#sandeepshettyaaronpk: though I'll still provide the plaintext markdown in that hidden div
#aaronpksandeepshetty: that's a start. :) I still don't know how I can reliably parse markdown for only certain people's sites. most people aren't writing markdown
#cweiskesandeepshetty, can't you content-negotiate?
#cweiskewhen clients provide an "Accept: text/x-markdown" header, they get the markdown
#aaronpkarticles are still html. I think the long-form stuff is different enough that we can treat it differently in terms of stripping/sanitizing html
#sandeepshettyAlmost all social networks deal in text and external long-form content just gets shared with plaintext snippets.
#aaronpkyep, often as a "preview" of the content, so loss of information isn't as much of an issue
#sandeepshettyif we were to follow a similar model then it looks like we'll end up only syndicating notes some of which are related to "sharing" articles?
#sandeepshettyaaronpk: (I'm on the phone) can you add the discussion about syndicating only plaintext somewhere on the wiki? (maybe /plaintext)
#aaronpksure! I feel like we have some notes on that somewhere already
#aaronpkah it is probably on everyone's individual project pages
#sandeepshettyaaronpk: that leaves us with the issue of syndication other types of content like photos. We might have to revist the activity streams stuff again maybe
#sandeepshettyI obviously haven't thought this through so let me start by saying why I think sending a ping is good.. from the perpective of the author it's good to know when my content is linked to...
#sandeepshettyfor a majority of users that shouldn't be 1000's of pingbacks
#cweiskehow can I distinguish normal links from bookmarks?
#willnorrisaaronpk: what at IndieWebCamp still needs sponsoring? Lunch/Dinner on either day?
ShishKabab joined the channel
#aaronpkwillnorris: still need one lunch covered (or we can do lunch on your own at food carts one day) and so far we don't have dinner either day covered
#willnorrisGoogle will pick up dinner Saturday night
#sandeepshettycweiske: That's the problem (Re can differ in different contexts) with generic solutions (like bookmarking software)
#sandeepshettywith our indieweb stuff I think we have very specific use-cases
#willnorrismostly +1 to tantek's point. The fact that I am no longer is any actual AS implementation (either personally or at Google) is the main reason I try and stay out of current discussions
#willnorriserr, *involved in any AS implementations
#sandeepshettytantek: it's the legacy arguments that interest me... I'm hoping there are some lessons there.
#sandeepshettyfor example with RSSB I have in-reply-to, like and repost at the same level while with AS they have a comment object
#cweiskeanother context is that bookmarks are only once in the system, and can't be bookmarked twice. when that's the case, voting is the only way of expressing that the link is useful
#willnorrissaying that AS has zero dogfooding my spec authors is certainly unfair. James Snell uses it extensively at IBM. I know Monica *was* using it extensively at socialcast, though I can't speak to what she's doing now
#sandeepshettyyeah I've seen that :) just curious about why the AS folks didn't just add a reply verb (replied to a post) and have this weird indirection of "posted a comment with a inreplyto to a post"
#willnorriswhat would the object of the activity be if the verb were "reply"? The new comment, or the thing being replied to? Either way, where would you express the other one?
#tantekif you want to document examples of corporate dogfooding perhaps we can learn from that
#sandeepshettyI see what you mean. I guess the difference between what I'm trying is embedding the info in the "new" object and not an abstract activity..
#willnorristantek: that's at least more accurate, though I don't agree with the statement that you must be selfdogfooding, as opposed to corporate-dogfooding in order to have something relevant to say with regard to "present implementation / standards / building-blocks"
#tantekthis feels like a lot of rehashing of existing issues
#tantekif there's a new issue, sandeepshetty - please add it to the end of that list
#tantekwillnorris - if you can show me a counter-example, where only corporate dogfooding has helped a standard actually pragmatically advance, I'd love to hear about it / see it documented
#sandeepshettytrying to figure out if it's an issue or if I'm plain wrong :)
#tantekuntil then, I'm on the selfdogfood or go home team.
#tantek(my experience has been that corporate dogfooding = 2-3 year lifespan tops, until the corp gets bored or moves on or forgets, employee turnover)
friedcell joined the channel
#tanteksandeepshetty - document it as a possible issue then - and indicate your uncertainty :)
#tantekless so with selfdogfooding - that seems more sustainable
#willnorristhough exactly to your point, Chris and I no longer work on Activity Streams at Google… the folks working on activity related stuff (albeit mainly under schema.org) for both G+ and Gmail are newcomers to the space /shrug
#willnorrissandeepshetty: yeah, that makes sense… just having the object. But having a lone object like that is simply interpreted as having an implied activity with the verb "post". So you're really right back where you started, albeit with less syntax
#sandeepshettyright but that "post" is the difference between user posting the object vs the system posting an activity object... that post (verb) is also there with the activity?
#melvsterJames Snell is doing some great stuff at Activity Streams lately...
#melvsteri see his work as a basis to unite the whole social web in 1-2 years from now
#willnorrisin an implied activity, the actor is unknown unless it can be inferred from the object, so it's not really interpreted as the system posting the object… it's just unknown
#melvsteralthough istr you need to identify it some other way in that case
#melvsterblank subjects is a multi year topic of debate
#willnorrisXRD doesn't *really* have empty subjects. The <Subject /> element is optional, but only so that the logical subject can be identified by other means
#willnorristhe only real use case at the time was for host-meta, which used an <hm:Host /> element. No one really cares about that use case anymore though, since webfinger doesn't use host-meta
#melvsterwillnorris: yeah that's what I meant, but in the xml validation it's allowed to be empty i mean
#melvsterFor quite a long time I was against them, but then I got persuaded
#melvsterimho, when in doubt, always provide a subject
#sandeepshettywillnorris: I see what you mean... I guess the difference is more fundamental then: in my case it's an object with some context (verb) vs. with AS it's an activity that describes an object and it's context. Question is will I encounter issues by not having the abstract activity?
#willnorrisan implied activity is different though, in that it really does have an unknown subject. An activity with no actor still makes sense and can still be useful (though maybe not as useful as if the actor were known). An XRD document without knowing what it is that you're describing it meaningless
#willnorrissandeepshetty: where are you identifying the verb? Are you letting it be implied by the fact that the object as an in-reply-to, or are you making it explicit (in which case, I would assume that you have an actual activity object, no? even one without an actor?) ?
#sandeepshettywithin the h-entry (object) I have an p-author/h-card (subject) and links in the h-entry are marked up with verbs (in-reply-to, repost, like)
#willnorrisare those links to perform those actions or to get a list of historical activities of those types?
#willnorrisin other words, are they for future or past actions?
#willnorriserr… sorry. I think I misinterpreted what you meant
#sandeepshettythe fact that the h-entry was posted implies that the link was liked/reposted/commented on
#willnorrissemantically, this maps to AS nicely. syntactically, it deviates a bit to the point that an AS parser probably wouldn't recognize, though I do like the simplicity of it
#willnorrisnot off the top of my head, but to Tantek's point earlier… don't take it from me, as I'm not actually implementing this stuff myself :-D
#sandeepshettyI'm already seeing some issues around how to interpret the content of h-entry
#sandeepshettyfor example you ignore it for likes and repost but you use the content for replies
#melvsterim going to start implementing activity streams 2.0 quite soon
#willnorriswhy is that a problem? Not all actions necessarily need content.
singpolyma joined the channel
#sandeepshettyjust realized that the problem I had in mind was because of someone else's incorrect implementation... you're right that should not be problem.
#sandeepshettyon the rw list I can't understand most of things they are saying because of the jargon.... I need like a one page jargon index :)
#melvsterthat's good feedback ... id like to try and make it more accessible
#sandeepshettythanks that would be great.. I've always had interest but never really got into it... I would love the get your "it's real easy" perpective (assuming it's not coming from the I've crossed this jargon hurdle so it's simple for me now
#sandeepshetty!tell cweiske: IMO, send a webmention when a link is bookmarked but don't send one for voting because it seems like an action on "your" software rather than on the link that was bookmarked. Does that make sense?