#indiewebcamp 2012-10-23

2012-10-23 UTC
donpdonp joined the channel
#
tommorris
I can't edit the wiki.
#
tommorris
!tell tantek I can't edit the wiki!
#
Loqi
Ok, I'll tell him that when I see him next
tantek joined the channel
#
Loqi
tantek: tommorris left you a message 31 minutes ago: I can't edit the wiki!
tantek_ joined the channel
#
tantek
tommorris - looking into it
tilgovi and dascher joined the channel
#
aaronparecki.com
edited /Main_Page (+42) "/* Where */ update venue to say Esri Portland"
(view diff)
dascher joined the channel
#
aaronpk
!tell tantek here is a post based on tonight's conversations! http://aaronparecki.com/2012/296/article/1/creating-content-on-the-indie-web
#
Loqi
Ok, I'll tell him that when I see him next
tantek joined the channel
#
Loqi
tantek: aaronpk left you a message 29 minutes ago: here is a post based on tonight's conversations! http://aaronparecki.com/2012/296/article/1/creating-content-on-the-indie-web
#
tantek
reads
#
@aaronpk
Some fun discussions about today's @AWSCloud outage: http://news.ycombinator.com/item?id=4684384 #indieweb #ownyourdata (http://aaronparecki.com/2012/296/note/1)
#
tantek
Nice article and summary of our discussion aaronpk!
#
aaronpk
thanks! Hopefully it will be useful for the FSW discusisons later this week!
#
tantek
yeah, we need to figure out a constructive way to deconstruct "activity streams" as they are currently framed
#
tantek
I'm liking the idea of leaning towards posts with just different sets of attributes
#
tantek
as a more incremental way forward
#
tantek
than the actor-verb-subject model
#
tantek
*rather than
#
aaronpk
yes, me too. It seems like most of the content I publish online will fit that new model
#
tantek
so I think maybe that was AS's mistake. they pushed a more complex (richer) model than was absolutely necessary for the folks who actively cared about implementing
#
aaronpk
"over-abstraction" as you called it
#
tantek
the AS model was based on what publishers appeared to be doing, but when said publishers showed little interest in implementing AS, it should have been scaled back. drastically scaled back.
#
tantek
I know that one of the "use cases" that drove the verb model was "liking" / "favoriting"
#
tantek
another one was for more passive activities like "listened to a music track"
#
tantek
perhaps those should have been pushed to a v2?
#
tantek
or is there way of at lease squeezing in likes/favorites into the posts model?
#
tantek
just as a checkin can be a textless note with venue information
#
tantek
how do you post a "favoriting" of a URL?
#
aaronpk
The way I would prefer to do that on my site is to have a completely separate feed of objects that I've "liked", "favorited", or "listened to" which is more like the bookmarks I currently publish at http://aaron.pk/bookmarks/
#
tantek
whereas I'd like to have the ability to have a composite feed of everything
#
tantek
on a single timeline
#
aaronpk
The major differentiator being content I've created vs consumed
#
tantek
"favoriting" is creating a single bit of content
#
aaronpk
"favoriting" is really just noting that you've consumed a particular piece of content and find it particularly interesting
#
tantek
and there's also "bookmarking" or especially "bookmarking with tags" (is that just "tagging" ?)
#
tantek
in my own anecdotal querying of users of what does "favoriting" or "liking" mean to them - they're not very consistent about it
#
tantek
except for - it means I clicked the icon
#
tantek
"favoriting" means I clicked the star icon
dascher joined the channel
#
aaronpk
My inclination is to have a stream of things I've consumed, and I would tag each item with my action towards that item, "favoriting/liking" being things I like, "saving" for things I want to come back to, "bookmarking/collecting" for things I'm organizing into sets
#
tantek
"like" means I clicked the "thumbs up" or "heart" icon
#
tantek
some people think of it as "saving"
#
tantek
others think of it as giving props
#
tantek
since the action is public
#
tantek
and it increases a public number on/about the item
#
aaronpk
currently services don't make the distinction the way I want to express an action towards the item. So on Twitter I sometimes hit "favorite" for things I like and sometimes because I want to go back and read it again later
#
tantek
being liked/favorited
#
tantek
yeah - the meaning is inconsistent
#
tantek
so it devolves to just meaning "clicked the star"
#
tantek
I'm not even sure users mean the same thing when they favorite a tweet vs. favorite a flickr photo (even though both use a star)
#
tantek
Nor "liking" a FB post vs. an IG photo.
#
tantek
where both use the term "like" but the icons are different "thumb up" vs. "heart"
#
tantek
I think there was a study somewhere that showed people clicked heart icons more than star icons regardless of the verbage (favoriate or like)
#
aaronpk
I believe that
#
aaronpk
Regardless of the UI, I would like to mark content I've consumed into three different categories: favorite/like, save, bookmark. Bookmarks would have tags.
#
aaronpk
and that is based on the way I interact with current services, and the awkwardness/inaccuracies of the UIs and actions in them currently
#
zztr1
what about 'read later'? is that 'save'? I find that the 'read later' check box on pinboard is very conducive to cleaning out my open browser tabs
#
aaronpk
"read later" and "save" are similar for me
#
tantek
yeah people differ on those
#
zztr
not for me. if i bookmark something without bothering to tag it that's essentially a 'save'. I can find it if i'm looking for it but i won't ever be reminded without remembering
#
tantek
I think the public/private uses are different too
#
tantek
I "save" plenty of URLs privately either as a "I read this" (personal logging) or as a "Read it later" or even as a "Reprocess this later and incorporate into a note/article/wiki page"
#
tantek
publicly I tweet about URLs with a bit of text that typically amounts to a set of tags
#
tantek
for me the "verbs" are quite different both between and among the public and private use cases
#
tantek
and that's *just* for "saving/bookmarking" URLs!
#
tantek
nevermind other content types
#
tantek
I do like aaronpk's distinction between creating and consuming
#
tantek
that dichotomy resonates a bit with me
#
tantek
active vs. passive
#
tantek
checks http://indiewebcamp.com/checkin to see if anyone else has contributed to it.
#
aaronpk
i'll add a link to my post
#
aaronparecki.com
edited /checkin (+160) "add link to my blog post about content types including checkins"
(view diff)
#
aaronparecki.com
edited /User:Aaronparecki.com (+115) "/* related blog posts */"
(view diff)
#
aaronpk
I'm heading to bed, got an early start tomorrow!
#
aaronpk
goodnight everyone!
#
Loqi
night night
#
zztr
night aaronpk
#
zztr
tantek: did you see my response to your question? https://plus.google.com/104255768651498070358/posts/GEsaqFUZT9M
#
tantek
hmm - hadn't seen that
#
tantek
I don't have any expectation that IIW is good for getting any kind of "message across"
#
tantek
I think posting on your own site is a better way to "get the message across"
#
zztr
heh. if people read your site
#
tantek
last year at TPAC I actually had the opportunity to question reliance on DNS with Tim Berners-Lee at lunch - it was kind of an epically fun conversation. Kevin Marks was there too.
#
tantek
zztr - that's what syndication, in particular, POSSE, is for
#
tantek
for now "fixing DNS" is kind of a "too hard" problem
#
tantek
until you want to solve some *really* interesting peer-to-peer crypto/trust/discovery problems
#
tantek
which seems *possible* (i.e. not impossible), but certainly very hard
#
tantek
however, once done, has a lot of potential
#
tantek
I'm ok with letting other folks more into crypto/peer-to-peer try to solve that
#
tantek
for now I'm happy focusing on existing indieweb publishing/syndication problems
#
zztr
yeah. i see that it's "kind of a too hard problem" but i still want to talk about it. and from the perspective of augmenting, not replacing.
#
tantek
you'll definitely have no shortage of people wanting to "talk about it" at IIW
#
tantek
IIW is good for talking, not much for doing
#
tantek
which brings me to your next para
#
tantek
re: "I want to see a shift toward technologies of integration and storage as a means to data ownership." - I think the shortest path to that is you building and deploying and depending on it yourself.
#
tantek
DIY, IndieWeb, self-dogfood
#
zztr
i'm not a p2p advocate. i do think p2p is boxed too often as ideological, but it's not my drum to pound.
#
zztr
yup. i'm doing that
#
tantek
awesome
#
zztr
just finished crawling my bookmarks. just reaching alpha stability with my library for personal data management BaseParadigm
#
tantek
zztr - what's your personal URL?
dascher joined the channel
#
tantek
awesome
#
tantek
got a few rel=me in there
#
tantek
but why not all the links to your other profiles?
#
tantek
could use some hCard
#
tantek
(looking at the first URL)
#
tantek
is traviswellman.com a static single page? or are there other pages there? posts with permalinks?
#
zztr
i just don't spend a lot of time on it. certainly welcoming constructive criticism. not a lot of visitors either
#
tantek
btw I agree with: "That essentially means the entirety of the data has to be outside such services."
#
tantek
that's a key component of POSSE
#
zztr
tantek: glad you do. that's important and overlooked
#
zztr
i will look into POSSE
#
zztr
this is the first time i've heard of it
#
tantek
you've found the right set of folks for this :)
#
tantek
*a* good set of folks at least
#
zztr
:)
#
tantek
I can't demonstrate uniqueness obviously, nor is it necessary to do so ;)
#
zztr
POSSE is more of a guideline than a protocol or spec huh?
#
tantek
regarding: "I don't see people talking about this enough. I want to find more people who are, or who will listen." - you found #indiewebcamp - welcome and hope to see you here more often!
#
tantek
POSSE is a methodology
#
tantek
the wiki calls it a syndication model
#
tantek
it was necessary to distinguish it from PESOS because people on the broader interwebs a) kept confusing the two, b) didn't understand how hugely different they were in terms of implications, data ownership, reliability, fragility, etc.
#
zztr
do you remember talking in Oregon at the camp about citation formats with Nicholas?
#
tantek
I think I even have a citation for that
#
zztr
:) wasn't sure if the nick zztr represented me well
#
tantek
a-ha!
#
tantek
is connecting instances of humans
#
tantek
did you see my follow-up to that session on the microformats wiki?
#
zztr
so effing cool you have twitter.com/t btw
#
tantek
oh it's almost a historical accident that I have @t
#
zztr
about microformats. do you know of any thought/work towards canonicalization so that identical micro documents would hash the same?
#
zztr
in that's one of two problems on top of my mind right now. equality is difficult, but necessary for connecting/integrating a sharded graph
#
zztr
so if two people cite the same paper but write the microformat by hand, and a third party (say archive.org) crawls both sites among others and tries to link up references... i suppose having the same permalink in there would be mostly sufficient...
#
zztr
but possibly not for all microformats
#
tantek
I don't think equality is necessary for connecting/integrating a sharded graph
#
tantek
yes, URL equivalency is likely sufficient
#
tantek
but then again, we won't know until we get implementations parsing publishing examples, and won't get that until we have a simple enough spec to use
#
zztr
another q, do you know of any efforts to link specific functionality to specific microformats? like, a catalog of rest apis that are looked up by microformat?
#
tantek
there have been a few efforts started (or just proposed maybe?) for microformats for APIs
#
tantek
but I'm not sure there's been much followthru put into actually doing and documenting research
#
tantek
there's been tons of "wouldn't it be cool if" type questions/proposals for various microformats but nearly all of them lack actual follow through
#
tantek
which in my mind means they lack actual motivating use cases
#
tantek
and/or real world examples
#
tantek
still - people can document research on any of those they like
#
tantek
we can switch to #microformats for that
#
zztr
have you used the 'share' functionality of Android?
#
tantek
since we're kindof off the indieweb topics
#
tantek
I've watched people use the 'share' functionality
#
tantek
and been particularly confused when that's what you have to do to email something
#
tantek
have you seen my rants on share?
#
zztr
no maybe i should
#
zztr
anyway that's kind of what i'm getting at. you encounter a bit of data, and there are things you can do with it. the list of things is filtered based on the format
#
tantek
the list of things should be based on use cases
#
tantek
and user research
#
tantek
not on the format
#
tantek
format driven UI is a terrible anti-pattern
#
tantek
and a really common developer-mistake
#
zztr
yeah you're right 'share' is over overloaded
#
zztr
the problem is that the developer has to include a button for it, and since it's so amorphous, it needs to just be one button. meaning, all content needs to be able to meet up with all consumers through that one 'share' mechanism
#
tantek
"problem is that the developer has to include a button for it" - sounds like the developer's problem
#
tantek
and misframed at that
#
tantek
rather than solving a *user's* problem
#
zztr
i'm glad it's there personally, but it would be better if the one button listed a range of actions, and each action could be implemented by different providers
#
zztr
well... the developer can't plan for the range of potentially installed software on a phone
#
tantek
personally I think it's a tough problem space that's been ill considered by current solutions
#
tantek
so maybe limited options is ok
#
zztr
so it has to be just a single 'do other stuff' kind of functionality
#
tantek
that's typically the case in user interfaces
#
tantek
limited and simpler is better (more usable) than extensible
#
tantek
but programmers always err on the side of extensible
#
tantek
and thus mistakenly create UIs accordingly
#
zztr
ooh. yes, simpler. yes, more usable. no, not better. highly frustrating in fact to me.
#
zztr
but yes i am a developer, and one who very much cares about being able to use all the data in all the ways
#
tantek
the other thing is that you don't need a giant list of actions every time
#
tantek
lots of stuff being needlessly reinvented by android - all the "intents" stuff
#
tantek
it's as if these people never heard of file types vs. applications
#
tantek
invented in the 80s I believe
#
tantek
do you get a giant list of applications everytime you double-click a new document on your system? no.
#
tantek
not even windows 3.1 had that dumb of a UI
#
zztr
hold on. yes you do. you have to associate a file type with an application to open it
#
tantek
no you don't
#
tantek
the system has defaults
#
tantek
that "just work" 99% of the time
#
zztr
yes
#
tantek
ironic that android etc. are a *regression* in that regard
#
tantek
like I said, it's as if the developers ignored all previous related work
#
tantek
they thought there were inventing something new, but they weren't, and instead they just invented a suckier version of existing technology
#
zztr
i hear ya. it's not good enough.
#
zztr
do you think file extensions with default application associations is an ideal solution?
#
tantek
ideal? who care about ideal? how about, best so far?
#
tantek
Mac file types and app types are a decent solution
#
zztr
well i ask because it's not good enough for me, so i wonder what you think would be better
#
tantek
and don't depend on extensions
#
tantek
file extensions are a distant second
#
tantek
"intents" crap is somewhere off in the weeds
#
tantek
I did do some work on improving on the simple Mac type/creator model back in the 1990s
#
tantek
when I was working on OpenDoc
#
zztr
i don't know if you want to be this far into this conversation in this channel this early on a tuesday ... but my thinking leads me in the direction that there are three things to worry about. 1) establishing the type of data 2) establishing what can be done with a type of data and 3) managing implementations of those actions
#
tantek
I think they're orthogonal
#
tantek
but that's much more a #microformats type conversation
#
zztr
yeah
sandeepshetty joined the channel
#
zztr
ok getting some z's now. I look forward to lurking and sporadically ranting in this channel later :)
#
tantek
excellent
#
tantek
welcome and looking forward to your constructive contributions
#
tantek
would enjoy more iteration on the citation microformats discussions and ideas
#
tantek
thanks and good night!
#
Loqi
gute nacht!
#
zztr
i leave you with the factoid that all my 4000 bookmarks crawled one hop off (respecting robots.txt) are about 35 gigs and took about 3 days to download politely
#
zztr
night tantek and Loqi :)
#
Loqi
dude
#
zztr
... over dsl with 4 threads
#
tantek.com
edited /2012/Academic_Citations_Web (+84) "see also - web citations microformats proposal"
(view diff)
#
sandeepshetty
So I've started to write some code for my personal publishing system so that I can move away from Blogger: https://github.com/converspace/converspace/commits/master
#
sandeepshetty
Once I've frozen on the schema (which should be today: https://github.com/converspace/converspace/wiki/schema), I should be able to start publishing...
#
sandeepshetty
Do any of you publish/open-source your indieweb personal publishing system or at least the schema you use?
#
tantek
sandeepshetty - I use the hAtom scheme :)
#
tantek
and my publishing system is Falcon (http://tantek.com/w/Falcon) and I open-source the stable bits of it in the #cassis.js library: http://cassisproject.com/
#
sandeepshetty
Yeah I read about :)
#
sandeepshetty
So I'm guessing you use file-system?
#
tantek
yes - I think the wiki page notes that: http://tantek.com/w/Falcon
#
sandeepshetty
is making all sorts of typos due to lack of sleep
#
sandeepshetty
I haven't seen tagging on tantek.com, guessing it's not implement?
#
tantek
it's not, not yet
#
tantek
I haven't figured out what UI I want to provide for it
#
sandeepshetty
I've just decided to merge categories and tags into channels that allow selective subscriptions and like twitter they can be present anywhere in the post.
#
tantek
that's probably a good simplification to start with
#
sandeepshetty
The file-system approach was awesome for commodity web (shared) hosting. Most of the cloud providers don't providing persistence stores by default these days.
friedcell, tilgovi, tantek, adactio, borior, dascher, spinnerin and barnabywalters joined the channel
#
barnabywalters
adactio: I visited the UK’s second largest device testing lab today. It was a hugely educational experience!
#
adactio
barnabywalters: where was that?
#
barnabywalters
Exeter, in the little attic which is the nimbleworks studio
#
barnabywalters
they have a little table full of devices
#
barnabywalters
and an open door policy
#
barnabywalters
tantek: fab, thanks for the clarification
#
tantek
anytime! it's good to write these down and answer them as we encounter them - you're likely not the only one to come up with such questions, likely just the first.
#
@ngsgene
Patient &amp
#
Loqi
their info stays together! MT @23andMe @GeneticAlliance patient access to clinical lab results: http://blog.23andme.com/news/23andme-supports-better-access-to-health-records/?utm_source=twitter&utm_medium=tweet&utm_campaign=blog #ownyourdata
#
tantek
do we need to start indiegenecamp?
sandeepshetty, tilgovi, BjornW, barnabywalters, friedcell, dascher and friedcell1 joined the channel