#microformats 2011-08-10

2011-08-10 UTC
jrodgers, tantek and nonge_ joined the channel
#
tantek edited /existing-rel-values (-332) "removed rel-lytebox|demo value. Neither of the referenced GRDDL nor WCLR defines "demo" or "lytebox"." (view diff)
voxpelli, csarven, andr3, manu`, Amorphous, Atamido, ChiefRA, Soopaman, mwunsch, voxpelli-laptop and kennyluck joined the channel
#
tantek edited /search-engines (+368) "move defunct search engines to new section" (view diff)
mwunsch joined the channel
#
tantek edited /rel-home (+495) "added example in wild: use rel-home as part of feed discovery from post permalink pages" (view diff)
mwunsch joined the channel
#
tantek so I think I finally found a good use for rel="home"
#
tantek when a post permalink page links to the feed of the *site* for auto-discovery purposes, there's something not quite right about using just rel="alternate"
#
tantek that is
#
tantek alternate implies that the destination URL is an alternate version of the *current* page
#
tantek which it isn't in the case of a post permalink page linking to the *site* feed
#
tantek thus one solution would be to combine 'alternate' with 'home'
#
tantek on such links
#
tantek cc: Kevinmarks
#
tantek I've started doing this on my site - per suggestion to include feed auto discovery on my permalink pages.
KevinMarks, mwunsch and dglazkov joined the channel
#
dglazkov tantek: got a sec?
#
tantek dglazkov - got an opinion on rel="home alternate" as noted above http://krijnhoetmer.nl/irc-logs/microformats/20110810#l-61 ?
patcito joined the channel
#
dglazkov tantek: sure, "home alternate" sounds reasonably intuitive
#
tantek ok, given that actual practical use case for rel-home, I'm going to add it to the registry for HTML5
mwunsch joined the channel
#
tantek edited /existing-rel-values (+307) "added rel-home to HTML5 registry given the actual use case of machine discovery of a feed for a site from a permalink page" (view diff)
#
tantek edited /existing-rel-values (-2) "/* HTML5 link type extensions */ typo" (view diff)
#
singpoly1a @tantek: that rel values (to me) reads: "this other page is *both* an alternate of this page *and* the homepage"
#
singpoly1a I'd rather leave the rel off altogether. discovery should be more based on type="" anyway, for most of the feed discovery use cases
#
KevinMarks singpoly1a: you men an alternate of the homepage?
#
KevinMarks *mean
#
tantek edited /rel-home (+1253) "added suggested use of home+alternate for linking to site feeds from permalink pages to a new brainstorming section" (view diff)
#
singpoly1a KevinMarks: no. tantek is using it to mean alternate of the homepage. It seems more like alternate of current page and is the homepage
#
singpoly1a rel values combine with "and", no?
#
tantek singpoly1a - no
#
tantek not quite
#
tantek e.g. rel="
#
tantek sorry
#
tantek rel="alternate stylesheet"
#
tantek they also combine with media
#
singpoly1a people use alternate and stylesheet on the same thing? what does that mean?
#
tantek e.g. rel="alternate" media="handheld" is how you do auto discovery to the mobile version of a site
#
tantek singpolyma: http://www.w3.org/TR/html4/present/styles.html#h-14.3.1
#
singpoly1a wow, I had never seen that before
#
tantek semantics of rel values combine
#
singpoly1a alright then. I thought rel was more like class
#
tantek so in this case, "alternate home" is the meaning, and the type makes it clear it's a feed
#
singpoly1a right. in that case, cool
#
tantek rel is like class in that it is a space separated set of values
#
tantek but class values are undefined in the HTML spec
#
tantek however rel values are both defined in the HTML spec, and now have an official registry: http://microformats.org/wiki/existing-rel-values
#
singpoly1a so, how does xfn work with that? rel="friend met" mean both friend and met, no? I guess in this case "the kind of friend that I have met" is basically the same semantic
#
tantek right
#
singpoly1a does order matter in rel values, then?
#
tantek no
#
singpoly1a ok
#
tantek per http://www.w3.org/TR/html5/semantics.html#attr-link-rel
#
tantek dglazkov - I've added the brainstorm proposal for use of rel home+alternate here: http://microformats.org/wiki/rel-home#use_with_rel-alternate - please feel free to add your opinion :)
#
dglazkov tantek: :)
#
singpoly1a interesting that the URI for that page is .../semantics.html
#
dglazkov tantek: are you ever in MTV these days?
#
Hixie tantek: dude, don't use the TR/ page, that's always out of date
#
tantek no I'm unfortunately too old to be on MTV which I think has an age cap of ~25?
#
dglazkov tantek: heh. How about the age cap for Mountain View?
#
singpoly1a Hixie: I expect your version doesn't use the word semantics in any URIs? ;)
#
tantek Hixie, noted. :)
#
tantek http://dev.w3.org/html5/spec/semantics.html#attr-link-rel
#
Hixie that's better :-P
#
Hixie (the whatwg.org copy has a few more fixes still, but for most purposes the dev.w3.org and whatwg.org copies are equivalent)
#
tantek cool
#
tantek Hixie, in other news, I found an actual machine-readable use case for rel="home" when combined with rel="alternate" (see above http://krijnhoetmer.nl/irc-logs/microformats/20110810#l-61 )
#
singpoly1a So, that section of the spec sort of implies my initial intuition: " if a single link element has a rel attribute with the value next stylesheet, it creates both a hyperlink (for the next keyword) and an external resource link (for the stylesheet keyword)"
#
Hixie i saw
#
Hixie anyone planning on implementing it?
#
Hixie that's the thing that killed rel=feed :-(
#
tantek not sure, I just started publishing it :)
#
Hixie (lack of implementations)
#
tantek well rel=feed was unnecessary - all of its use-cases were already covered by existing solutions
#
tantek in this case I'm pretty sure I've identified a problem that existing solutions don't solve
#
tantek or I'm welcome to being shown the existing solution :)
#
tantek Hixie - feed discovery implementations are plenty buggy and problematic btw
#
tantek even Google Reader fails to parse link rel correctly (treats it as a literal rather than space-separated set of tokens)
#
Hixie not sure "even" is the word i would there :-P
#
tantek would appreciate if you could ping the Google Reader folks to fix that - I've reported the bug years ago but to no avail
#
Loqi pong
#
singpoly1a most autodiscovery implementations tend to ignore <a> as well
#
Hixie tantek: i think in practice the use case is "subscribe to a feed", and rel=alternate type=rss/atom seems to handle it fine
#
Hixie tantek: the semantics of this are rather dubious of course
#
Hixie tantek: but since when has that ever won the day :-(
#
tantek Hixie, the problem is that rel=alternate type=rss/atom on a permalink page would/should literally mean a feed for *that* page, e.g. a feed of the comments on the post
#
tantek \
#
tantek and some CMS's put out multiple rel alternates for a post permalink now
#
singpoly1a and there is in practise often *also* such a feed
#
tantek and thus it's ambiguous which to use
#
Hixie sure
#
Hixie like i said, the semantics are quite dubious
#
tantek so for the feed readers / indexers *that care*, rel="home alternate" provides a way to distinguish the feed for the *site*
#
Hixie but in practice people seem to be happy with just disambiguiting by title=""
#
Hixie so...
#
Hixie *shrug*
#
singpoly1a A machine can't really diambig by title
#
Hixie anyway, what matters at the end of the day is whether implementations decide this is a good idea
#
tantek yeah, I tend to agree. however, I'm also fine with trying new semantic markup to see if anyone ends up caring years later :)
#
Hixie that's pretty much my philosophy too :-)
#
singpoly1a Hixie: what actually matters is if publishers decide it's a good idea
#
Hixie no
#
tantek oh great Hixie, now you're just feeding the conspiracy folks. ;)
#
Hixie singpoly1a: what matters is whether users end up using it. for that you need publishers and implementors -- and implementors tend to be the harder nut to crack, as there's always far more publishers than implementors.
#
tantek see, they have the same philosophy, they *must* be colluding!
Askarii joined the channel
#
singpoly1a Hixie: if enough publishers implement it, then I can use it. If implementors only have it, that does nothing for me or my users
#
singpoly1a where by "implementors" I assume you mean your usual "the big N browsers"
#
tantek singpolyma - when you say "I can use it" - are you implementing something?
#
singpoly1a tantek: yes
#
Hixie singpoly1a: by "implementors" i mean anyone writing consuming code
#
singpoly1a Hixie: well, if that's all you mean then having publishers is sufficient for small-time implementors to use it in useful code
#
tantek singpolyma feel free to try it on my site, e.g. try it on this post permalink: http://tantek.com/2011/220/b1/web-actions-a-new-building-block
#
Hixie singpoly1a: i.e. anyone writing code to implement the feature in a user agent (i.e. any software working on behalf of a user)
#
tantek (I just added this today because of a bug report that someone couldn't automatically find my feed from one of my posts)
#
singpoly1a tantek: I'll certainly roll support into my various things that do autodiscovery next time I'm working on them
#
Hixie singpoly1a: what matters is whether any small-time implementors or big-time implementors do use it
#
Hixie singpoly1a: there are plenty of features that have been used by billions of pages but never used to any significant extent by implementors
#
singpoly1a Hixie: like what?
#
kevin marks edited /rel-home (+111) "/* use with rel-alternate */ check home in supported browsers" (view diff)
#
Hixie singpoly1a: rel=feed for example
#
Hixie actually that one wasn't adopted by authors, bad example
#
Hixie um
#
tantek singpolyma - in practice yes, if you implemented it, it would make a difference for rel-home in terms of helping it advance from 'draft' to 'specification', as we now require at least 1 implementation (consuming code) not-from-the-author to advance to spec: http://microformats.org/wiki/process#Specifications
#
Hixie singpoly1a: rel=up would be a better example
#
tantek Hixie, did authors ever use rel=up?
#
Hixie yeah, there's plenty of pages using it
#
tantek I wonder what for
#
tantek breadcrumbs?
#
singpoly1a Hixie: I've used userscripts that supported rel=up. also, I think Opera does/did support it
#
singpoly1a though I haven't seen many pages actually supporting it
#
tantek singpolyma - feel free to add any documentation of specific support you know of (e.g. Opera documentation) to http://microformats.org/wiki/rel-up
#
Hixie rel=index is another example
#
Hixie rel=external was the 5th most-used value a few years ago
#
Hixie is rel=bookmark used by any consuming software?
#
singpoly1a so, up/index/next/prev/home are all things I've wanted to use / have used in the past, but the big problem for me has always been lack of publishers using it
#
Hixie that was the 4th most-used
#
tantek Hixie, rel=bookmark was indexed by Technorati looking for blog post permalinks
#
singpoly1a bookmark is used by tons of microformats consuming software (as part of hAtom ot very least)
#
Hixie k
#
Hixie anyway my point is just that just because it's published doesn't mean it's widely used
#
tantek agrees.
#
Hixie er
#
Hixie (used as in implemented)
#
tantek as in consumed/parsed
#
singpoly1a and by "widely" you mean "by the big 4/5" ?
#
tantek singpolyma - if you're asking me, I'd say by *anyone*
#
Hixie singpoly1a: not necessarily the major browsers, no
#
tantek and if you do know of implementations of any specific rel value, then please document/cite them on the wiki
#
tantek in an "Implementations" section on the page for each rel value that you know that is implemented/parsed
#
tantek without documentation of specific implementations, there's no evidence that there is any use/implementation.
#
Hixie singpoly1a: one tool used by 100 people isn't "widely". A few tools each used by a few hundred thousand might be.
#
singpoly1a use/implementation.
#
singpoly1a oops
#
tantek to some extent, even a tool used by a few people would be worth documenting, as that might be a sign/harbinger of further implementation (not always, but sometimes is)
#
singpoly1a Hixie: right. so, if that's the criteria than I agree. good things tend not to get deployed by major corps, and thus not to groups of that size
#
singpoly1a tantek: I'm poking around a bit right now to see if I can find some impls I've used
#
tantek singpolyma awesome
#
tantek and if there's no page yet for the rel values you're documenting, feel free to create it
#
Hixie tantek: certainly the trend is relevant, yes.
#
tantek even if you just start with
#
tantek {{stub}}
#
KevinMarks so rel-home and rel-up are the first two links in a breadcrumb chain
#
tantek == implementations ==
#
Hixie gotta go, bbiab
#
tantek * [URL name] - description, what it does with the rel value
#
tantek etc.
#
KevinMarks hm, or rathe the first and last links in the breadcrumb chain
#
KevinMarks what would you label intermediate ones?
#
singpoly1a up/down
#
KevinMarks <nav><a rel="home" href=''>home</a>:<a rel="???" href=''>2. formats</a>:<a rel="up" href=''>2.1 microformats</a></nav>
#
KevinMarks what's the ???
#
singpoly1a up up ;)
#
tantek KevinMarks perhaps it would be a good time to start documenting a few *real world* examples of breadcrumb link structures so we can analyze them, e.g. on http://microformats.org/wiki/breadcrumb-examples :)
#
tantek singpolyma - no that was rejected
#
KevinMarks down works if I'm on 2.1 and am pointing at 2.1.1
#
tantek rel is a set
#
tantek # of occurrences does not affect the semantic
#
singpoly1a that should go in the spec, then :P
#
KevinMarks strictly should have <ol> around the breadcrumbs
#
tantek it is
#
tantek that's what "set" means :P
#
singpoly1a looks at spec again
#
singpoly1a oh... it uses "set" very colloquially. I didn't even notice
#
tantek https://secure.wikimedia.org/wikipedia/en/wiki/Set_%28computer_science%29
#
singpoly1a Sure, I am aware of what "set" mean formally
#
dglazkov like a tea set
#
dglazkov tantek: have you been following the component model discussion at all?
#
tantek edited /rel-home (+73) "/* Discussions */ link to IRC discussion" (view diff)
#
tantek dglazkov - no idea what you mean - would need a URL for specifics.
#
dglazkov tantek: http://wiki.whatwg.org/index.php?title=Component_Model_Use_Cases
#
tantek ah, the XBL2 conversation
#
tantek no I'm not really following it as it's a huge problem to attempt to solve
#
tantek I'm focusing on simpler solutions for the simple cases in CSS3-UI
#
tantek I'm not even sure that any "component model" solution would be any simpler than just using Javascript libraries for such UI widgets
#
tantek and if it's not simpler for authors, then why bother.
Askarii joined the channel
#
dglazkov tantek: using JS libraries takes you a long way, certainly
#
dglazkov tantek: but the trip always ends in the abstraction layer leaking and the mirage breaking down
#
tantek edited /rel-home (+357) "/* use with rel-alternate */ link to an example in the wild for those that want to see/try it perhaps with an implementation, noted possible implementation support" (view diff)
#
dglazkov as described by http://wiki.whatwg.org/index.php?title=Component_Model_Use_Cases#Widget_Mix-and-Match
#
singpolyma created /rel-next (+614) "list some implementations of rel-next" (view diff)
#
singpoly1a http://help.opera.com/Linux/9.50/en/toolbars.html mentions rel="next" explicitly, and then says it supports some other kinds of navigation links, but doesn't list which ones :(
#
singpolyma created /rel-up (+204) "list some implementations of rel-up" (view diff)
#
Loqi [http://twitter.com/andyrutledge] RT @t: #aea @stubbornella CSS tips (@lukew http://j.mp/stbcss). one error: classes/#microformats *are* used by search: http://ttk.me/t4DG1
#
Loqi [http://twitter.com/smiller] RT @t: #aea @stubbornella CSS tips (@lukew http://j.mp/stbcss). one error: classes/#microformats *are* used by search: http://ttk.me/t4DG1
#
Loqi [http://twitter.com/justtobetrendy] RT @t: #aea @stubbornella CSS tips (@lukew http://j.mp/stbcss). one error: classes/#microformats *are* used by search: http://ttk.me/t4DG1
#
Loqi [http://twitter.com/valerierose] Food Bloggers: What are your thoughts on recipe formats e.g. microformats & hRecipe? cc @wubbahed @foodtechconnect
#
Loqi [http://twitter.com/feastie_dot_com] Food Bloggers: What are your thoughts on recipe formats e.g. microformats & hRecipe? cc @wubbahed @foodtechconnect
#
tantek Thanks Loqi for the tweet mentions of microformats. And while I appreciate the RTs, I think you can leave them out.
#
Loqi you're welcome
mwunsch joined the channel
#
Loqi [http://twitter.com/t] @tabatkins that's not how rel works. already covered in IRC http://krijnhoetmer.nl/irc-logs/microformats/20110810#l-84 (ttk.me/t4DG3)
#
tantek edited /rel-home (+774) "/* use with rel-alternate */ added FAQ regarding misconception that rel values act independently (since now two smart people have asked the question)" (view diff)
#
singpoly1a Tantek: good to know I count as "smart people" ;)
#
tantek now only if we could folks to discuss things on IRC instead of Twitter (ahem, tabatkins)
#
hober you could awlays /invite TabAtkins #microformats
#
tantek hober, I tried
#
hober ahh
#
tantek then I pm'd him
TabAtkins joined the channel
#
TabAtkins tantek: The sole example of interacting rel values is "alternate" and "stylesheet". In every other case, rel values are independent. Thus, "home" and "alternate" shouldn't combine either.
#
tantek well I'm going for a run. but for now, in short, in reply to https://twitter.com/#!/tabatkins/status/101399365033271296 - there are of course non-sensical rel value combinations - that's what happens when you combine predefined values. and this is also a key outcome of how rel is different from class, since the keywords are defined and registered, deliberate can be taken to use them in combination and define special meanings when u
#
tantek to do so.
#
tantek deliberate *care
#
TabAtkins tantek: I'm not talking about nonsensical combinations. I can imagine situations in which "next stylesheet" makes sense.
voxpelli joined the channel
#
TabAtkins Similarly, I can imagine situations in which "home stylesheet" make sense. If you throw an "alternate" in there, does it apply to one or the other, or perhaps both?
#
TabAtkins tantek: Combined processing without expliciting grouping is a bad thing. "home-alternate" gives a combination rel value that's completely unambiguous.
#
tantek it has to be defined, but that's the point, because rel values have to be defined anyway, combinations are open to definition, which is what we're doing with "home alternate" because it's useful (has a practical machine-readable use case)
#
tantek disagreed - explicit hyphenation of potential combinations of values leads to too many values.
#
TabAtkins I understand the use-case of defining a feed for the homepage (separate from the feed for the current page). I disagree with your attempt to assign meaning to rel combinations, which is unprecedented aside from a single example.
#
tantek unprecedented aside from a single example = there's a precedent
#
tantek and a well established one
#
tantek thus it's reasonable to follow it
#
TabAtkins tantek: "Single precedent" and "legacy mistake" sound awfully similar.
#
tantek instead of your general "dislike of"
#
tantek if you think "alternate stylesheet" is/was a mistaken then make that point
#
tantek rather than appealing to "you don't like it"
#
TabAtkins I pointed out a technical weakness of combined-rel semantics.
#
tantek theoretical weakness that doesn't seem to have affected "alternate stylesheet"
#
tantek so we're fine there
#
tantek as KevinMarks noted, it would be interesting to see what existing rel="home" implementations do
#
voxpelli TabAtkins: eg. http://tools.ietf.org/html/rfc5988.html#section-4 is on your side - but not sure if I personally agree
#
TabAtkins tantek: Yes, because there's only one single example of it. This is the sort of problem that gets worse with scale. If every combination is potentially meaningful, it becomes more difficult to avoid meaning conflicts as you add more meaningful combinations.
#
voxpelli tantek: Did you see my tweet about the semantics of alternate in Activity Streams auto discovery?
#
tantek KevinMarks' comment: http://microformats.org/wiki/rel-home#use_with_rel-alternate
#
tantek voxpelli - did not - URL?
#
tantek TabAtkins - no one ever said design was easy
#
tantek ;)
#
voxpelli tantek: here's the thread on Google Groups: https://groups.google.com/forum/#!topic/activity-streams/XJm-AwQJigo/discussion
#
TabAtkins tantek: Glibness aside, do you accept that the problems caused by rel combinations don't show themselves strongly when there are few valid combinations, but do when there are many?
#
tantek I think you can make that statement about combinations and complexity and problems about *any* space.
#
tantek e.g. CSS border related properties ;)
#
tantek doesn't mean we shouldn't try and do something useful with the simple combinations
#
TabAtkins I don't think the analogy is relevant. Combining orthogonal things is good - the combinatorial explosion works well for you. Combining *non*-orthogonal things may not be.
#
tantek I'd rather try simple things and find out than not try for fear that the complex would fail.
#
tantek in fact that's usually how things work
#
tantek folks say, that will never work because of x complexity
#
TabAtkins I would rather go for a scheme that doesn't have the same technical problems but has equivalent semantics. ^_^
#
tantek and then other folks say (or are too ignorant to know better) let's try this simple thing
#
tantek and then it works
#
tantek combinatorial complexity exists whether you use hyphens or not
#
TabAtkins Frex, defining that "alternate" within a comma-separated rel name has a particular meaning.
#
tantek voxpelli - not sure what the point of that thread is
#
tantek rel="feed" is unnecessary
#
tantek that's been shown already
#
TabAtkins tantek: No, the problem is distinct. rel="home stylesheet alternate" is ambiguous. rel="home-alternate stylesheet" (or one of the other two possibilities) is not.
#
TabAtkins s/comma-separated/dash-separated/
#
voxpelli tantek: The point I was trying to make there was that the "alternate" relation might not be a proper way to link to a stream of related actvities
#
tantek depends on from what page!
#
voxpelli tantek: which was kind of the same issue that you brought up with "home alternate" - "alternate" means an alternate representation of a specific page - not a link to something related
#
tantek from a page that shows HTML updates, then using rel="alternate" to link to the feed or activity stream is correct
#
tantek e.g. from tantek.com to /updates.atom - rel="alternate" is correct
#
tantek but yes, alternate means an alternate representation of *the* page
#
tantek not the site
#
voxpelli tantek: wouldn't rel="related" or similar be more correct for eg. post sites?
#
tantek rel="related" is redundant
#
tantek every link is somehow related
#
tantek what's a post site?
#
voxpelli post page*
#
tantek no
#
tantek because those can have feeds too
#
tantek e.g. of their comments
#
tantek that's note in the brainstorm: http://microformats.org/wiki/rel-home#use_with_rel-alternate
#
tantek *noted
#
voxpelli tantek: of course, but for eg. a twitter stream on a tweet page
#
tantek every page that shows a stream in HTML could have an alternate in AS-JSON
#
tantek regardless of the scope
#
voxpelli of course
#
tantek even a tweet page
#
tantek could have an AS-JSON of the folks that RT'd it
#
voxpelli and that should be a del="alternate"
#
tantek because it's shown on the page
#
singpolyma edited /rel-home (+342) "/* use with rel-alternate */ quote the spec" (view diff)
#
voxpelli The common use case though is: 1. I visit a tweet. 2. I think the tweeter is brilliant and wants to follow him.
#
voxpelli So it was the case of the activity stream related to that specific use case that I was referring to
#
Loqi [http://twitter.com/getwired] @GrumpusNation I wrote up an XML microformat spec a year ago for restaurants. I wish Google would help small restaurants succeed.
#
voxpelli tantek: would you in such a use case use rel="home alternate" type="application/stream+json"?
#
voxpelli or perhaps even rel="author alternate"?
#
tantek there's no such thing as an XML microformat spec
#
tantek edited /rel-home (+665) "/* use with rel-alternate */ noting difference between non-normative examples (what you can't imply from) and what the spec actually normatively says." (view diff)
#
tantek voxpelli what's the use case specifically?
#
tantek i.e. linking from *what* URL to what *other* URL?
#
singpoly1a tantek: thanks for the clarification on the wiki :)
#
tantek singpolyma, np. reading specs is hard. and annoyingly so. I sympathize.
#
voxpelli tantek: From: A page for a user activity. To: A feed containing that users identitiy.
#
voxpelli To: A feed containing that users _activity_
#
tantek I think that's what rel="home alternate" solves
#
tantek assuming their home page shows their activity
#
tantek which, e.g. in Twitter, it does
#
tantek if you have other examples, then we'll have to look at specific URLs
#
voxpelli tantek: Take a multi-author blog - the "home" of the multi-author blog is very different to that of its author feeds.
#
tantek hard to discuss / reason about without looking at a specific example
#
tantek e
#
singpoly1a tantek: rel="index alternate" maybe? (index means page where the activity is, yeah?)
#
tantek no index does not mean that
#
tantek that's what I tried first actually :)
#
singpoly1a ah
#
tantek then I looked at the specific definition of rel-index
#
voxpelli tantek: never mind - you are losing me in thoughts here. If you want to add your thoughts on "home alternate" and such to the activity streams mailing list you at least know about the discussion there now.
#
tantek edited /rel-home (+55) "see also" (view diff)
#
tantek rel-index "Refers to a document providing an index for the current document. "
#
voxpelli tantek: I'm not trying to complicate things - just trying to wrap my head around how to do discovery in a social landscape and using simply rel="alternate" everywhere and nothing else seems like too little semantics
#
tantek voxpelli - I think that's what I was saying (losing me in the example) :)
#
tantek hence I needed a specific example to take aloof at
#
tantek a look
#
tantek damnyouosxlionautocorrect
#
tantek voxpelli - I think we mostly agree
#
tantek I'm just trying to build up the specific proposed solutions based on specific examples / problems to solve.
#
voxpelli tantek: I understand, but there's so many missing bits and different bits in this social discovery space right now that it's easy to get lost.
#
tantek rel="home alternate" solve the blog post permalink page linking to the feed for the blog problem. examples, personal site, like tantek.com, and hosted content, like twitter.com/t
#
tantek if you want to solve a "multi-author blog" problem - then it will help if you can link to a specific instance of one
#
tantek so that we can better understand the structure / links that we're dealing with
#
TabAtkins Does the twitter stream page count?
#
TabAtkins For a mroe traditional example, lifehacker or gizmodo or something like that.
#
TabAtkins reddit is a slightly non-traditional one. lesswrong.com looks more traditional, but is built on reddit.
mwunsch joined the channel
#
voxpelli tantek: on a related question - has there been any discussion on rel="author" in microformats lately? Are the community behinds Googles usage of it?
#
tantek voxpelli, btw I am opposed to rel="activitystream" - I think rel="alternate" type="application/activitystream+json" works fine
#
tantek rel="author" predates Google's proposal/usage btw
#
tantek from my cursory understanding it looks like Google adopting the existing semantic
#
voxpelli tantek: I'm feeling more and more the same - I guess I'm better at asking questions then at finding solutions
#
tantek voxpelli - btw, feel free to invite the activity streams folks to discuss rel values here
#
tantek in #microformats
#
tantek TabAtkins - what's the Twitter stream page?
#
voxpelli will do
#
TabAtkins tantek: The page you get when you first enter twitter after ahving logged in.
#
tantek sighs. you ask for a URL and you get names. will assume ".com" unless otherwise specified.
#
TabAtkins Uh, yes.
#
tantek your logged in twitter home page doesn't have a feed. so therefore not an example for this problem space.
#
tantek whoa - Twitter mentions tab now includes when people favorite your tweets!
#
tantek twitter.com/mentions
#
tantek lol at the timing
#
tantek now goes to: https://twitter.com/your_activity
#
tantek and there's also https://twitter.com/activity
#
tantek TabAtkins, Lifehacker.com is a sad example of a multi-author blog - they don't provide feeds for specific authors :/
#
tantek e.g. on a page of posts by a specific author: http://lifehacker.com/people/Gyroscope352/ it only provides feed links to all of Lifehacker (or Lifehacker "top posts)
#
tantek so, crappy example, nothing we can reason with.
#
TabAtkins tantek: Try lesswrong.com. They provide feeds for individual authors.
#
tantek yup that one looks fine. an author page: http://lesswrong.com/user/KPier/ links to the RSS just for that author with rel="alternate". that's correct. and doesn't bother to link to a global feed (which is fine, and thus, not our problem)
#
tantek and a post by that author http://lesswrong.com/lw/70d/theory_of_knowledge_rationality_outreach/ links to a feed of comments on that post with rel="alternate" - also correct. and doesn't provide a feed to that author as a whole which is again fine, and also not our problem.
#
tantek (this is why we ask for specific examples)
#
tantek (it may turn out that in practice, a problem doesn't exist)
#
tantek voxpelli - so again, if you want to explore the multi-author blog case and possible challenges, try to find actual example URLs of multi-author blogs that have such challenges.
#
tantek (none of the provided examples have the mentioned theoretical challenges)
abki|home joined the channel
#
Loqi [http://twitter.com/yetanotherjosh] @stlhood @lmorchard srsly. I thought he'd start ranting on the evils of microformats. I sensed he hated kittens, too.
#
tantek lol
#
lmorchard Ah, yes, my evil twin was apparently encountered in SF
KevinMarks joined the channel
#
tantek voxpelli - I've added some additional material/links to http://wiki.activitystrea.ms/Autodiscovery based on our discussion in IRC (linked to it)
#
hober makes a sad noise about the demise of rel=feed
#
tantek voxpelli - I still don't understand what specific problem you're trying to solve with rel="author alternate" in your latest message here https://groups.google.com/d/topic/activity-streams/XJm-AwQJigo/discussion
#
tantek hover, feed was too late - authors/publishers and implementers had already adopted rel="alternate" type="application/foo+xml" and that was "good enough"
#
tantek rel="feed" didn't provide enough justification for its existence (significant advantages etc.)
#
singpoly1a I'm not sure what the point of a rel="feed" would be
#
hober yes, it was too late