#indiewebcamp 2012-10-05

2012-10-05 UTC
dascher joined the channel
#
sandeepshetty
Would love to get feedback on http://activitypingback.org/
#
sandeepshetty
it's part of a larger project I'm working on: http://activityweb.org/
dascher joined the channel
#
tantek
sandeepshetty - welcome! what's your personal site?
punkprogrammer, danbri and sandeepshetty joined the channel
#
tantek
sandeepshetty - interesting site - are you using your own domain which is then running Blogger?
#
sandeepshetty
Own domain pointing to Blogger
dascher joined the channel
#
tantek
how does that setup work? we have some folks that are using Tumblr that way, and Wordpress.com
#
tantek
(feel free to add a list item there for Blogger right next to * Tumblr, * Wordpress.com )
#
sandeepshetty
ItCNAME to blogger
#
tantek
are you able to login to indiewebcamp.com with your personal domain?
#
tantek
cc: aaronpk
#
tantek
(wondering if personal domains CNAME hosted with blogger work with IndieAuth)
tantek-ipod joined the channel
#
sandeepshetty
Yes. Thats also my openid URL (thanks to blogger so super easy to remember and use)
#
sandeepshetty
<link rel="me" href="http://www.blogger.com/profile/05858379451002808518" /> <link rel="openid.server" href="http://www.blogger.com/openid-server.g" /> <link rel="openid.delegate" href="http://sandeep.shetty.in/" />
tantek-ipod, punkprogrammer, danbri and sandeepshetty joined the channel
#
sandeep.shetty.in
edited /Getting_Started (+108) "/* Hosted Solution */"
(view diff)
josephboyle, sandeepshetty, Nadreck, dascher and zztr joined the channel
#
@benatkin
When people build their own websites, they get a better understanding of Internet Freedom. Check out http://indiewebcamp.com/Main_Page #Internet2012
dascher, zztr, Nadreck, josephboyle, dietrich, aaronpk, donpdonp, Alphi, hober, lmorchard, mkowens, catsup, reidab, singpolyma, benward, tommorris, jancborchardt, elf-pavlik, wajiii-afk and brianloveswords joined the channel
#
@benatkin
;@zekeweeks @alexisohanian was glad to hear about #indieweb. The name makes it easy to explain. /cc @aaronpk @indiewebcamp #Internet2012
#
@benatkin
@zekeweeks be sure to sign up for next year and visit the IRC! #indieweb on freenode
dascher, danbri, tantek, friedcell, adactio, sandeepshetty, spinnerin, barnabywalters and brennannovak joined the channel
#
barnabywalters
hey sandeepshetty — good to see you here!
#
barnabywalters
I had a look over activitypingback
#
sandeepshetty
Great! What do you think?
#
barnabywalters
overall it looks to be a lot easier to implement than salmon would be/is
#
barnabywalters
the big benefit of using public keys, however, is that you get encryption as well as verification
#
barnabywalters
e.g. how would private messages be implemented using activitypingback?
#
sandeepshetty
I haven't thought this through because I was focusing on public activities like on blogs and twitter (because I think there are other channels for private secure messaging), but at the top of head you could do it with either HTTPS (only if the receiver has an HTTPS endpoint will the sender send the notification) OR a new content-type approach?
#
barnabywalters
if activitypingback is purely for public data, why not just use ordinary pingback? The same amount of requests are involved (sender -> receiver, receiver -> sender to verify content)
#
sandeepshetty
ActivityPingback has richer activity semantics (because of the activity streams payload) than the "someone linked to you semantics" of pingack...
#
barnabywalters
that behaviour could easily be added onto existing pingbacks, and ideally should be expressed by the sending endpoint
#
barnabywalters
e.g. if I favourite one of your posts, I send you an activitypingback…
#
barnabywalters
…with activitystreams data in (even though it’s an ordinary XMLRPC pingback)
#
barnabywalters
if you support the AS data, you can verify it
#
barnabywalters
if you don’t, it just counts as a link, and so legacy blogging applications can interop too
#
barnabywalters
even the HMAC auth could be piggybacked onto XMLRPC pingbacks
#
sandeepshetty
I see what you mean. I've always cringed at the XMLRPC based interface with it's hex return codes :)
#
barnabywalters
sandeepshetty: sure, me too :) but it already exists and is *really* widely supported
#
barnabywalters
like, every wordpress blog uses it!
#
barnabywalters
at the very least I plan on using it as fallback for whatever more advanced protocol becomes popular
#
sandeepshetty
The added semantics require a little more security than what Pingback provides... to avoid things like replay attacks
#
sandeepshetty
Senders should be able to verify that the "Activity" originated from them... I'm guessing it could still be done via regular pingbacks and if the receiver gets an AS, it can be pretty sure that the sender support a verification endpoint..
#
barnabywalters
which is another good reason to send a dumb ping and let the receiver crawl the useful information off the page if it wants to
#
barnabywalters
I am increasingly leaning toward that approach, actually
#
sandeepshetty
The nature of activites (lots of them) means that not all info will be "on page" and crawlable...
#
barnabywalters
sandeepshetty: then I would ask "why not", and/or "why is a <link rel="alternative" type="application/json"> version being supplied"
#
barnabywalters
add .json to the end of the URL (and pardon the incorrect mime type, it will be fixed in a min)
#
barnabywalters
the only reason all the extra data isn’t being exposed in the HTML is there is no agreed serialisation for AS data in h-atom yet
#
barnabywalters
and each activity has it’s own URL, so no need to load all the activities and have to crawl them
#
sandeepshetty
Yeah I was just thinking that..
#
sandeepshetty
makes sense..
#
barnabywalters
is fixing that MIME issue
#
barnabywalters
fixed. Should be nice JSON data (soon to be actual activitystrea.ms JSON data) now
#
sandeepshetty
barnabywalters: I'm going to things about just piggybacking on Pingback and get back to you...
#
barnabywalters
sandeepshetty: fab. if you’re up for it we should build a prototype
#
sandeepshetty
Yep. What's the _list in the json?
#
barnabywalters
in the meantime I’ll have a look over AS-serialisation in HTML (tantek? you mentioned this previously, I believe?)
#
barnabywalters
sandeepshetty: residue from the framework I built. I refer to the view the activity’s being rendered in as list
#
barnabywalters
then the different versions for content types are list.html.php, list.atom.php, list.json.php…
#
sandeepshetty
I was actually going to write to tantek about that (saw point 4 here http://tantek.pbworks.com/w/page/21743425/Falcon)
#
barnabywalters
ah, that’s where I saw it mentioned!
#
sandeepshetty
BTW, is the publishing platform you're using/develeping open source?
#
barnabywalters
sandeepshetty: sort of. I’m against releasing the entire thing (it’s not a product, I don’t really think people will want to use it as it is), but I’m releasing the useful parts of it
#
barnabywalters
or, more accurately, I’m developing external modules for use with it and releasing them
#
sandeepshetty
(Y), BTW, that isn't a valid json activity stream right?
#
barnabywalters
not yet :) Haven’t made it valid yet due to lack of consumer apps and time. Now the spec is almost finalised I’ll see if I can make it valid
#
barnabywalters
at the mo it’s just a JSON representation of what’s in my DB
tilgovi joined the channel
#
sandeepshetty
barnabywalters: Time to go back to the drawing board.. be back soon...
#
barnabywalters
nice talking to you. goodnight (in the UK)!
#
Loqi
night
#
sandeepshetty
Way past midnight for me (in India) :)
#
aaronpk
sandeepshetty: I'd be up for hacking on the pingback thing too, did you already see my Ruby version? https://github.com/aaronpk/Pingback
#
sandeepshetty
I'm thinking of changing my spec (https://github.com/converspace/activity-pingback/) to just extend existing pingback
#
aaronpk
that would be awesome!
#
sandeepshetty
Just adding you guys to the github prject team..
#
sandeepshetty
what's your github username?
#
aaronpk
aaronpk
#
sandeepshetty
Do you know barnaby's username?
#
sandeepshetty
It's not on his website: http://waterpigs.co.uk/
#
tantek
scrolls up
#
sandeepshetty
yeah found it :D
#
tantek
great discussion about extending pingbacks!
#
tantek
sandeepshetty - awesome that you're publishing your thoughts in progress on all this stuff
#
tantek
and welcome to #indiewebcamp :)
#
tommorris
I think the pingback stuff is probably too complicated.
#
sandeepshetty
tankek: Thanks :)
#
tommorris
Very simply put, all you need is "hey, this resource <a> is about resource <b>"
#
sandeepshetty
aaronpk: Done. Can you check if you have access now?
#
tommorris
all the other stuff should be discoverable from the URLs. ideally, in the HTML, but conneg-able too. ;-)
#
aaronpk
sandeepshetty: yep, looks like it worked
#
tantek
tommorris - that's not all you need, as Trackback was essentially that and just got spammed to uselessness
#
tantek
it's one of those cases where RDF reasoning fails quite quickly in the real world of spammers and other bad actors ;)
#
tommorris
wait, <a> is about <b>. okay, go have a look at <b> and see if there's actually a link back from <b> to <a>.
#
tantek
tommorris, and now you've essentially described pingback
#
tantek
(with that extra step)
#
tommorris
well, that's up to the recieving server whether they follow that.
#
sandeepshetty
tantek: I don't have an answer for the spam problem except for moderation
#
tantek
not really optional, in that if you omit it, all you get is spam
#
tommorris
some people might want spam. ;-)
#
tantek
not really " up to" that is
#
sandeepshetty
You can still get spam with Pingback
#
sandeepshetty
The spammer just chooses to host the content...
#
tantek
if you just want spam, then just publish your unverified HTTP referrers
#
tantek
sandeepshetty - it's a higher barrier for spam though
#
aaronpk
also see http://aaronparecki.com/2011/213/article/1/indieweb-messaging for an interesting approach to combating spam
#
tommorris
Other than the presence of XML-RPC, I kind of quite like Pingback's simplicity. Wondering if we could do Pingback2 which would eliminate that and just use multipart/form-data instead
#
tommorris
part of the problem is also making the UI nice. compare Wordpress/Moveable Type pingback to something like Tumblr's notes
#
tantek
tommorris - feel free to write it up! or replace the XML-RPC with JSON
#
tantek
for the modern kids ;)
#
tommorris
meh, don't even need JSON
#
aaronpk
I'm not a huge fan of XML-RPM either, but once you abstract it out you don't even need to care that it's XML anymore
#
aaronpk
RPC* oops
#
tommorris
you just need to POST back the two URLs to the server
#
aaronpk
I wonder why pingback used xml-rpc in the first place... form-encoded would have been just as good
#
tommorris
because back then XML-RPC was hot shit.
#
tantek
" The problem with this is that you don’t know if what you are posting could irk FB."
#
aaronpk
exactly the reason to not use JSON right now :)
#
tommorris
well, also YAGNI
#
tommorris
there's no point adding extra cruft to handle JSON when the only fields you actually need are "source" and "destination"
#
tommorris
and you may not even need "destination", that should be discoverable by parsing. ;-)
#
tommorris
I'm getting very close to splitting off my 'posts' and my 'notes'.
#
tantek
tommorris, I think it's more like this: MetaWeblogAPI used XML-RPC, TrackBack used RDF (in model, philosophy, and implementation) and got spammed, Pingback was designed as a replacement in the context of being another "standard" blog API, and since the existing such standard used XML-RPC, it did too, for consistency with that API model.
#
tantek
not a matter of "hot shit", just that XML-RPC seemed like the way to build successful openly standard blog APIs back then.
#
tommorris
I hate to go all No True Scotsman, but you can't blame RDF for not-checking-for-spam. ;-)
#
tantek
tommorris - you kind of can, because most RDF discussions have an implicit assumption that you can just trust all the triples
#
tantek
hence why I'm blaming RDF philosophy
#
tantek
hahaha
#
tommorris
not at all. the RDF people worry about trusting things, they just hide it behind a lot of academese.
#
tantek
not really. I mean, yes they worry about trust, but they don't really do anything about it in practice, nor in the promotion of the use of the technology
#
tantek
it's that mythical upper layer in the seven-layer RDF semantic burrito
#
tantek
looks for the infamous diagram
#
tommorris
I think that was put in the blender long ago. we now drink linked data smoothie.
#
sandeepshetty
aaronpk: indieweb-messaging messaging is pretty neat especially the factoring problem.. just needs discovery (link header or link element) and it be easy for anyone to implement..
#
tantek
and I was right that Trust was at the top
#
aaronpk
oh my god i thought you were kidding
#
tantek
someone should make a parody of that graphic where all the labels are burrito / taco ingredients
#
aaronpk
just needs a little cheese
#
tantek
aaronpk ^^^ perhaps "where all" should not trigger the memeification
#
aaronpk
heh, yea it's a pretty broad regex
#
aaronpk
ah--they call it a cake?
#
Loqi
ah has -1 karma
#
tantek
and just to make it worse, yes "Figure 1. The Semantic Web layer cake: the technology stack for the W3C Web architecture"
#
tantek
not just any cake
#
tantek
the W3C Web architecture
#
tommorris
layer cake. it's yummy.
#
aaronpk
gives loqi a Semantic Web layer cake
#
Loqi
thanks, aaronpk!
#
tommorris
regretfully locates an XML-RPC Ruby gem and implements Pingback.
#
tommorris
Ah, stdlib.
#
tommorris
I was gonna do it per-URL
#
tommorris
as in, if you are responding to /posts/47, you send your XML-RPC responses to /posts/47
#
aaronpk
ah yea. I didn't want to integrate pingback into all my sites individually so I set it up as a service
#
tommorris
but that's helpful
#
@t
"The problem is you don't know if what you are posting could irk FB"/G+/Tw http://www.thecoolhunter.net/article/detail/2126/when-facebook-disables-your-fan-page #ownyourdata #indieweb (ttk.me t4LJ1)
#
tantek
tommorris - FYI as of 2009 TBL was still including the 7-layer semantic burrito diagram in presentations: http://www.w3.org/2009/Talks/0120-campus-party-tbl/#(14)
#
tommorris
sure, and it's still broadly applicable.
#
tantek
(though maybe it's a historical reference, hard to tell without seeing the talk itself)
#
tantek
disagreed. any model that puts trust at the is missing it
#
tantek
at the *top
#
tantek
you have think about trust / dependability at all levels
#
tantek
have *to
#
tommorris
given my current exposure to Semantic Web stuff at the moment is building storage stuff for data that is inherently trustworthy (it's official and governmenty), I don't really keep up with what people are saying about trust on the open web. it tends to be that if you get a bunch of data from gov.uk, you kind of trust the source. ;-)
#
tantek
tommorris, except when it *is* wrong, and then impossible to correct, because of such assumptions :(
#
tantek
(insert Brazil / Hitchhikers Guide kind of scenarios here)
#
tommorris
I think trust in the broad sense is something people have to think about at all layers
#
tantek
I'm not even sure you can ever have "trust" per se with data, all you can have is a verifiable provenance chain, and then the application has to surface said provenance (citations) to the user and the user has to make a trust judgment themselves.
#
tantek
so yeah, I do question that fundamental assumption of the RDF layer cake
#
tommorris
so, for me it basically boils down to: in RDF land, merging linked data doesn't require application-specific data, because the data model mostly solves that. unique identifiers, open world assumption, additive merges.
#
tommorris
trust basically boils down to "do I merge these two graphs?", and, yes, that's a hard problem. it's in that special class of HS-Hard.
#
tommorris
like NP-hard, but HS-hard: homo sapiens hard.
#
tantek
I think that difficulty (HS-hard) makes it an invalid question
#
tantek
basically, forget about merging graphs, and just keep track of provenance
#
tommorris
but we are merging what are essentially graphs all the time. when you want to see a list of all the murders on a Google Map, programmers are doing elaborate merges by hand
#
tommorris
one of these days, I'll write a long rant about RDF, but today isn't that day.
#
tommorris
need to get back to implementing pingback
#
tantek
sounds like a good prioritization :)
#
tommorris
well, it's as much getting things done as in about a week or so, at work, I'll be eating triples for breakfast.
#
tommorris.org
edited /projects (+174) "/* experimental */ adding my little contribution to this effort"
(view diff)
danbri joined the channel
#
tantek
Tommorris - be sure to add a link to your actual instance of Ferocity running on your personal site :)
#
tantek
in an "Attendees that are using it on their own site:" area
#
tantek
like the other projects
#
tommorris
it's just running on heroku at the moment. i'm hoping to actually point tommorris.org there in the next week or so.
#
sandeepshetty
tantek: Did you make any progress on point 4 (http://tantek.pbworks.com/w/page/21743425/Falcon)? re: figure out how to add Activity Streams semantics to hAtom
#
sandeepshetty
aaronpk: Haven't done much XML-RPC. Do you know if servers complain if they receive more than the expected parameters?
#
aaronpk
Not sure actually, I would assume not.
#
tantek
sandeepshetty - I made some progress, but then realize any additional work on enhancing hAtom should really go into enhancing h-entry instead, since that (microformats2) is the path forward.
#
sandeepshetty
aaronpk: seems like it might be implementation specific though...
#
tantek
ah you're there, good
#
tantek
feel free to ask about microformats in #microformats also
#
sandeepshetty
is rethinking piggybacking on Pingback
#
sandeepshetty
Looks like it's a common recommendation to disable pingback in the wordpress community
#
sandeepshetty
" In short, we have found 99% of all trackbacks/pingbacks to be SPAM. This is the reason why we have disabled it entirely. It is not worth the time to moderate a lot of SPAM like this." -http://www.wpbeginner.com/beginners-guide/what-why-and-how-tos-of-trackbacks-and-pingbacks-in-wordpress/
#
aaronpk
I enabled pingbacks on all my sites recently to see what happened. I do get a lot of spam, but there's also a bunch of valid ones
#
tantek
aaronpk - any patterns to the valid ones?
#
tantek
like are they from URLs that contain representative hCards for example
#
aaronpk
great question! I haven't looked in detail yet
#
tantek
just wondering if there is something we can determine about the non-spamminess by analyzing the referring page
#
tantek
(because spammers tend to be dumb/lazy)
#
tommorris
tantek: so, whitelisting. ;-)
#
tommorris
shame Google Social Graph API isn't working anymore.
#
tantek
tommorris - yeah
#
tantek
worried that whitelisting will lead to too much echo-chamber
#
tantek
if you only accept comments/pingbacks from those you already know
#
tantek
maybe that's ok?
#
tantek
like a living room discussion
#
tommorris
well, the alternative is to broaden it out a bit
#
tommorris
so, n degrees of separation
#
tantek
odds are, you already know everyone you've invited to your living room to have a discussion
#
tantek
like 2
#
tantek
not n
#
tommorris
this is one of the things the closed social networks solve to some extent
#
sandeepshetty
tantek: re:echo-chamber this is what the twitter stream feels like
#
tantek
not really - I get FB spam friend-adds
#
tantek
(in reply to tommorris re closed social)
#
tommorris
on Facebook, if someone friends me and they know 60 of my friends, I'm likely to trust them not to be crazy.
#
tantek
right, that's just pre-caching the XFN in common friends stuff
#
tommorris
and I have friended some crazy people on the basis of this and then gone and told the people whose trust they were also abusing
#
sandeepshetty
No I meant following is type of whitelisting...
#
tantek
right
#
tantek
i'd almost be ok with 2nd degree following I think
#
tantek
anyone I'm following, or anyone they follow
#
sandeepshetty
and you peek at the mentions once in a while (moderating)
#
tantek
hah - mentions are useless for me :/
#
tantek
(on twitter)
#
tantek
either for @t or tantek
#
sandeepshetty
you mean a netwrok of trust
#
tantek
not really a network, just a 2 level tree
#
tantek
followings of my followings
#
tommorris
(we had a crazy banned wikipedia nutter a while back who managed to friend most of the Foundation people on the basis that if they know a few people, they could friend more on the basis of "friends with N people" stuff on Facebook. which given people use Facebook for personal stuff is kind of scary.)
#
tantek
and not really "trust" per se, but more, some degree of approval/reputation
#
tantek
"trust" is too loaded a word
#
tantek
e.g. I dont' trust everyone I follow on all topics, they're just not experts on everything.
#
tantek
(another one of the methodological flaws in RDF - including an abstract/generic notion of "trust" as a building block, when no such thing exists in the real world short of math proofs)
#
tantek
tommorris - yes, I believe that is a form of social engineering
#
tommorris
well, in this case, it's more, "if someone Jeremy Keith says is a rel=friend pingbacks, fuck it, go ahead and post it"
#
tantek
one interesting (social/cultural) aspect of twitter is that they've lowered the energy for social follow churn
#
tantek
that is, adding/removing people from your "blogroll" or other explicit "friends list" (regardless of format) doesn't stay current enough, because there is too much stigma with dropping people
#
tantek
yet on Twitter, following/unfollowing is much less of a (dramatic) problem
#
tommorris
I never really got into blogrolls. there's kind of a scalability problem. ;-)
#
tantek
sure. the only point with blogrolls is that they seemed to organically emerge, so it made sense to add a little markup to them.
#
tommorris
I can follow 2,000 people on Twitter just fine. adding 2,000 people to my blogroll means adding 2,000 links to the sidebar of every blog post.
#
tantek
*asking* people to create / maintain blogrolls was too high of a barrier
#
tantek
Twitter used to show you all your followers / following in the sidebar too
#
tantek
then they showed only a 6x6 grid
#
tantek
then only one row
#
tommorris
then Justin Bieber came along and broke it all
#
tantek
and now, zero, with a link to view them all AJAXy
#
tantek
no I think it was Scoble
#
tommorris
Robert Scoble: the original Justin Bieber. ;-)
#
tantek
for social networks anyway
#
tommorris
contemplates the wisdom of creating a little spider-my-site script that'd go through posts and find XFN links and aggregate them on one "friends" page
josephboyle joined the channel
#
tantek
tommorris - XFN rels in posts are temporily scoped
#
tantek
they were relationships at the time of the post, but likely have some half-life of reliability
zztr joined the channel
#
sandeepshetty
The other thing I like about Twitter is that replies are only visible to you, so spammers have less incentive or they are limited to just "you"
#
tantek
I see *tons* of spam on Twitter
#
tantek
they're not even close to "solving" that problem yet
#
sandeepshetty
You mean @replies?
#
sandeepshetty
and followers right...
#
sandeepshetty
any other type of spam?
#
tantek
well there's Indonesian mentions of my name which I have no idea what they're talking about
#
sandeepshetty
that still @replies/@mention
#
sandeepshetty
which are limited to you.. will blog comments spammers are targetting all your readers/visitors
#
tantek
no it's not replies/mentions
#
tantek
hold on will show you
#
tantek
if you (or anyone) can explain what is going on there, I'd appreciate it
#
tantek
I tried Google translate on "tantek" and it doesn't do anything
#
tommorris
incidentally, icerocket seems like a nice way of getting stuff out of Twitter without the API these days: http://www.icerocket.com/search?tab=twitter&q=from%3Atommorris&rss=1
#
sandeepshetty
haha That seems more like a search problem (like if you search for PHP you'll also get posts with links ending in .php that might be relevant to PHP)
#
tantek
wow ice rocket is still around
#
sandeepshetty
and it's using webintents :)
#
tantek
where?
#
sandeepshetty
* might not be relevant to PHP
#
sandeepshetty
Visiting that link in chrome poped up the web intents dialog for me..
#
aaronpk
it just makes me download a file search.rss.webintents
#
sandeepshetty
"Which service should be used ofr viewing?"
#
sandeepshetty
I'm on ubuntu/chrome
#
tantek
weird - I didn't see anything in the markup
#
aaronpk
os x chrome 22 here
#
sandeepshetty
Maybe its chrome doing something based on content-type?
#
sandeepshetty
which is cool if that's the case..
#
tantek
I for one don't want to see a dialog pop-up when I visit a site
#
tantek
that kind of Kanye-UI has no place
#
tantek
"Imma let you finish browsing that web page, but this dialog box is the most important thing in your face ever!"
#
sandeepshetty
I know what you mean.. but in this case if it renders the content-type better and I can save the preference, I wouldn't mind it
#
tommorris
ooh. just had a thought:
#
tommorris
current situation with separation between web browsing and web publishing : TimBL's vision of a web browser/web editor hybrid application :: current situation with blogging and RSS readers being separate : how it should be.
#
sandeepshetty
tantek: Thats pretty close to what I'm doing with http://activityweb.org
#
tantek
well, it's a hand-waving attempt at describing the problem
#
tantek
it's not really anything built
#
tantek
it's someone realizing they need to do something
#
tantek
they're walking along the same path that many here have traversed
#
tantek
"doing" consists more of actual coding, deploying on your own site, sharing said code (or design or UX)
#
tantek
beyond position-statement blog posts
#
sandeepshetty
I agree. I'm just getting started putting my thoughts down, but you're right, the proof is in the implementation, which I'm working on :)
#
tommorris
I wonder if Hashcash is a potential solution for the Pingback/Trackback spam problem
#
sandeepshetty
goes back to working on implementation
#
tantek
and sandeepshetty - don't get me wrong - blog your thoughts on these things as much as you like! it's useful for us to share thinking even if we haven't figure out how to code what we want
#
tommorris
I think Tantek is saying we shouldn't say "doing" when we mean "talking about"
#
sandeepshetty
the thing I like about the current federated infrastructure that exists is that you can do this incrementally..
#
tantek
agreed. enabling incremental adoption (rather than a monolithic "stack", ahem) is the way to make all this stuff succeed.
#
sandeepshetty
the doing begins with an idea :)
#
tantek
even HTML, HTTP, URL was incremental
#
tommorris
yes, the web had backwards compatibility. <a href="telnet:…"> <a href="gopher:…">
#
tommorris
'twas only recently Mozilla removed the gopher client from Firefox
#
tommorris
on incremental adoption, sure. this is one of the many reasons Diaspora et al. won't work: because they rub straight up against Metcalfe's law.
#
tommorris
Join Diaspora: the social network for people who think distributed social networks are awesome and people geeky enough to run a Rails app.
#
tantek
to be fair to Diaspora, they did implement a bunch of open/federated formats/protocols
#
tantek
IIRC, hCard, WebFinger, Salmon
#
tantek
so they had some degree of interop with other federated social systems
brennannovak joined the channel
#
tantek
much better to use HTML link relations for discovery than "special" paths like /profile, /followings etc.
#
tantek
special paths are a bit of a known anti-pattern
#
tantek
like robots.txt, favicon.ico etc.
#
sandeepshetty
and .well-known (yucks!)
#
tantek
yeah :(
brennannovak joined the channel
#
tommorris
!tell tantek initial design work on pingbacks: http://gyazo.com/f28d71f025e75637f5009af321be2ae5 - just using your site as an example.
#
Loqi
Ok, I'll tell him that when I see him next
tantek joined the channel
#
Loqi
tantek: tommorris left you a message 16 minutes ago: initial design work on pingbacks: http://gyazo.com/f28d71f025e75637f5009af321be2ae5 - just using your site as an example.
tantek_ joined the channel
#
tantek_
*click*
tantek joined the channel
#
tantek
so here's the thing
#
tantek
IMHO pingbacks don't deserve their own UI
#
tantek
pingbacks should merely be the plumbing behind distributed comments
#
tantek
so really, there shouldn't be any different in the UI, no matter where the comments came from
#
aaronpk
yes! I think people often confuse pingbacks with the way wordpress displays them
#
tantek
and comment UIs/designs on blog posts are already fairly well developed
#
aaronpk
and that causes a lot of people to hate them and think they are just a bunch of spam
#
tantek
right
#
tantek
one thing WordPress got right is to intermingle them with "normal" comments
#
tantek
MT was worse about that, by putting a whole different section for "Trackbacks"
#
tantek
plumbing never deserves a UI heading
#
tommorris
well, I'm leaning towards not having comments at all. just pingbacks.
#
tommorris
you wanna respond? here, blog. ;-)
#
tantek
then you're leaning towards having distributed comments only
#
tantek
and pingback is just your choice of plumbing
#
tantek
I often IRC before I blog ;)
#
tommorris
I might want more than just pingbacks in the future, in which case I'll redesign it to have more than just that. but I like minimalism.
#
tantek
minimalism is not reinventing comment display design