThat's been fixed in the new server, we need to update
#rhiarotantek: extensively on various whiteboards in our office, but presumably somewhere else as well, I'll check with dmitriz. I believe the main reason for openidconnect was the multiple existing implementations of services and libraries. I've been out of the loop on the actual dev of the auth stuff for the last month or so though
#tantekrhiaro, yes was wondering if anyone captured anything citable/linkable online
#rhiaroThe indieauth refinements and openidconnect refinements of oauth2 are actually pretty similar if I understand correctly
#tantekresists making his usual joke of ask LD folks for *links*.
#rhiaronotes that that emote was actually not resisting at all :p
#tantekThe biggest problem with openid connect is that it seems to be following the same multiple central silos (who don't accept each other's identifiers) antipattern that openid did
#tanteklet me know if you find two big "implementations of services" who actually "peer", that is, both produce and consume each others identities. until then, OpenID Connect is no better "socially" or in a distributed/decentralized fashion than OpenID 1 or 2 etc.
#rhiaroI'm not well versed in the details, but I thought it used emails as identifiers
#tantekalso seriously? "multiple existing implementations of services and libraries" [citation needed]
#rhiarooh okay, when I was looking I found a library in Python and PHP, and dmitriz is using a nodejs one in the solid implementation
#tantekanyway, yeah, just sounds like another decision made without documenting research in a place online citable by URL
#tantekrhiaro: "implementations of ***services***"
#tantekanyway, I'll wait for the online documentation, otherwise, I'll accept that "whiteboarding" is the decisioning making methodology that was used (and what that entails).
#rhiaroI would hate to mispresent how things are because I'm ill informed. I'll stop talking about it.
#tantekyou mean about "Also certs are going away altogether sooooon" ?
#rhiaroBut while I'm looking at it, there's a list of implementations here and to what type of conformance they have (to what extent they talk to each other, I think) http://openid.net/certification/