#elf-pavlik!tell harry will you stay in Paris one of next days, I plan to stay around here for couple more days and we could meet up :) also http://www.xwiki.org team would like to discuss possibilities for interop
#elf-pavlikrhiaro: if you can, please let me know ASAP if you and sandro can make WebEx calling out to me, for now based on email reply from Arnaud it doesn't sound promissing :(
#Arnaudelf, can't you use something like skype to call in?
#elf-pavlikArnaud: friend made an account (with some calling credits) available for me but now it doesn't have enough left to make another call, I guess I may end up IRC only for now
#ArnaudI guess it's hard to function in this world without any money :)
#elf-pavlikjust in thu current *state* of this world, future looks much more promissing since many people work on this issue :)
#tantekyeah, giving up on captive portal support on iOS9 - totally broken. DO NOT UPGRADE to iOS9 if you have an iOS device - it's wifi support is fundamentally broken.
#elf-pavlikaaronpk I thought no one has host code ... who called you?
#rhiaro_eprodrom: Are we going to ahve any further problems, or will we do more monthly drafts as expected?
#aaronpkelf-pavlik once you enter the meeting it has a button to call you. no host code needed
#rhiaro_jasnell: assuming we get the process down it should be fine. The tools have been getting in the way. Once we've got through it once, then it should be much more regular
#rhiaro_... There has been discussion and progress with text, but we require more than just a text document
#rhiaro_... We need test suites and implementation plans
#rhiaro_... People committing to implement the spec
#rhiaro_... There is an exit criteria for CR, we invite the world to implement and gather implementation reports, usually using the test suite
#rhiaro_... People can generate reports using the test suites against their implementation, send reports back, someone cosolidates the reports and we use this to justify that the spec can got to proposed rec
#rhiaro_... So the two aspects here that are important are the test suite and the plans to implement
#rhiaro_... Everybody knows we have had an effort made by IBM to start a test suite, but I was hoping there would be peopel who can help out and take it to the next step. We haven't seen that happen. It's unclear at this point who is planning to implement it
#rhiaro_eprodrom: I'd like to discuss the timeframe that we're in right now. We've been discussing among chairs, but I put it on the proposed agenda items for f2f, that we'll be discussing the progress of AS2.0 and what our way forward is
#rhiaro_... We'll need to make a decision if we've had enough progress with that by the f2f to justify going to CR
#rhiaro_... If we can't justify it, we need to discuss alternatives
#rhiaro_... Do we continue to work on it after the f2f, if ther'es progress we can postpone and make the decision later.
#rhiaro_... We need to define what the steps are with consumers
#rhiaro_... How do we validate a consumer of activitystreams?]
#rhiaro_... THat said, I think we have a lot to do
#rhiaro_... There are some philosophical differences
#rhiaro_... Some are of the opinion we are documenting the state of the industry. If we get to the f2f and there haven't been changes outside our group, our job is to represent that external reality and make our decision based on that
#rhiaro_... There are others who feel that as a WG we can be pushing this forward and it seems that we have a few clear paths to go forward
#rhiaro_... I think the test suite is something we can bring to the table
#rhiaro_... Those who are planning to implement, free open source implenenations, will definitely move us forward
#rhiaro_... And outreach to existing implemenations
#rhiaro_... We have 2 months to go forward. For those interested in seeing it get to CR, we have work to do
#rhiaro_... THis is my call to action to get us starting to do this work
#rhiaro_Arnaud: What we're tryign to tell everybody is that we are concerned we don't see much activity on those axis. We need to look at this seroiusly and come to the f2f with expression of support or not
#rhiaro_cwebber: I was working on an AS2.0 representation library, and having worked on it it made me think that the most technical aspect of it is the optional requirement of JSON-LD. Otherwise mostly it's just a serialisation in JSON
#rhiaro_... So that really makes me wonder what a test suite would look like
#rhiaro_... We've discussed this before, nobody gave a clear answer
#elf-pavlikq+ to ask cwebber2 if he used AS 2.0 extensibility
#rhiaro_... otherwise i'ts just json objects structured in a specific way. What is there to actually test?
#rhiaro_elf-pavlik: Question for chris: I wonder if you use some accessibility? You just use provided context, or you use other terms not in AS2 vocab?
#rhiaro_cwebber: I'm not interested in talking about my own implementation for this part of the call. But I ended up hitting the point where I wanted to implement types, and if we did have the option to extend with JSON-LD I needed to write a JSON-LD expander, so I did
#rhiaro_eprodrom: I want to answer that. I'm not sure if we want to go into that in depth in thsi call. Maybe we could start developing a wiki page here ^
#hhalpinApologies guys, was getting the deal re WebEx from Wendy and W3C Management - have a brief update. TL;DR If group has consensus, moving to Mumble is OK if there is group consensus
#rhiaro_... THe second is to make sure consumers are 'understanding' what the input is
#rhiaro_... I've been trying to look into some of the other test suites for other document formats
#rhiaro_... I'd like to see us produce something ideally.. soem sort of test driver that produces correct output
#rhiaro_... So we can test consumers. Somtehing like a commandline test driver so you can fire it at a library and let it parse a document and produce certain output
#rhiaro_jasnell: Evan's point about giving it those scenarios and test if they're valid, eg. Sally uploaded a photo, there are only ac ertain number of ways that can be encoded, we can test if that works properly
#rhiaro_Arnaud: Anything else? Otherwise evan's suggestion is a good one. I inviet everyone who has an interest in this to follow on with a discussion [on the wiki]
#rhiaro_Arnaud: elf-pavlik, I believe you added this, links broken
#rhiaro_elf-pavlik: The point I added about relevance to as2.0 vocabulary to social api, I would like to clarify if there's a strict requirement to use vocabulary in social api and federation, or if there will be another vocabulary
#rhiaro_... I think we may have to wait to finalise the vocabularly until we know what we will use in the social api and federation
#rhiaro_... I would like clarity on the approach. Separate vocabulary for api and federation, or we want to make sure to include everything in AS2.0 vocabulary?
#rhiaro_eprodrom: I think this is a concern that we don't even have consensus that we're going to use JSON in our social api, much less that we're going to use AS2.0 or JSON-LD. I don't think this makes any sense. I think if we're going to use AS2.0 we need to just make it go forward rather than holding off
#rhiaro_... I think of the spec that we have, by far the one we have farthest along is AS2.0
#rhiaro_... I don't want to wait for social API to get AS2.0 out
#rhiaro_Arnaud: We were just talking about the challenges we are having with moving AS2.0 forward, if we tie it to something even less defined that makes it even harder
#rhiaro_sandro: Is the question, are we committing the API to using AS2.0, or is it okay to design the API in such a way that it can use something other than AS2.0?
#rhiaro_eprodrom: Of the 3 candidates we put together, only of them uses AS2
#hhalpinJust to ask, "Would anyone be uncomfortable to be accepting AS2? *with other syntaxes being possible to accept?"
#rhiaro_... I agree with harry that it feels that the charter is we have the social data syntax, and an api that uses that syntax. If we went with an api that does not use that syntax, it would be pretty remarkable, we would have to justify it
#Zakimsees elf-pavlik, tantek on the speaker queue
#rhiaro_... I don't think that's a settled decision in this WG
#rhiaro_... If we do not take AS2 to CR then we need to look at the purpose of this group and if we have a mandate to go forward with API and federation protocols that do not use an existing syntax
#hhalpinI would say it should accept AS2 (assuming it is a Rec) and can accept other syntax choices, with the other two being pure unadulterated RDF and another being microformats
#rhiaro_tantek: Certainly AS2 is the most mature of all the different technologies and charter areas that we've been pursuing. Like evan, I'm conerned that if we're not going to make progress with AS2 then we need to take a hard look at the purpose of this group
#rhiaro_... On the other hand, regarding API candidates, one of the strong candidates which is micropub (strong on the basis of numerous deployed implementations interoperating clients and servers) does not reuqire AS2
#rhiaro_... I think there is potential for compatibility with AS2
#rhiaro_... One of the reasons I followed up with post-type-discovery is to explore areas for compatibility
#rhiaro_... I'm not concerned with being bound to AS2. I'd rather have a working proven API than one that is bound to previous decisions
#rhiaro_... But I"d like to see how we can make all these peices work together
#rhiaro_Arnaud: It seems lik there is agreement that we shouldn't tie the two together
#rhiaro_... It's still unclear what the protocol/API is going to be
#rhiaro_... Best if it could leverage AS2 somehow, to which degree is less to be defined
#rhiaro_eprodrom: I don't want to move off this topic... elf's original proposal is that we hold off on publishing vocabulary until api and federation are better defined
#elf-pavlikif you would like to speak about API, i can scribe for a bit!
#rhiaro_elf-pavlik: Some of the terms are related to API and federation. Terms like I've listed on agenda page, like paging and audience targeting
#jasnellgiven that we don't even have an API Editor's Draft, it would be insane to tie AS 2.0 progress to API progress.
#rhiaro_... I would like to clarify that if some of those terms are part of API or federation, if we want to publish AS2 we should remove terms that are not specific to modelling data
#rhiaro_... The whole conversation about AS2.0 and whether that's linked to standard stuff.. when I was in Boston over the weekend I talked to people in MediaGoblin community and others in this group, and had a lot of thoughts
#tsyesikaI'd also like to see discussed the Webex/mumble issue hhalpin added to the agenda, I am only participating via text as I had WebEx issues
#rhiaro_... I think elf is right, we do need to figure out if AS2.0 is linked, but that's a whole call in itself, so lets not do that this week
#rhiaro_... But the other side, int erms of implementation, tsyesika is close to landing a massive rearchitecting of MediaGoblin so we can support federation
#rhiaro_... THis has been holding this back, it's nearing actually working. This was all set up from the original plan of working towards pump API server to server stuff happening
#tsyesikaArnaud: I'm also hoping we can discuss WebEx issues
#rhiaro_... Second question, for implementation, would be useful to have python library that can consume and produce AS2.0. Is that something that's part of these next steps for MediaGoblin, and something youc ould share?
#tsyesikaArnaud: I and others have struggled joining over the last few months due to the changes
#rhiaro_cwebber: I think that's possible. Need to talk to tsyesika. I'm optimistic. Was thinking when you said earlier that we do a commandline test suite, it might not be too hard to write that as a little python application. Something like that might end up happening.
#tantekArnaud - can we assign the WebEx issues to the chairs to handle offline?
#rhiaro_... Let us talk about it and discuss next week
#rhiaro_hhalpin: I was just discussing this with Wendy Seltzer. Other groups have had this issue. No other group has resolved it successfully. But as long as the group has consensus on what software and everyone can use it, it is okay to swtich off webex
#cwebber2sandro: I sent him an email about the process of Invited Expert
#rhiaro_... So if we want to swtich to Mumble, that's fine by w3c
#ZakimAs of this point the attendees have been Arnaud, csarven, rhiaro, aaronpk, shanehudson, sandro, elf-pavlik, kevinmarks, wilkie, eprodrom, jasnell, ben_thatmustbeme, cwebber, tantek
#wilkie15 steps to decide an object type based on some arbitrary fields being present seems flimsy from an extensibility and a robustness point of view
#elf-pavliktantek, i will comment on the wiki in few hours
#KevinMarks_wilkie: what si flimsy in practice is explicit object types that are invisible
#aaronpkeven facebook and twitter don't have explicit object types
#tantekelf-pavlik: note that commenting doesn't stop when moving to w3c wiki, this is just about accepting it as an *editor's* draft since it's a resolution to one of the group's actions/issues!
#tantekelf-pavlik: I expect feedback/iterations to continue while working on it in the group
#wilkiewhat is the difference between this algorithm and an explicit field that says what it is meant to be rendered as? except for the complexity of it.
#KevinMarks_the assumption that objects fit into hierarchies is at odds with how people perceive the world
#KevinMarks_we actaully think of things as matching various prototypes to different degrees
#wilkiebesides, a 15-step algorithm is an ordering anyway, but that doesn't matter. a type is a type. not necessarily describes any subtypes.
#wilkieif you don't understand that explicit type, it's a 'note'
#KevinMarks_this draft models that process more empirically than requiring upfront labelling according to an existing model
#wilkienot sure what you gain with this algorithm, unless it is a fallback to an unknown explicit type. but the algorithm is too specific for that.
#elf-pavliktantek, it looks specific for http://microformats.org vocabulary but it says "Activity Streams (1.0 or 2.0) JSON, or JSON output from parsing [microformats2]"
#wilkiehow do you add new types to the algorithm? what would that process be?
#KevinMarks_you gain the ability to interpret things that have not been explicitly labelled according to your world model
#tantekelf-pavlik: do you have a specific question?
#KevinMarks_the process would be comparing examples and seeing where the decision point is
#elf-pavlikit also diesn't have links to AS2.0 drafts even that it makes it a normative reference
#tantekwilkie - so far based on publishing experience and evidence
#tantekelf-pavlik: a-ha was waiting for today's publication
#KevinMarks_having written a lot of crawlers, you start out with supposedly clear divisions and end up with heuristics. Documenting and converging heuristics is a good idea.
#eprodrom* Does it fit with our charter? In other words, can we argue that taking on this work as part of the WG is related to the work that we're supposed to be doing?
#eprodrom* Do we have the bandwidth for it? As a working group, do we have the time and attention to work on this document and move it forward?
#tantekeprodrom: it fits within the charter as its part of resolving an issue for AS2!
#eprodrom* How does it relate to our other deliverables? Is it a replacement for the JSON-based social data syntax, or kind of a preprocessing best practice?
#tantekeprodrom: it's a small document, and I have the bandwidth to work on it
#eprodrom* What are our success criteria? Are we intending to publish this as a Note or Recommendation? Or is it there to inform implementers of the other specs? Or are there other goals for continued work with it?
#tantekit's not a replacement for the JSON-based social data syntax, if anything it helps generate it
#eprodrom* Who will work on the document? Who will be shepherding this format?
#eprodromSounds like we have an answer to that last one, which is great
#elf-pavliktantek, I recommend taking with rhiaro about some major differences between how Microformats vocab and AS2.0 vocab work, I guess she understand them best out of all of us here
#eprodromAnd sorry for the copy-and-paste to the channel
#tantekeprodrom: next step after editor's draft is to publish as a WD. eventual destination whether NOTE or REC will be up to implementation experience and WG
#tantekit's definitely intended as a spec for implementers, so that we can get interop on Post Type Discovery for implementations that work without explicit typing
#elf-pavlikthose models have *major* differences so tantek's draft only stays relevant to Microformats vocab
#tantekelf-pavlik: I think you may be misunderstanding - this draft is showing how implementations can discover the types which can the be treated as AS2 types
#elf-pavlik"The Microdata, RDFa and Microformats examples included in this document are purely informative and may not currently reflect actual implementation experience or accepted best practices for each format. These alternate serializations may be removed from future iterations of this document and moved to a separate informative WG Note."
#KevinMarks_in this case, hat like does represent actual implementation experiences
#elf-pavlikKevinMarks_, but you don't find like-of in *normative* AS2.0 Vocabulary
#elf-pavlikone can consider it 'background noise', at least as of its current state
#KevinMarks_that implies the vocabulary needs changing to match reality to me
#elf-pavlikif you like to change it, please propose it
#elf-pavlikat this moment, as of nomative part of the draft AS2.0 DOES NOT include *like-of*
#aaronpkelf-pavlik: I'm pretty sure tantek has repeatedly said he is willing to add examples of AS2.0, why are you insisting otherwise?
#elf-pavlikfolks, in many ways i enjoy this dance here but my belly demands Activity with property verb:eat and maybe first i need Activity with property verb:cook
#elf-pavlikBTW this may actually work with proposed discovery :)
#csarventantek Not opposed to it, but that's rather out-of-band, i.e., everything else is self-descriptive
#elf-pavlikin context of current version of AS2.0 working draft (just published today) which doesn't need to reflect some experience of some peopel
#tantekcsarven: was attempting minimal (re)invention, by re-using 410
#rhiaro_So... point of post-type-discovery is if you are a consumer, and you understand AS2 with explicit types, and you consume some data which *doesn't* have explicit types, you use this algorithm to get an explicit type so you can move on
#tantekcsarven: not "out of band" as all other HTTP responses must be accepted also, e.g. 200
#elf-pavliknow really gone for Activity with action verb:cook and Acitivity with action verb:eat (which can support type discovery...)
#rhiaro_but if you have a bunch of data which *otherwise* conforms to AS2, just without explicit types, post-type-discovery doesn't actually help for that yet
#rhiaro_it does help if the non-typed data you have found is mf2
#csarventantek Is rhiaro's paraphrasing above accurate?
#tantekrhiaro_: right, so I asked for real world publishing examples of "a bunch of data which *otherwise* conforms to AS2, just without explicit types" so that I could specify the algorithm accordingly
#rhiaro_In order for post-type-discovery to be complete, would there need to be a path to follow to reach every explicit type in AS2?
#aaronpkcsarven: would a "deleted at" timestamp suffice?
#tantekcsarven: likely for display purposes yes, just as HTTP 404 responses typically include HTML that says something about the Page not found
#tantekrhiaro_: also another reason to not worry about "every explicit type in AS2" is that there is no evidence to suggest all the explicit types in AS2 are needed - that is, I think a lot of them could be dropped based on absence of interest / implementation / publication
#csarventantek aaronpk I was hoping for something more "machine-readable" than natural language. So, yes, an explicit property for deleted would go in that direction.
#tantekas in you want the 410 page to have the data in it?
#rhiaro_tantek: sure, I agree that not everything in AS2 is needed. I meant, assuming that the final result of AS2 is a version in which core things remain
#rhiaro_So, to me right now, post-type-discovery reads as a bridge between non-AS2 publishers and AS2 consumers. Which seems very useful - but is that in the scope of the group?
#KevinMarks_(which I coincidentally saw a minute ago)
#tantekcsarven: that makes sense to me. I'd likely store such "deleted" h-entrys myself in my own storage as tombstones of deleted entries!
#tantekwhich my CMS would use to then know to return a 410 response
#csarventantek I don't want anything in particular. I'm just trying to understand whether the proposal or what it is headed towards has concerned or working on a solution that is *to be an alternative* for how typing is derived in AS2 (I'm not sure at this moment whether the proposal is for AS2 only or everything else... I've only noticed the ISSUE being categorized under AS2)
#aaronpkoh that's a good point, once the HTTP request is done, there is nothing in the storage of the post that indicates it's deleted
#tantekdt-deleted provides a nice complement to dt-published
#tantekand there have also been brainstorms about dt-created for offline posts before they are published
#csarvenAs rhiaro was trying to rephrase tantek's point earlier; if in the absence of explicit types, one can derive the type via explicit properties -- that sounds reasonable to me. However, two things to consider: 1) whether that alternative is just necessary or not 2) whether that is of interest or should be part of something that the WG needs to deliver. I don't have a strong view on this.. just hoping that it is not overlooked.
#aaronpklike if i'm storing a copy of tantek's posts, I would need to store the HTTP status code somewhere if there was no "deleted" property
#KevinMarks_right, that's the issue - having a marker that this is a tombstone is csarven's point
#tantekpoint is, how does it justify a separate dt property from dt-updated?
#tanteksure - try to avoid them, just like CSS does
#KevinMarks_it could be an explicit type, h-tombstone, but that seems counterproductive
#ben_thatmustbemegah, i go to add my site to the indieweb examples of deleted, and there is another page of text to catch up on
#KevinMarks_as an h-entry that says it's deleted is better
#tanteklet's start with the real world question - does anyone who supports "deleted" posts keep track of their datetime deleted *separate* from the last time they updated their content?
#aaronpkthe benefit of the h-entry is consumers that don't understand the "deleted" property or HTTP 410 will likely treat it as an update and replace the text
#tantekthat's good for immediate feedback/consideration
#tantekhere's a possible use-case for separate deleted vs updated dates: a separate updated date may provide the consuming code enough information to retrieve the last updated version of the post from an (indie) archive, purely by URL + date lookup.
#tantekhowever, the question is, is that use-case of sufficient interest to anyone besides say, the Internet Archive?
#tanteke.g. would you use that to show/freeze the last version of a comment or reply-context of a deleted post?
#csarven410 is orthogonal to self-descriptive data. The point of the activity streams is to broadcast or dip into the information as to what happened - and this is at the heart of it. 410 on the other hand is only provided when the resource is requested - and this is a good practice.
#tantekI think I prefer an explicit h-entry property as well, rather than trying to fake it with <meta http-equiv STATUS 410 > kind of hack
#rhiaro_Perhaps if you don't want to provide tombstone content - ie. you just want to nuke the post, but still want to track when that happened, dt-deleted will help a reader to know what's going on better than dt-updated - otherwise it'll just update what it displays to an empty post? Or maybe the reader should know not to display anything if there's no content property... </braindump>
#tantekcwebber2: did you find any real world benefits to the added work to support JSONLD? or was it just busywork to satisfy the spec?
#tantekI think there's a lot of resistance to JSONLD by implementers because it seems like a big YAGNI
#tantek(especially how strongly it is promoted by those who are not actually implementing it themselves, but for pseudo-architectural reasons)
#cwebber2tantek: for me it's useful, because I'm planning on working on something a bit outside the normal realm of "blogpost style federation", which is a little game, and I tried to mock it up, and I couldn't figure out how to do the types
#cwebber2I tried also doing a type mapping myself, then realized I was duplicating the expansion work
#cwebber2anyway, I think json-ld is just fine for extensibility, as long as languages have libraries... but python, java, ruby, javascript, php, c# all do
#cwebber2and that's most of the languages people use for webdev
#cwebber2tantek: notably, I think a lot of things are complex; extracting microformats from html is also a bit too much to ask everyone to write a parser for, but one nice thing is, indieweb already provides that stuff!
#cwebber2so there's some inclusion within this sphere, whether it fits as a spec under this umbrella, I'm not totally sure, but I like it, and it very well may.
#tantekcwebber2: cool - well you can certainly work with / on it in # indiewebcamp until/if/when the WG accepts it
#cwebber2tantek: I'll read it again and give it more thought!
#cwebber2okay, awesome, pump.io group scheduled, now to churn through my email, then write some emails myself!
#tantekwow that's hilarious. b, p, bd are voticons
nicolagreco and tilgovi joined the channel
#elf-pavlik!tell tantek, if find interest in Microformats -> AS2.0 conversion, https://indiewebcamp.com/post-type-discovery seems to provide some steps in that direction, it would still need few extra steps! One should have possibility to convert all data published with Microformats into AS2.0 but to my understanding converting from AS2.0 to Microformat most likely will loose information (sort of gracefully degrade)
#elf-pavlikKevinMarks_, I think we still have a long way to go before all participants of this group will start really understand each other and various technologies represented
#elf-pavlikAs well as everyone can define one's own custom vocab for snowflakes ;)
#elf-pavlikURI based vocabs recommended by AS2.0 also allow converging terms over time, while non unique string based vocabs like Microformats as of today besides possible collision of indepedently defined properties/types also don't provide path to disover conversion of terms
#elf-pavlikURI based vocabs simply take advantage of 'follow your nose' to for example find out that give term merged into another one http://schema.org/supersededBy
#aaronpkyou know I have never actually seen any of these vocab namespacing URLs lead to an actual web page. most of the time they are either unresolved domains or 404 pages.
#aaronpkhappy to be shown one that actually goes somewhre
#cwebber2maybe someday someone will square zooko's triangle in a way that doesn't require people to download a never-ending growing blockchain and then they will point me to it
#aaronpkall the computers are going to stop working then anyway, then we can all go home
#elf-pavlik"here are a growing group of organizations that have pledged responsibility to ensure the operation of this website. These organizations are: Digital Bazaar, 3 Round Stones, OpenLink Software, Applied Testing and Technology, Openspring, and Bosatsu Consulting."
#tantek!tell elf-pavlik you provide the FOAF Person URL as an example to aaronpk of a vocab URL that actually shows something in a browser, and yet, it provides a good counter example of your assertion that URL based vocabs allow/encourage convergence, because if they did, FOAF Person would have been collapsed back into vCard (prior vocab, more standard). Your faith in URL based vocabs does not bear out in practice.