sandrostrugee, thanks for your interest, but the meeting are only for people in the group (or by special invitation of the chair, if there's a relevant topic)
aaronpk... i ended up spending a lot of time preparing other projects for releases and preparing talks, but did not advance the stuff in the group itself
aaronpk... i was looking at which features had been implemented in which reports, but realized that amy had a fantastic script to generate that automatically
aaronpk... two things: if there are parts of the vocab that are unimplemented, that people feel cannot be left behind, my review of that table is that there was nothing on there that I wasn't willing to see drop off
tantekKevinMarks: except that the people here on the call are about the same as recent telcons at the old time - so actual data of attendance disputes your assertion of survivorship bias
aaronpkevanpro: my main concern is i don't want to leave this open forever. i would like to work quickly to receive implementation reports from activitypub and web annotation implementations to show that we meet the threshold for the properties they depend on
aaronpktantek: the one question i wanted to ask is i can't tell if the green X's mean publishing or consuming or both and i find that pretty concerning
aaronpk... for testing activitystreams, as we discussed before, for something to count, a consumer has to do something interesting rathre than just parse it and do a syntax transformation
aaronpkben_thatmustbeme: : there was a field in the implementation report to indicate this, they don't split it by term but it's for the overall implementation
aaronpk<aaronpk> i definitely remember this discussion and that's why it's taken me so long to build the test suites and reports because i'm specifically testing both sides
aaronpkevanpro: we need to document that we have publishers and consumers on each side of the equation. we should have at least a yellow marking around the ones that are published but not consumed, etc. even if there are multiple publishers.
tantekalso, I'd say 2+ consumers per property should be required, otherwise we're just testing a monoculture (whether you can publish to *one* implementation that consumes it) which is also not a standard
aaronpktantek: the requirements that we're putting in place for AS2 is that the summary indicates whether each term is published or consumed, and that we get impelmentation reports from the annotations folks with that level of detail
aaronpk... the point is if you only have one implementation then you don't really have a standard. i would have a hard time saying that's vetted by implementations
aaronpkevanpro: i don't think we're making assurances that the properties they need are part of the spec. however we can make the assurance that if there are properties they need that are in the spec that have implementation reports that that will be part of the spec.
sandroPROPOSED: We will retain, unchanged, in AS2 everything for which we get 2+ implementations reports (possibly setting the bar at 2+ producers and 2+consumers)
aaronpkevanpro: this isn't necessarily moving the goalposts. we may have resolutions not reflected in the exit criteria, so that would be the communication issue here
aaronpkRESOLVED: We will retain, unchanged, in AS2 everything for which we get 2+ implementations reports (possibly setting the bar at 2+ producers and 2+consumers)
rhiarofrom plh about process: Besides updating the transition requirements to accommodate Process 2017, I also updated them to make sure we're applying Process 2015 correctly.
cwebberbtw, if someone's interested in seeing something I've been working on which is vaaaaguely related to activitypub / federation, connect in firefox: http://dustycloud.org:8888/
aaronpkright the issue is that there needs to be at least one publisher and one consumer in order for anything to be considered interoperable. what i was hearing is trying to make it okay for the 2 implementations to both be publishers.
sandroaaronpk, I completely understand your point, but lost on that many times in many WGs over the years. People were comfortable with "we're producing data, and someday it'll be consumed". So I'm used to just shrugging, at this point.
rhiarobut also it's kind of a vicious circle isn't it. it's not like our audience right now feels big enough to assume we definitely cover what people need. If there are two publishers, that seems to indicate it's useful to two people/projects
ben_thatmustbemethe point is to get implementation experience, so its pretty obvious that 2 producers and no consumers is pretty much useless. at the same time, i see the arguement for having 2 consumers, as one comsumer means that its verified as consumable by an application, not as generically consumable. same thing with producers
ben_thatmustbemeif we change it to 2 producers and 2 consumers of each, it would definitely effect which rows are green, in a quick look through, it would effect 10 entries
ben_thatmustbemeits correct that we did not put those requirements on in the first place, but I think its certainly something we SHOULD strive to have. otherwise it feels like we aren't doing our jobs
tantekso I'm pretty upset and surprised that anyone here would be claiming that *less* than 2 implementations (i.e. only *one*) of either a publisher or consumer, or either a client or server would be sufficient to exit CR.
tantekat a minimum, anything less than 2+ implementations IMO is very much against the spirit of what a CR is supposed to be for, which is to demonstrate implementation interop experience
tantek!tell rhiaro in your AS2 impl report summary, can you separate / cluster columns for impls by editors, wg members, outside of wg, similar to the way Webmention impl report summary has? https://webmention.net/implementation-reports/summary/
tanteknot a requirement to exit CR, but I think it demonstrates just how mature a technology is and how much interest there is beyond the folks immediately working on it
tantekand since meaningful consuming implementations are harder to build anyway, that's *probably* good enough. I would be surprised by any feature that had 2+ consumers but not 2+ publishers
tantekto be clear, I'm ok with going to PR with that as long as we can demonstrate that impls in the WG are actually *different* / independent (e.g. not just sharing libraries with multiple impls)
csarvencarry any weight? At the moment, we are not capturing that by merely looking at the implementation counts, regardless of their granular categorisation (in/outside the group). Case example: WA pointing at AS2. While I do see value in having legit interest and proof of work through implementations (whether that's 1 or 50, that's still arbitrary), it should be on a case by case basis or at least be in consideration of other dimensions that may be at play which
csarven could essentially prove the spec's worthiness. Not all tech is strictly here and now, nor meant to be the end of exploration or innovation. A lot of both W3C specs and non-specs out there played a role in building up on new ideas. Sometimes leaving those threads around so that they are a little bit more visible for the next cool thing to pick up from is important. :)
csarvenIMO, implementation counts alone may not say much about its fitness or appoval for wider use. It is arbitrary in a sense because it only reflects on the reality that the spec was visible/accessible to the implementor and they had the means to implement. It may be useful to broaden the scope of what constitutes a spec that's worthwhile to have. For instance, if other specs and development have built up a normative dependency on the spec at hand, should that
strugee hey all
@t #100DoPP d25: Updated @W3C @SocialWebWG Post Type Discovery ED https://tantek.github.io/post-type-discovery/ Fix #13 #16 #18. Proposing WD tomorrow (ttk.me t4mq1) (twitter.com/_/status/831407018921570308)
Loqi Rhiaro made 1 edit to [[Socialwg/2017-02-14]] https://www.w3.org/wiki/index.php?diff=101322&oldid=101001
ben_thatmustbeme i cannot scribe, i'm sorry, things going wrong at work
RRSAgent logging to http://www.w3.org/2017/02/14-social-irc
KevinMarks -1 to new telcon time, (survivorship bias)
tantek rhiaro: any chance you can regenerate http://as2.amy.gy/reports.html with latest reports?
rhiaro 1. Updating a Candidate Recommendation with only editorial changes: See https://services.w3.org/xslt?xmlfile=https://www.w3.org/2005/08/01-transitions2017.html&xslfile=https://www.w3.org/2005/08/transitions2017.xsl&docstatus=cr-editorial-tr Note that your publication request must assert that changes are only editorial.
aaronpk just updated the list of meetings https://www.w3.org/wiki/Socialwg#Next_Meeting
sandro Loqi, tell tantek I found the minutes: https://www.w3.org/wiki/Socialwg/2016-01-12-minutes#AS2_Exit_Criteria
csarven carry any weight? At the moment, we are not capturing that by merely looking at the implementation counts, regardless of their granular categorisation (in/outside the group). Case example: WA pointing at AS2. While I do see value in having legit interest and proof of work through implementations (whether that's 1 or 50, that's still arbitrary), it should be on a case by case basis or at least be in consideration of other dimensions that may be at play which