#social 2015-07-21

2015-07-21 UTC
jasnell, bblfish, tilgovi, the_frey, the_frey_, jaensen, jaywink and melvster joined the channel
#
melvster ok so i think no one in the indieweb community really cares about indexicals
#
melvster in that case, i propose using a default indexical of #this
#
melvster ok proposed it on the mailing list
#
melvster it's kind of an auto correct for http, much like how browsers will correct <br> and <br/> tags to mean the same
Arnaud, pfefferle, pfefferle_, jasnell_, bblfish and tilgovi joined the channel
#
sandro REMINDER-- no meeting today, happy summer
#
Loqi REMINDER has -1 karma
shepazu_ and bblfish joined the channel
#
ben_thatmustbeme lol
#
ben_thatmustbeme REMINDER++
#
Loqi REMINDER has 0 karma
melvster joined the channel
#
wilkie mattl cwebber2: I leave portland tonight! but I'll probably walk around today since the pass was free. heh.
#
cwebber2 wilkie: oh! if you can come near the convention center I'd be happy to meet up for coffee or the like
#
cwebber2 wilkie: I am mostly free-floating at oscon today
#
wilkie ah cool
#
cwebber2 I'm heading down there now
#
melvster aaronpk++
#
Loqi aaronpk has 898 karma
#
melvster for putting your url in your github account
#
melvster amazing how few people do that
#
mattl wilkie, cwebber2 -- just landed, head over to OSCON.
#
mattl Heading
bblfish joined the channel
#
wilkie mattl: I'm heading there too. but I'll be there in like an hour :/
bblfish joined the channel
#
mattl wilkie: expo seems dead until 5pm
#
ben_thatmustbeme melvster: yes, we do that for rel-me-auth
#
cwebber2 wilkie: mattl and I are here
#
cwebber2 wilkie: SMS me when you get close or etc
#
Loqi Pelf made 1 edit to [[Socialwg]] https://www.w3.org/wiki/index.php?diff=85169&oldid=85050
bblfish joined the channel
#
aaronpk good afternoon
#
aaronpk mattl: cwebber2: up for breakfast tomorrow? I'm thinking at The Original
#
KevinMarks melvster: is #this actually an id on the page you are linking to?
#
melvster KevinMarks: yes
#
melvster it's modeled on the 'this' keyword in javascript
#
KevinMarks OK. are you suggesting that everyone should add such an id to their page?
#
KevinMarks if you just want a fragment for rule14 reasons, just use plain #
#
KevinMarks if you use #this, anyone runnning fragmention will have it scroll to the first occurrence of 'this' in the page
#
KevinMarks which is probably not your intent
#
melvster KevinMarks: it depends what you want to do, if you want an entity that is distinct from the page, in general an indexical how it is done, #this is an emerging pattern, as it emerged as a special keyword in some programming languages, there's not a real consensus right now ...
#
KevinMarks if # isn't enough of a signifier for you, use #webidindexical or #cecinestpasunepipe or something, not a word that will occur in a lot of pages and cause collisions with fragmentions
#
melvster KevinMarks: it's not a question of what I want, it's a question of naming style ... if you have a document say, cv.doc about you ... sometimes you'll want to make "cv.doc" the filename and "you" the person, distinct identities
#
KevinMarks I know the background
#
KevinMarks I'm saying don't append short fragments arbitrarily, as you are going to link to hinkgs you don't intend to
#
melvster what's the reasoning?
#
melvster that's how it's done in js, java, php objects etc.
#
KevinMarks http://www.w3.org/2000/09/dbwg/details?group=72531&public=1#Carvalho links to you
#
KevinMarks it's nto how URLs work
#
melvster KevinMarks: I think I get what you are saying, but the intention is to link to those URIs
#
KevinMarks http://www.w3.org/2000/09/dbwg/details?group=72531&public=1#this links to the occurence of "this" in the 3rd para
#
KevinMarks if you are just appending a fragment to say "I don't mean the page I mean the person" use #
#
melvster KevinMarks: that's what facebook do, and I used to do it also, but then timbl pointed out that it breaks some things called qnames : https://en.wikipedia.org/wiki/QName
#
melvster so some software will break
#
melvster KevinMarks: it's a good idea, just that qname thing is an issue ... annoying
#
KevinMarks how does qnames break this? they have colons not #
#
melvster because you cant have a qname without a right hand side apparently
#
melvster * dont ask me why *
#
KevinMarks OK, so in that case if you are defining an arbitrary fragment to append to mean 'link to the base URL but mean the main thing referred to by the page" I'd suggets you pick a relatively long one that is not going to cause false matches
#
KevinMarks or at least not often
#
KevinMarks whereas #me or #this is going to match a fair few actual id's and a huge amount of text with fragmentions
#
melvster if #me is there, problem solved
#
melvster it's only when there's no fragment at all do we have to 'guess'
#
melvster which is a surprising number of cases
#
melvster if someone later adds #this, even better!
#
KevinMarks for eg, look at http://www.kevinmarks.com/microformatschema.html#this
#
KevinMarks that links to the para starting 'For the sake of daftness'
#
ben_thatmustbeme wouldn't #this if there is an id=this on the page, still have to guess if they mean the block of text or the "this" of the semantics?
#
KevinMarks quite
#
KevinMarks and fragmentiosn make this happen for amy page that contains the text 'this'
#
melvster KevinMarks: i see your point, but that was quite a rare false positive
#
KevinMarks no
#
KevinMarks if you use #thecontainerforthethingcontained or something, it will be rare
#
KevinMarks if you use one of the most common words in the english language it won't be
#
ben_thatmustbeme this is why i don't think #anythingatall will really solve the issue. Its all about context in which the url is given anyway.
#
ben_thatmustbeme best bet is to look at the actual item in the page and look for its id (if there is one)
#
melvster yes
#
melvster that's what we do
#
ben_thatmustbeme so then whats the proposal with #this ?
#
ben_thatmustbeme is confused
#
melvster just in the case that there is no fragment at all
#
ben_thatmustbeme if there is no id, you cannot assume anything at all, which is why it really makes more sense to just look at the context of the usage
#
melvster yes, you have to make an educated guess
#
melvster like with spell checking
#
ben_thatmustbeme so really the proposal should be, "if you want to be able to clearly reference the person the page represents, make sure you set an ID
#
melvster sometimes you're going to get it wrong
#
ben_thatmustbeme no "#this will always mean the semantics"
#
ben_thatmustbeme you are better leaving the ambiguous and you can always go back to the UI level and ask "did you mean you like this person or this page?"
#
melvster #this (informally) means what "this" page is "about"
#
mattl aaronpk: is that near our dorms?
#
ben_thatmustbeme #this would not mean any part of this page with id="this" then any text with the word "this" if you happen to be using fragmentions or perhaps you mean "what this page is about"
#
ben_thatmustbeme its still ambiguous, so it doesn't help
#
ben_thatmustbeme and yes, ambiguous in all cases, i would say
#
melvster ben_thatmustbeme: let's put it another way, lets say the 'this' keyword didnt exist in javascript ... now someone finds a problem that when working with an object they want to reference the content of that object and asks 'what shall we name it?' ... what would you answer be ... bear in mind there's no right or wrong answer ...
#
ben_thatmustbeme this isn't a programming language we are talking about, $this exists in most languages, yes. URLs are something very very different
#
ben_thatmustbeme if you really want to maintain that style, $this would refer to the container (page), not the contents (semantics)
the_frey joined the channel
#
ben_thatmustbeme what you are referring to is really metainformation about the url, this is where XFN used rel values
#
ben_thatmustbeme which makes a lot more sense then arbitrarily tacking on #something
#
ben_thatmustbeme rel="friend" would clearly mean the semantics, not the page
#
aaronpk It's a bit farther north, like maybe half a mile
#
melvster ben_thatmustbeme: it doesnt really affect XFN because I dont think you guys use fragments ...
#
aaronpk How's this look? http://aaronparecki.com/events/2015/07/22/1/w3c-socialwg-breakfast
#
ben_thatmustbeme that was exemplary, not meaning it conflicts at all... also, all urls, anywhere can use fragments
#
melvster this is true
#
melvster KevinMarks: I think there's a slight confusion, in linked data we're not trying to avoid collisions, we're trying to *create* collisions
#
KevinMarks you are trying to link to the wrong thing on purpose?
#
KevinMarks I thought the whole point of this rigmarole was dismbiguation - now you are saying it is intended as obfuscation?
#
melvster trying to link to the main object in the page, that could be a person, a book, a product etc.
#
melvster ie to break out of silos
jasnell joined the channel
#
melvster in foaf we have the "primaryTopic" relation .. .but many people forget to use it
#
melvster primaryTopic means : "what is this page about?"
jasnell, KevinMarks and tilgovi joined the channel
#
cwebber2 aaronpk: I'm down for breakfast
jasnell, the_frey and bblfish joined the channel