#indiewebcamp 2013-01-30

2013-01-30 UTC
#
aaronpk
is now reading pingbacks in IRC as they come in in real-time
#
tommorris
aaronpk: is that going into a channel? ;-)
#
aaronpk
on my private irc server yea
#
aaronpk
going to get pingbacks to indiewebcamp.com showing up here too, but I need to fiddle with Loqi first
#
tommorris
should really use indiewebcamp as his public todo list.
shaners joined the channel
#
tommorris
just got to log on to indiewebcamp.com from another computer and used the incredibly sexy indieauth process.
#
aaronpk
I like the "got to" part of that
#
tommorris
I mean, it's not quite as sexy as West Ham winger Matt Jarvis in his briefs, but within the limited scope of web authentication experiences, I approve.
#
tommorris.org
edited /User:Tommorris.org (+184) "adding personal site todo"
(view diff)
scor and zztr joined the channel
#
tantek
tommorris - nice
bnvk joined the channel
#
donpdonp
higher praise has never been heard.
mxuribe, bnvk and lmorchard joined the channel
#
@benwerd
Seems like there's significant (good) overlap between #pcloud and #indiewebcamp.
#
@Johannes_Ernst
@benwerd: Seems like there's significant (good) overlap between #pcloud and #indiewebcamp.” Yes! Connect us?
andreypopp, lmorchard, bnvk, zztr and Stevef joined the channel
#
zztr
are all thee indie web guys aware of the 'personal cloud' guys? I just got back from an event with them and I can't help but notice philosophical alignment.
andreypopp joined the channel
#
@BarnabyWalters
RT @Johannes_Ernst: “@benwerd: Seems like there's significant (good) overlap between #pcloud and #indiewebcamp.” Yes! Connect us?
tantek joined the channel
#
zztr
hey that's my line
#
zztr
i said it at the thing. someone must have heard me
#
zztr
which is the whole object of saying things, actually
bnvk and eschnou joined the channel
#
tantek
zztr - do any of the 'personal cloud' guys actually create, run, depend on their own dogfood?
#
tantek
philosophy is one thing, action is another.
#
zztr
tantek: personal cloud tech is early, but it looks to me like it's being built because it's what they want to use. I do appreciate the dogfood indieweb philosophy much
#
zztr
a few 'personal cloud' projects: kynetx.com connect.me respect.io
donpdonp, friedcell, tilgovi, Stevef, andreypopp and adactio joined the channel
#
tommorris
spots an adactio.
#
tommorris
Greetings.
#
adactio
Greetings
mxuribe joined the channel
#
tommorris
really ought to do some personal-site-hacking today.
andreypopp, stereoket, Stevef, scor, tilgovi, tantek, indiewebcamp-vis, bnvk, elioat, aaronpk, zztr and barnabywalters joined the channel
#
tantek
zztr, "a few 'personal cloud' projects: kynetx.com connect.me respect.io" - none of those seem to be used by their creators on their own domain (lack of self-dogfooding), and in addition, the first two appear to be centralized and silo-like in architecture - not sure of any evidence of anyone using respect.io on their own domain.
#
tantek
so I'll ask again, do any of the 'personal cloud' people/creators actually create, run, depend on their own dogfood? I'm pretty much ignoring projects that don't meet that bar.
#
tantek
until something is useful/interesting/critical enough for its creator to be using it and depending on it (preferably daily), it's not worth bothering with IMO.
eschnou, bnvk and tantek joined the channel
#
ianloic
tantek, connect.me is more of a repuation service iirc
bnvk joined the channel
#
tantek
ianloic - why would I trust a silo as a reputation service? the web is my reputation service.
#
zztr
tantek: gotta run, but to some degree you're quite right, but arguments can be made. i'm not going to make such arguments, because they're not my projects. But there's enough there to keep me interested.
#
tantek
zztr - I guess perhaps I'm a bit more cynical, having seen too many such optimistic projects just flounder and fail, typically because the creator(s) didn't actually have the passion/commitment to put their own identities on the line behind them in real world use.
#
tantek
so yeah, there's tons of open source projects out there. I say focus on the ones where the creator(s) are living and breathing them every day on their own domains. everything else is just a science project.
#
tommorris
if it isn't actually being used on the web, it's not a solution to the problem of publishing on the web.
#
ianloic
tantek, I don't think it'll work, but I don't think that "the web" works well to connect reputation over degrees of separation either.
#
ianloic
tantek, at least it's paying for my friends to screw with Erlang :)
#
barnabywalters
ianloic: the way I would represent reputation on the web is via XFN within h-cards with p-category set for the various areas with which I associate that person
#
barnabywalters
combined with suitable parsing/crawling code, a reputation graph could easily be constructed
#
tantek
barnabywalters is right. and if you need to consume that information, just crawl it, use one of the available open source microformats2 parsers, and cache it as you deem fit.
#
ianloic
barnabywalters, could be, but to my knowledge hasn't been. and to the best of my knowledge there hasn't been enough reputation information published to be useful.
#
tantek
age the data and refresh as necessary
#
barnabywalters
I already tag people within my address book and publish XFN+h-card, I should add p-category
#
aaronpk
isn't google a reputation system?
#
tantek
google caches the reputation system of hyperlinks
#
tantek
google itself isn't the system
#
ianloic
anyway, I hope these silo approaches beget a more open distributed system
#
aaronpk
yea that's what i meant
#
tantek
ianloic - from past history of such approaches I share no such hopes at this point.
#
tantek
maybe they'll get some papers written on reputation systems and hosted behind paywalls
#
ianloic
yeah
#
tantek
silo thinkers gonna silo
#
ianloic
I mean, my whole experience of the social web started with advogato which was built primarily as a repuation system. And it never went anywhere...
#
tommorris
problem with all reputation systems is they are gameable
#
tommorris
a banned troll on Wikipedia managed to Facebook friend large chunks of the senior Wikipedians, which meant Facebook's UI inherently gave them reputation
#
tommorris
because when you get a friend request from someone who has five people you trust as friends, you sort of trust them too
#
tommorris
the UI explicitly says something like "Friends with: X, Y, Z" on the friend request page
gardnr joined the channel
#
tommorris
after I worked out they were a banned troll, I went to all those people and said "look, you being friends with this guy means others trust that endorsement and give that person access to private information"
#
ianloic
I think a big problem is that they're more interesting to game than to use.
#
tommorris
if private data is in the mix, getting your endorsement system (especially if it is implicit) wrong means horrible ghastly things like giving stalkerish trolls access to information that could, say, allow them to stalk people IRL, out them, etc. etc.
#
tantek
I treat FB friending kind of like a newer more popular LinkedIn in that regard
#
tantek
only accepting/adding people I've met and determined are non-creepy
#
tantek
for exactly the reasons you provide tommorris
#
tommorris
this, incidentally, is why every big company doing social networking ought to have a panel with people who have been stalked, subject to threats of violence, or spent time living in the closet, as a privacy advisory panel.
#
tantek
perhaps even someone who's lived in an oppressive regime.
stereoket joined the channel
#
tommorris
indeed. religious, sexual and ethnic minorities who have lived in unfree societies where revealing themselves, their beliefs or activities would lead to execution.
#
tommorris
I'm presuming people saw http://actualfacebookgraphsearches.tumblr.com/ the other week. ;-)
#
ianloic
devil's advocate here: why? are we supposed to be focusing on the 80% rather than trying to build the perfect thing for every fringe use-case?
#
barnabywalters
tommorris: yeah, that’s great
#
tommorris
Pareto's principle is wonderful. but I'd strongly argue that "the user not being murdered by roving bands of religious fundamentalists" is a non-fringe use case.
#
ianloic
the vast majority of people aren't as concerned about privacy as people who have experienced stalking, threats of violence, etc
#
ianloic
tommorris, I sure hope it's a fringe use case. If it's a non-fringe use case we have bigger problems than social network privacy settings :-/
#
tommorris
in which case, it's a use case anyone building social web stuff needs to consider highly even if it goes against Pareto's principle.
#
tommorris
"Islamic men interested in men who live in Tehran, Iran". That's why Facebook scares me.
#
tantek
yeah - saw that in the graph searches example thing/post
#
tommorris
also, the fact that users don't prioritise their own privacy is not a reason that developers shouldn't prioritise their privacy.
#
tommorris
I'm sure if you took a survey of hairdryer users of things they wanted in a hairdryer, not electrocuting them wouldn't be something most people would explicitly state as a preference.
#
tommorris
that doesn't mean that we don't rightly insist that hairdryer manufacturers comply with electricity safety standards.
#
tantek
ianloic - pareto principle doesn't work when it comes to representing human diversity
#
tantek
hence we were able to get a much better "gender" property added to vCard4 (and thus now h-card)
#
barnabywalters
general rule: assume that "live a long, safe, happy life" is one of your user’s goals
#
tommorris
barnabywalters: if that isn't one of their goals, they'll find a way of adapting the software to their own life-shortening, unhappiness-inducing and safety-reducing ways without the developer helping them. ;)
#
tantek
and to be clearer, human existential diversity, not necessarily the diversity of human creative output - very different things
#
ianloic
tantek, gender is an easy one - for the majority of us who identify as male or female adding flexibility doesn't create any disadvantage to us[*], other steps we might take to support the diversity of humanity might.
#
tommorris
well, the nice thing about the web is that individuals and groups can just decide that the current way of representing humanity sucks, take it and change it
#
ianloic
[*] actually if you're engaging in business with other companies they'll often want to know demographic information, especial genders. That provides a strong incentive for companies to demand a self-identification as male or female
#
tommorris
don't like how microformats does it? use microformats plus your own custom properties.
#
tommorris
don't like how schema.org does it? roll your own properties
#
ianloic
tommorris, you can do that, but that doesn't make it useful
#
ianloic
tommorris, free-form structured information is a bit of an oxymoron.
#
scor
ianloic: until others follow the same practice and use the same custom properties :)
#
tommorris
well, it's more, you keep the stuff you agree with, and mix'n'match the stuff you don't
#
scor
(which then slowly become less custom)
#
barnabywalters
at least with microformats there’s a change your changes will become “official”. No such luck with schema.org
#
tommorris
tantek: on pages like that on microformats, can one document things like official government standards?
#
tantek
research and contributions welcome (for interested-in, looking-for, relationship-status, and any other similarly challenging fields)
#
tantek
tommorris yes on the page with "-formats" appended to it
#
tommorris
tantek: a while back, Wikimedia UK did a survey that asked "How do you identify your gender?" and "How do you identify your sexual orientation?" rather than "What is your gender?" and "What is your sexual orientation?"
#
tantek
not just can one, but it is encouraged
#
tommorris
As a result, I looked into standard demographic survey questions
#
tantek
and if there isn't enough to create a new page, feel free to just create a "Previous Formats" section on the page
#
tommorris
and the UK Office of National Statistics have freely available standards that define how to ask demographic questions in a sensible way, that's backed up by research
#
tantek
tommorris - I'm still pretty amazed we managed to get a properly (research based) decomposed "gender (bio/identity)" property into vCard4
#
tantek
which means it will then flow downstream into all other contact schema
#
tantek
well except for #schema . org folks who just make up their own crap
#
tommorris
imagines the IETF probably wouldn't be amenable to radical queer activists. :)
#
tantek
tommorris - it was a long battle of building up support from various engineers and various companies until we had a strong representation on the VCARDDAV mailing list
#
tantek
and publicly documenting feedback
#
tantek
(that page still shows up (to me) as the top result when you google for "vcard4")
#
tommorris
so, for demographic survey questions, which a lot of websites ask, I put this together - https://meta.wikimedia.org/wiki/Survey_best_practices
#
tantek
based on that feedback I was able to write up and get a bunch of support for, we have this:
#
tantek
never really saw myself as a "radical queer activist" but perhaps that describes the task accomplished. now y'all can point to an RFC ;)
#
tommorris
tantek: no, not you. ;)
#
ianloic
more of a "radical humanity activist"? vcard is perfectly able to represent non-queer genders :)
#
tantek
tommorris - that's an interesting survey guide - perhaps add it to http://microformats.org/wiki/gender-examples
#
tantek
thanks ianloic :)
#
tommorris
tantek: done
#
tommorris
it wasn't so much for gender/orientation but also for things like question ordering in surveys, age bracketing, anonymisation
#
tommorris
so for instance, you might have a small local Wikimedia organisation with 10 members. only one of them is female. you separate out gay and lesbian into two separate categories on your survey, and you set up your age brackets in the wrong way and you not only know you have one female member, but that you have one 15-year-old lesbian member. because of lack of anonymisation and shitty survey design
#
tommorris
because volunteers who haven't had training in statistics and so on may be writing the surveys, it seemed sensible to say "hold on, governments and big companies have worked out how to do this in a more optimal way"
#
tantek
makes sense
zztr, andreypopp, gardnr, bnvk, tantek and mxuribe joined the channel