#indiewebcamp 2013-06-28

2013-06-28 UTC
andreypopp and tilgovi joined the channel
#
www.sandeep.io
edited /responses (+1208) "/* object of action */ Clarified what I meant by common usage, why thing over object/post, liked is past tense and note about object-of-mention."
(view diff)
#
bret
i really dig the mozilla logos http://blog.seanmartell.com/brands/ especially the hobbes one
#
bret
is there any advantage or disadvantes to using or not using www.?
#
aaronpk
your root domain cannot be a CNAME record, it has to point to an IP address. that has implications for where you host your site, but that's really the only technical reason I can think of
#
bret
so basically if the site was going to change IPs and you had a free dyndns or something www would be usefull
#
aaronpk
no not quite
#
aaronpk
you can still change the IP just as easily with an A record
#
aaronpk
it's more of an issue for larger-scale systems that use load balancing and such
#
bret
something I have never dealt with
#
KevinMarks
it can be an issue hosting on heroku or other AWS based
#
aaronpk
if you host your root domain on heroku, they have to give you an IP address to point to. that means they're more limited in how they can shuffle things around behind the scenes
#
aaronpk
whereas if you point www.example.com -> example.herokuapp.com they can change things easier on their end
#
aaronpk
but more and more these providers are adding better support for root domains
#
tantek.com
edited /deleted (+1560) "continue debunking excuses for 404"
(view diff)
#
@joshuajuran
RT @benwerd: I'm going to eat burritos and talk #indieweb tomorrow night: http://werd.io/event/51ccc9ecbed7deaf58e8ff5e/indieweb-burritos
#
aaronpk
hmm, it does occur to me that it's far more difficult to send back a "410" code than it is to serve a static file
#
aaronpk
you'd have to implement the 410s in an htaccess file if you're using jekyll for example
#
aaronpk
in that sense it's slightly more difficult to maintain than the original comment
#
aaronpk
oh there's already a note about jekyll
benwerd joined the channel
#
aaronpk
bret: i'm heading to side door right now, then stopping by the appfog party later tonight fyi
#
tantek
scrolls up
#
bret
yeah im still really buys around the apartment right now, not going to be able to be free for a while
#
tantek
pdurbin - I don't see the point of "linked data" above and beyond
#
tantek
"just use JSON"
#
bret
is the appfog party a hack session?
#
tantek
or rather, JSONLD that is
#
aaronpk
no it's just a party :)
#
tantek
it's just yet another DRY violation design - I mean really? who cares?
#
aaronpk
which is why i'm stopping by there on my way home, not starting there
#
tantek
just use HTML+microformats2. If you want JSON, parse it with an open source parser. Done.
#
bret
ahh i see, well have fun! we should try another time for sure though
#
tantek
aaronpk - nice UI!
#
bret
especially if you are going to be around portland
#
aaronpk
bret: i'm around portland for another week, then will be out of town a bunch
#
aaronpk
back later!
#
bret
doh, ill drink some portland coffee for you homee when you are away
#
tantek
aaronpk - re: far more difficult to send back a "410" code than it is to serve a static file - then put the 410 in the static file
#
tantek
first meta in the static file: <meta http-equiv="Status" content="410 GONE"/>
#
bret
tantek i wish google let you do 301 redirects via static files
#
bret
i was unable to do a domain change in the webmaster tools tool
#
bret
its still showing my old domain
#
bret
in search results
#
tantek
search results take a while to update too
#
tantek
I think even when I *am* returning a 410 status code, I'm going to put that meta in the returned HTML.
#
bret
i like the static approach because the html itself communicates the issue
#
tantek.com
edited /deleted (+144) "/* Brainstorming */ in addition"
(view diff)
#
bret
rather than relying on a webserver to get it right
#
tantek
bret indeed. and it's the machine equivalent for the static prose in the file saying "this content has been deleted" or whatever
#
@sandeepshetty
Every time I look at /mentions and see the decentralized conversations and likes, I get a smile on my face #indieweb http://www.sandeep.io/93
#
pdurbin
tantek: oh good... it's not just me :)
#
tantek
next time someone pitches you something like that just ask them how they're using it on their own personal site
#
tantek
and if they're not, then just walk away
#
tantek
(hint: nearly all the time they're not)
#
tantek
(hint2: instead they'll ask you to join a discussion on an email list)
b0bg0d joined the channel
#
pdurbin
well, it's tim berner lee's idea :)
#
tantek
so Tim shipped an implementation of HTML, HTTP, URL.
#
tantek
that's why those took off as "his idea"
#
tantek
same cannot be said about RDF, or linked data
#
tantek
no one gets a pass on selfdogfooding.
#
pdurbin
hmm, somehow i thought folks in this channel are into rdf. maybe I'm thinking of something else
bnvk joined the channel
#
tantek
tommorris has some history there. but no, RDF is one of the big rocks that gets in your way of actually getting things done.
#
tantek
as nearly all DRY violations are.
#
benwerd
once got booed for talking down RDF on stage
#
pdurbin
tommorris: I did notice you in #swig ... logs at http://chatlogs.planetrdf.com/swig/2013-06-27.html
#
tantek
tommorris does actually use/deploy stuff on his own site, which is why his word carries a lot more weight.
#
tantek
than the usual RDFer.
#
pdurbin
benwerd: booed huh? :)
#
benwerd
well, by the proportion of the crowd that cared
#
benwerd
my formative Oxford Geek Night experience *nod*
#
benwerd
I would have joined in!
#
benwerd
can't believe i've never been to dConstruct
#
tantek
this would be an excellent year to go
#
benwerd
might go this year, now that I live 5000+ miles away, ironically
#
tantek
since we're doing an IndieWebCampUK on the Sunday 2 days after it
#
benwerd
indeed - you'll see me on the attendee list
#
benwerd
doing my best to arrange to be there
tantek joined the channel
#
tantek
LOL hadn't seen this version before: http://www.flickr.com/photos/yoz/6270904049/
#
aaronpk
tantek: yea cool, just saw that after reading the /deleted page
#
tantek
I don't think sandeep gets the design principle of starting protocols/formats small, restricted, conservative, and only expanding them as needed by real world use cases.
ozten joined the channel
#
tantek
I'll keep following up on his arguments, but I think there's a larger methodological / principled misunderstanding there.
#
aaronpk
he did pretty good with webmention
#
tantek
which is ironic, because in the stuff HE comes up with, it's VERY small, minimal, conservative.
#
tantek
and he takes other stuff (like pingback) and SHRINKS it
#
tantek
right
#
tantek
so he's thinking/designing properly when he's creating, but not when arguing on the wiki
#
tantek
it's strange
#
tantek
usually people have the opposite problem
#
tantek
they talk a good talk, but then falter when it comes to action
#
aaronpk
interesting how my prioritized list of itches changes from day to day on p3k
#
tantek
aaronpk - exactly - that's what happens when you write them down publicly :)
#
tantek
ah here's the one with the longer discussion thread: http://www.flickr.com/photos/29022619@N03/4968410475/
#
tantek
if only there were a way to relate all those photos together as being of the same event, perhaps some way to LINK their DATA :P
#
aaronpk
lol! i was waiting for the punchline
#
aaronparecki.com
edited /p3k (+266) "re-prioritizing itches"
(view diff)
#
aaronparecki.com
edited /p3k (+73) "/* Features */"
(view diff)
#
benwerd
oh, I like that idea.
#
KevinMarks
I like Danny O'Brien's comment "RDF was the LSD of the 90s"
#
tantek
KevinMarks, it wasn't nearly that visionary.
#
aaronpk
tantek: lol!
#
@t
federated #indieweb comments, likes, event + RSVPs demo by @benwerd:
#
tantek
benwerd - your POSSE'd @-replies are not threading!
#
benwerd
tantek - they aren't when there are two targets. I broke something! I had made the primary reply the post on your site
#
benwerd
I'm going to change the way that works, so you never have a "primary" reply target in my interface
#
tantek
benwerd you can automatically determine the POSSE tweet copy via rel-syndication
#
benwerd
yep, haven't implemented rel-syndication discovery yet
#
benwerd
that's on the to-do
#
benwerd
I'm writing my itches list right now, inspired by aaronpk ;)
#
tantek
benwerd - that's exactly why I added the rel-syndication links to my posts at IndieWebCamp!
#
tantek
it's why it was the top feature on my to-do list
#
tantek
so that folks with *better* reply code could reply to my posts, and have their POSSE tweet replies thread with the POSSE tweet of my original post.
#
bret
too bad my site was so ugly the day benwerd shot that video
#
tantek
bret - nothing like vanity to stoke those selfdogfood priorities huh? ;)
#
bret
it looks remotely better a day later
#
tantek
you never know when someone might make a screencast of your site and freeze your look&feel
#
tantek
bret - nice URL design ;)
#
bret
I need to fix the page title though
#
werd.io
edited /idno (+1590) "Added itches section, briefly discussed IndieFriends"
(view diff)
#
KevinMarks
heh, or send yo a pull request with better CSS like fred did for me
#
KevinMarks
does anyone know node express and sessions, BTW?
#
aaronpk
yea I even do that to myself: http://aaronparecki.com/notes/2012/09/09/3/indieweb (re: freezing the look/feel)
#
KevinMarks
my kludges seem to be preserving state globally, ratehr than per user
#
KevinMarks
which is problematic when you're auth'd as them to twitter
#
werd.io
edited /idno (+32) "/* IndieFriends */"
(view diff)
#
bret
tantek or aaronpk, could I move a generic version the github repository I use for my site into the indiewebcamp group? Im thinking about doing something similar to octopress where people interested in using it can fork off of a generic version easily and track developemnt updates
#
KevinMarks
did y'all see heroku fork today?
#
bret
no what happened?
#
tantek
so far we're only using the indieweb repo for building blocks - not for any entire sites or platforms
#
tantek
re: bret
#
bret
oh i see
#
tantek
we'd like to avoid any even implicit monoculture tendencies
#
bret
that make sense
#
benwerd
KevinMarks: oh, that's cute
#
tantek
building blocks good. monoculture bad.
#
benwerd
so you can fork an application and immediately have an installed version
#
aaronpk
tantek: I like that criteria for the indieweb github group
#
tantek
aaronpk - feel free to add it to the group description!
#
KevinMarks
this is edging close to what I want which is 'spawn me a running instance of this github repo'
#
benwerd
KevinMarks: if they were able to give the source project a kickback for new accounts forking their software, I'd be all over that (I'm not too proud to say)
#
KevinMarks
yeah, that makes sense, and does fit in with the existing plugins model
#
bret
tantek what was the deal with these death to the semantic web
#
bret
pictures
#
benwerd
Goodnight all
#
Loqi
sleep tight!
#
bret
bye!
#
bret
are we doing semantic web stuff?
#
KevinMarks
only in lowercase
#
tantek
aaronpk - perhaps we should grandfather those projects rather than delete them.
#
tantek
They haven't been updated in a while
#
tantek
so people will know to treat them accordingly
#
tantek
Maybe edit their descriptions accordingly.
#
aaronpk
well the indieweb repos are just forks
#
tantek
I thought those might have been originals
#
tantek
especially helloworld from mime
#
aaronpk
yea they're all forks. I think sandeep just forked a couple relevant projects when we set up the org so that it wouldn't be empty
#
aaronpk
but now that we're actually developing useful building blocks there those are less relevant
#
@kevinmarks
#indiewebcamp @t: we'd like to avoid any even implicit monoculture tendencies
#
KevinMarks
wonders if I should give up on cookies and just use local storage
#
tantek
hey Kevinmarks you can always link to the IRC logs to cite permalinks to IRC quotes ;)
#
aaronpk
lol "BTW if this all makes sense then feel good because it really is rocket science."
#
tantek
KevinMarks I tend to be skeptical of physics analogies to anything virtual / CS.
#
bret
I wonder if a https://svbtle.com/ type news site might more appropriate than the way indienews is set up right now. Like more of a community weblog where the submission process is the same as indie news, but the emphasis is more on the order of submission and medium sized previews rather than voting on titles
#
bret
also the emphasis would be on getting more people to adopt indie web technologies rather than asskissing silicon valley types
#
bret
in order to publish
#
KevinMarks
did you get https://www.potluck.it/ invites too?
#
bret
who me?
#
KevinMarks
very silo'd
#
aaronpk
bret: I think you mean more like https://en.wikipedia.org/wiki/Planet_(software) than svbtle
#
KevinMarks
planetplanet is very webmentionesque
#
bret
I have never heard of this
#
aaronpk
and possibly not pulling in the full content, just a preview kind of like facebook size previews
#
bret
right
#
aaronpk
oh god i'm getting old
#
KevinMarks
heh, Jon Abrams just joined potluck, talking of getting old
#
bret
what is potluck
#
KevinMarks
it's a link-sharing silo built by branch
#
aaronpk
potluck is a link-sharing silo built by branch
#
KevinMarks
that you have to create an account at by verifying your email
#
aaronpk
there, now loqi will answer
#
aaronpk
what is potluck?
#
Loqi
potluck is a link
#
aaronpk
dammit loqi
#
bret
What is potluck?
#
Loqi
potluck is a link
#
aaronpk
potluck is a link sharing silo built by branch
#
bret
What is potlick?
#
aaronpk
overly aggressive regex
#
aaronpk
what is potluck?
#
Loqi
potluck is a link sharing silo built by branch
#
aaronpk
better
#
bret
whoa
#
KevinMarks
potlick is what you do when the soup was really tasty
#
bret
What is potlick?
#
Loqi
potlick is what you do when the soup was really tasty
#
KevinMarks
I didn't realise Loqi did implicit jibot defs
#
KevinMarks
the http://planet.intertwingly.net/ could be a good model for an indeiweb reader
#
KevinMarks
the memes feed is great usually
#
aaronpk
haha, but yea I get your point
#
KevinMarks
the memes feed is anything that 3 or more blogs link to, wiht refs
#
aaronpk
KevinMarks: that's a neat idea
#
aaronpk
PlanetIndie?
#
aaronpk
maybe I can make a config option in the IndieNews source code to turn it into more of a Planet than a Reddit, then I can launch one at planet.indiewebcamp.com
#
KevinMarks
fear of an indie planet
#
aaronpk
sounds like an album title
#
aaronparecki.com
edited /p3k (+1380) "/* Photos */ additional notes about photos"
(view diff)
#
aaronpk
currently doing research on event RSVP pages
fmarier joined the channel
#
@BillSeitz
federated #IndieWeb comments, likes, event + RSVPs demo by @BenWerd:
fmarier, melvster and erikmaarten joined the channel
#
aaronpk
wow there really isn't an RSVP icon
#
aaronpk
well this took an unexpected direction
#
aaronpk
well that wasn't quite what I expected, but I'm pretty happy with it:
#
aaronparecki.com
edited /p3k (+21) "RSVPs are now visually marked on their permalink pages"
(view diff)
#
bret
aaronpk looks great! I like the traffic light analogy
andreypopp joined the channel
#
bret
I better get an RSVP out to that event, I already claimed bens free drink
eschnou, bizybee and josephboyle joined the channel
#
aaronpk
haha how did you claim the drink if you didn't rsvp yet?
pfefferle joined the channel
#
bret
ben bought me coffee the day before he left
#
neuro`
Good morning UGT
#
pfefferle
Good Morning
#
aaronpk
good morning ;)
#
tommorris
pdurbin, and (absent) tantek: I self-dogfood RDFa. I'm using it to express extra semantics
#
tommorris
and have been doing RDF stuff for years
#
tommorris
anyway, I've just woken up. ;)
#
pfefferle
what extra semantics exactly?
eschnou joined the channel
#
tommorris
sexual orientation and dietary preference
pfefferle joined the channel
#
pfefferle
Ah ok, but only as Markup, Not via webmention?
#
pfefferle
I read your Post about that
#
bret.io
edited /IndieAuth (-140) "/* To do */ Removed the add persona as an identity provder item from the indieauth todo list"
(view diff)
pfefferle joined the channel
#
aaronpk
thanks!
#
bret
you did all the hard work :p
seyz joined the channel
#
aaronparecki.com
edited /IRC_People (+7) "update to named timezones"
(view diff)
andreypopp joined the channel
#
bret
!!!!
#
bret
cool!
#
aaronpk
magic!
#
bret
sparkles!
#
bret
great work aaronpk!
eschnou joined the channel
#
aaronpk
well this should be interesting to see what happens with this
#
bret
it has been a firehose of new ideas
#
bret
aaronpk, i have an idea to speed up indieauth
#
aaronpk
me too
#
bret
enable cookies? so I dont have to log in every time
#
aaronpk
i guess that would work, but right now i'm not storing any state in indieauth which is nice. i'd rather keep it that way.
#
aaronpk
there is no "logged in" state
#
bret
ah, i presume thats a bit easier to deal with from a security standpoint?
#
aaronpk
my thought was to cache the list of providers on your domain and whether they are verified/supported, so that you don't have to wait for them to be checked every time. I would also provide a "refresh" button so you can force reload it if you change your site
#
bret
is tommorris displaying comment replies?
#
aaronpk
can't remember
#
aaronpk
ok it's late. that's all for me tonight!
#
bret
Bye!
#
bret
great work!
morrocco_mole, bizybee and erikmaarten joined the channel
#
bret
gnight
tantek joined the channel
#
pdurbin
tommorris: cool, cool. I'm open to new ideas :)
barnabywalters joined the channel
#
Loqi
barnabywalters: tantek left you a message 12 hours, 33 minutes ago: re: wrapper around PHP DOMDocument - totally agreed. That's an itch of mine too, since I used PHP DOMDocument to read/write my storage files. I want to wrap it with actual JS DOM written in #cassis.js naturally.
#
tommorris
pdurbin: so, you can layer RDFa and microformats-2 (and microformats original - is there a name for that? classic?) quite nicely.
#
tommorris
there's a bit of DRY fail there
#
tommorris
but meh.
#
tommorris
I'd like to do RDFa-based webmention too. shouldn't be too hard especially now the RDFa parsers have left the realm of XHTML theory and entered the world of HTML5-based reality. ;)
#
pdurbin
ah, I think microformats were the thing I was thinking of. that people in this room are into, perhaps. rather than RDF. sounds like they're pretty different
#
tommorris
RDFa lets you represent RDF in HTML. Microformats original have their own data model. Microformats-2 have a consistent JSON representation.
#
pdurbin
ok. I guess I'm not sure why I would want either one. maybe I'll go read up on them a bit :)
bnvk joined the channel
#
tommorris
pdurbin: so you want one or both to markup the page to be parseable
#
tommorris
so, if you run my page through an RDFa parser, you get back data like this - https://gist.github.com/tommorris/5474597
#
tommorris
run tantek's site through a microformats2 parser and you get back data like this - http://microformat2-node.jit.su/?url=http%3A%2F%2Ftantek.com%2F&callback=&filters=
#
tommorris
you can then use that data for interesting things. ;)
#
@paulbooker
RT @t: federated #indieweb comments, likes, event + RSVPs demo by @benwerd:
#
pfefferle
tommorris: do you know a good rdfa parser written in php?
#
tommorris
pfefferle: haven't looked recently.
#
pfefferle
what do you use?
#
tommorris
I mostly these days use rdflib in Python
#
tommorris
I may try Ruby's RDFa parser at some point
#
tommorris
arc2 might work. haven't tried it for a while
#
tommorris
arc2 is the main PHP parser AFAIK
#
pfefferle
but i think i have read that development has stopped :(
#
tommorris
these days, I'm lazy enough that when something isn't available in a language I want, I just make a simple HTTP server to serve it up.
#
pfefferle
barnabywalters: thanks! should spend more time with the whole composer stuff
#
tommorris
that's how I'm using microformats-node
#
pdurbin
tommorris: I'm looking at your gist. seeing some foaf stuff, which I've looked at before
#
tommorris
easyrdf looks like the solution in php
#
tommorris
pdurbin: so, the difference between RDF-world and microformats is with microformats, schema development is centralised and community-driven
#
tommorris
whereas anyone can make their own RDF schema/ontology
#
tommorris
so, in the JSON example I've given there, I've used a mixture of FOAF, Schema, DBPedia (for 'sexualOrientation') and schema.tommorris.org/diet (for 'dietaryPreference')
#
tommorris
you can mix and match ontologies
#
tommorris
also, GND (mostly to prove a point) and Geo (which is the geographical schema most people use and is blessed by w3c)
#
pdurbin
tommorris: sounds very flexible
#
tommorris
yep. well, you can represent in most of the semantics-in-HTML solutions something like an RSS feed
#
tommorris
but without the don't repeat yourself of publishing a separate XML version
melvster joined the channel
#
pdurbin
hmm, I guess publishing an RSS feed as well would be repeating yourself... but RSS is still pretty popular. people like being able to subscribe to RSS feeds
#
grawity
publish an Atom feed :P
#
tommorris
yep. it's a philosophy vs. practicality decision
#
tommorris
ideally, HTML would be everything you need. but in practice, you probably need to publish an Atom feed.
#
pdurbin
I'm not even sure if I publish atom or rss from my wiki
#
barnabywalters
pdurbin: there was talk of building a mf -> ATOM proxy server which would solve both problems
#
pdurbin
atom, looks like: http://wiki.greptilian.com ... to me it doesn't matter
sandeepshetty joined the channel
#
sandeepshetty
!tell tantek: Will try to login later tonight (your morning) and discuss 404. I'm all for constraints and starting small... just coming from the angle that 404 is a valid response code for delete.
#
Loqi
Ok, I'll tell him that when I see him next
#
pdurbin
barnabywalters: cool. I guess I'll just stick with straight atom for now. I don't even use the feed myself but maybe some day I will. I doubt anyone else cares when I update my wiki :)
#
sandeepshetty
Se I spent sometime working on a uf2 reader yesterday and boy are there issues....
#
sandeepshetty
it's looks like either there isn't a standard for publishing feeds in uf2 or no one is following it :)
#
pfefferle
sandeepshetty: i am trying to build something very similar, based on a pubsubhubbub client...
#
Loqi
Ok, I'll tell them that when I see them next
#
sandeepshetty
pfefferle: sent a webmention and it looks like it filtered the tag out.
#
pfefferle
once again?
#
sandeepshetty
yeah.. check the reply on my site and yours.
#
sandeepshetty
pfefferle: I think php-mf2 is doing some filtering.... /cc barnabywalters
#
pfefferle
still see the md syntax
#
barnabywalters
sandeepshetty: filtering what?
#
pfefferle
still markdown
#
pfefferle
you have to change your syntax first ;)
#
pfefferle
you have still the markdown on your site ;)
#
sandeepshetty
so I'm using <http://example.com> which is markdown for linking without link text and I syndicate plaintext markdown so it will have the tag that way. Ideally if the receiver is picking up p-summary it should be encoded and not filtered or atleast encoded first and then filtered
#
sandeepshetty
right not it's just stipping <http://example.com> out of the reply which means loss of context
#
sandeepshetty
ideally should just appear as <http://example.com> where <> are encoded to &tl
#
pfefferle
aaaaah ok, then the parser does the right thing because it looks like a html tag
#
barnabywalters
sandeepshetty: I didn’t add an option to turn on encoding of non e-* properties to my web UI yet but it’s in php-mf2
#
pfefferle
and p- does not contain any html
#
sandeepshetty
pfefferle: this is wrong. because I could be giving you html code example and it would strip it out..
#
pfefferle
why using md instead of html?
#
sandeepshetty
it shoudl encode and *not* filter
#
pfefferle
for the view?
#
sandeepshetty
I prefer syndicating md because if I send html and the receiver filters it then there will be loss of context (See indiewebcamp.com/plaintext)
#
pfefferle
but if it doesn't filter, you have to strip tags for every p- element
#
pfefferle
this is very hard on cms systems
#
sandeepshetty
the correct way to do this is to encode any p-* properties with htmlspecialchars() because < is plaintext mean < (but represented as &lt;)
#
sandeepshetty
s/< is/< in
#
Loqi
sandeepshetty meant to say: the correct way to do this is to encode any p-* properties with htmlspecialchars() because < in plaintext mean < (but represented as &lt;)
#
sandeepshetty
you could choose to filter it after encoding
scor joined the channel
#
pfefferle
but that means that i can't use p-summary or p-name on my view, because they have html in it?
#
sandeepshetty
though encoding should be enough.. BTW, the encoding should be done only while displaying/echoing
#
sandeepshetty
pfefferle: it's not html after you encode it...
#
pfefferle
yes, but it looks like crap if every <p> in my code is displayed like that
#
pfefferle
because i have html in my summary
#
pfefferle
to mark it up
#
sandeepshetty
pfefferle: and like I mentioned there other are situations (like when I ahve html code examples in my comment) where you *will* get html tags in the comment
#
sandeepshetty
pfefferle: there shouldn't be <p> in p-* properties unless it is meant to be displayed that way
#
barnabywalters
sandeepshetty: pfefferle: if properties have been marked up with p-* and you’ve told the parser to act consistently THEY WILL NOT HAVE HTML IN
#
sandeepshetty
you shouldn't put html content in p-* properties.. thats what e-* are for
#
Loqi
it is probable
#
pfefferle
that means the parser crops it?
#
barnabywalters
pfefferle: p-* parsing rules usually = take the innerText / textContent of the node
#
barnabywalters
e-* means concatenate HTML representations of each child of the element
#
sandeepshetty
so when my comment had <http://example.com> it was meant to be displayed that way
#
barnabywalters
only reason php-mf2 doesn’t escape none e-* properties by default is because it’s not in the parsing spec
#
barnabywalters
it will be default in v0.2
#
pfefferle
but i don't do any cropping!
#
pfefferle
i take what the parser returns
#
pfefferle
nothing more nothing less
#
pfefferle
that means that innerText is cropping your stuff
#
barnabywalters
as a mf consumer you have no control over whether or not the content is marked as p-* or e-*
#
barnabywalters
that responsibility lies with the author
#
pfefferle
yes, but will <http://example.com> be cropped or not?
#
barnabywalters
hence the importance of the parser behaving consistently for all property types
#
barnabywalters
pfefferle: no
#
pfefferle
but why i don't get it as a result?
#
barnabywalters
pfefferle: can you send me example code so I can actually see what you’re having problems with?
#
pfefferle
sandeepshetty can you please witch back to your old code with <url>
#
pfefferle
so we could test it?
#
pfefferle
now i am on track
#
pfefferle
and i understand what you mean!!!
#
pfefferle
oh man i am so a noob today!!!
#
pfefferle
sorry sorry
sandeepshetty joined the channel
#
sandeepshetty
pfefferle: I'm out (picking up my daughter from school). Be back in 40 min.
eschnou joined the channel
#
pfefferle
!tell barnabywalters the parser works totally fine, i had a little blackout and haven't mentioned that <url> uses &lt
#
pfefferle
and &gt
#
pfefferle
and so it couldn't be cropped!
#
Loqi
Ok, I'll tell him that when I see him next
#
pfefferle
!tell barnabywalters, sandeepshetty I take everything back! I am right though!! you use a not <http://example.com> encoded instead &lt;http://example.com&gt
#
pfefferle
and that is skipped by the mf2 parser because he thinks it's a html-tag and ignores it! http://pin13.net/mf2/?url=http%3A%2F%2Fwww.sandeep.io%2F92
#
Loqi
Ok, I'll tell them that when I see them next
#
Loqi
Ok, I'll tell them that when I see them next
#
pfefferle
thanks Loqi
#
Loqi
you're welcome
texburgher, punkprogrammer, brbcoding and sandeepshetty joined the channel
#
pfefferle
as far as i can see, the latest pubsubhubbub spec should also work with webmentions
#
pfefferle
you need to ping the server but it is not specified how
#
pfefferle
!tell barnabywalters I like the new rel-parser a lot, thanks!
#
Loqi
Ok, I'll tell him that when I see him next
bnvk and sandeepshetty_ joined the channel
#
www.sandeep.io
edited /IRC_People (-1) "Fixed my timezone."
(view diff)
#
sandeepshetty
pfefferle: so it's php-mf2 that is stripping out the tag?
#
pfefferle
yes, it it is totally right, that it does it
#
pfefferle
because he had to skip every html tag
#
pfefferle
and <url> looks like one
#
pfefferle
you have to encode it
#
pfefferle
even the browser can't display it properly
#
pfefferle
because he thinks it's a tag
#
sandeepshetty
pfefferle: think of it as a plaintext file.. you won't receive it with encoded < and >, it is the responsibility of the client that is displaying it to encode it based on the context (e.g., as part of a HTML document) in which it is displaying it.
#
sandeepshetty
what php-mf2 is doing is not right if it is stripping what look like tags from p-* content
#
sandeepshetty
argh! *looks
#
pfefferle
as barnabywalters and i understand the spec, "p-" attributes has to be parsed like plain text, not they have to be plain text.
#
pfefferle
you can use html tags to have a nicer view, but the parser should only interpret the text and has to ignore the tags
#
sandeepshetty
that cannot be right... given the simple scenario of having example html code in the content that is meant to be displayed as is and not stripped... if that is what the spec says then the spec is wrong. Where does it say you have to strip html in the spec?
#
pfefferle
if you have sample code, you have to encode it
#
pfefferle
vocabulary generic parsing, element text in general, treat certain HTML element/attribute combination as special and use those first, e.g. img/alt, abbr/title.
#
pfefferle
that means it can have html tags, but ignore them
#
sandeepshetty
pfefferle: encoding *cant* be done by the publisher because it does not know the context in which it will be displayed it.
#
sandeepshetty
that link doesn't say anything about stripping html
#
pfefferle
that said "treat certain HTML element/attribute combination as special and use those first, e.g. img/alt, abbr/title." it can include html and you should use the alt or title instead of the source
#
sandeepshetty
no that means pick up the plain text from the alt/title attribute....
#
sandeepshetty
see examples of h-card where it picks up the author name from an img alt attribute
#
sandeepshetty
it's taking about how to figure out what to treat as a p-* attribute.. not what it contains
#
pfefferle
yes and not: encode it
#
sandeepshetty
the spec should not and cannot talk about encoding..
#
sandeepshetty
it is the responsibility of the client that is displaying (to protect itself from XSS attacks, etc)
#
pfefferle
if you don't use encoding your reader can't see it anyways
#
pfefferle
it has to be encoded to be displayed
#
sandeepshetty
what is the "reader" here?
#
pfefferle
a person
#
pfefferle
what do you see if you display your p-summary
#
sandeepshetty
if the client that is displying it encodes it properly the reader *will* see it as it was meant to be seen
#
pfefferle
everything exempt the <http://example.com>
#
pfefferle
*except
#
pfefferle
because it is treated like a tag
#
pfefferle
you have to encode it, otherwise it's not displayed properly
#
sandeepshetty
it should not be treated like a tag because it was explicitly marked as plaintext with p-*
#
pfefferle
not for the person who reads the post, nor for the parser
#
sandeepshetty
yes the client (receiver) has to do that
#
pfefferle
let's talk about users not parsers
#
sandeepshetty
pfefferle: again, the sender cannot encode it because it does not the context in which it will be displayed (html, plaintext)
#
pfefferle
what will a user see if you display your p-summary
#
sandeepshetty
and this is about parser so talking about users doesn't help
#
pfefferle
no, it's about dry
#
sandeepshetty
which is what was sent
erikmaarten joined the channel
#
pfefferle
please disable the display none
#
pfefferle
he can't see it
#
pfefferle
because the browser thinks it's a tag
#
sandeepshetty
if I was sending html, yes I would encode the < and > in it...
#
pfefferle
no! your user won't see it!
#
sandeepshetty
ah wait .. I think we might have a disconnect
#
sandeepshetty
if the parser was picking up the e-content then it is right to strip it out because I am not encoding it
#
pfefferle
no, then it is not
#
pfefferle
because then the whole content is used
#
pfefferle
with html tags
#
sandeepshetty
the stripping is for security..
#
pfefferle
it is because of the spec
#
sandeepshetty
it *has* to filter before displaying
#
sandeepshetty
actually it is both..
#
pfefferle
if you would mark up your code with e-conent all your html tags would be used
#
pfefferle
displaying is another thing
#
sandeepshetty
you are right in that it will skip it because it looks like an invalid tag.. and I'm talking about what it has to do before displaying
#
sandeepshetty
so it's the domdocument that is skipping it?
#
pfefferle
microformats is all about dry, that means the parser is using what the user sees
#
sandeepshetty
I see what you mean now.. my bad..
#
pfefferle
because p- ignores all html elements
#
sandeepshetty
no this doesn't have to do with uf2 at all..
#
sandeepshetty
it's just invalid html for domdocument
#
pfefferle
if you would use e- it would work
#
pfefferle
try it with a <br />
#
pfefferle
the parser will skip it
#
sandeepshetty
writting a test page
#
pfefferle
you can check it with my webmentions… on my page, they are marked up with html, on your site they are plain text
#
pfefferle
otherwise you wouldn't be able to use any html in p-
#
pfefferle
it's easy take this page http://www.sandeep.io/39
#
pfefferle
and remove the display none of the p-summary
#
pfefferle
the p-summary is meant to be visible for users and your output is very cryptic
#
sandeepshetty
it has to be encoded in both.. notice how it's messed up in e-content..
#
sandeepshetty
it's just an html issue on my part that I need to fix..
#
sandeepshetty
it's not related to uf2
#
sandeepshetty
since I'm sure I'm only syndicating plaintext, I need to encode it.
#
pfefferle
on your demo site
#
pfefferle
that means the first is encoded
#
pfefferle
and than it works
#
pfefferle
that's what i menat
#
sandeepshetty
yes to show that encoding it will fix it
#
sandeepshetty
and the second is not to show that its not about putting it in e-content
#
pfefferle
no, it will work if you encode it
#
pfefferle
use <br /> instead
#
sandeepshetty
yes both will work if you encode it because otherwise <http://example.com> is treated as an invalid html
#
pfefferle
and don't encode both
#
sandeepshetty
pfefferle: I know that <br /> will work.. just trying to highlight that *this* issue is because of invalid html that I'm generating and not related to uf2
#
pfefferle
please try it
#
pfefferle
it will be skipped by summary!!!
#
pfefferle
your example only proofs that <http://example.com> is a wrong tag… not more, not less
#
sandeepshetty
maing changes for <br /> and deploying
#
sandeepshetty
pfefferle: yes.. I was specifically trying to test that only
#
pfefferle
ok, i will set up the test page for you
#
sandeepshetty
pfefferle: Mine is being pushed
#
sandeepshetty
give it two min and we can see what <br /> test looks like
#
sandeepshetty
ok you can check now
#
sandeepshetty
it stripped out <br />.. that it shouldn't have done
#
sandeepshetty
(from p-summary)
#
sandeepshetty
yeah same result..
#
sandeepshetty
that to me is not right..
#
sandeepshetty
the parser should not strip out html from p-summary..
#
pfefferle
but to display plain text WITH html unencoded isn't right eather
#
sandeepshetty
it could choose to encode it but it shouldn't.. it should leave it up to the client
#
pfefferle
that is what you show to your users
#
pfefferle
p-summary should be user friendly
#
pfefferle
micro formats shouldn't be hidden
#
sandeepshetty
pfefferle: if you put html in what you've said is plaintext, it means you want to show it encoded
#
sandeepshetty
p-summary *will* be hidden even if I have html e-content
#
pfefferle
yes, but if you want to show it encoded, then you have to encode it, otherwise the browser will suck it
#
sandeepshetty
no point in showing users both e-content and p-summary
#
pfefferle
but you can't show unencoded html on a html page
#
sandeepshetty
sorry did not understand..
#
pfefferle
you have to encode it, otherwise the browser interprets it
#
sandeepshetty
I'm not showing encoded html anywhere
#
pfefferle
because your markdown interpreter will do that for you
#
sandeepshetty
md doesn't do anything to p-summary
#
sandeepshetty
and that it what you will pick up.. so everything else is out of scope..
#
pfefferle
you are using encoded html here
#
pfefferle
in the code box
#
pfefferle
for p-summary you don't encode it… the user get's only text… no example code!
#
pfefferle
it is not readable in the browser
#
sandeepshetty
right.. we are on the same page on that... I already said that it's a problem that I'm not encoding my p-summary
#
pfefferle
yes, and if you do so, the parser would run perfectly fine
#
sandeepshetty
and that is the only issue here... it's a problem at the HTML level..
#
pfefferle
because then it isn't html any more
#
pfefferle
yes, because the parser skips html for p-
#
sandeepshetty
even if there was no uf2 on the page.. this is a problem.. so it has nothing to do with uf2 is all I'm saying
#
sandeepshetty
aaronpk: It looks like I broke the logs
caseorganic joined the channel
#
sandeepshetty
!tell aaronpk: It looks like I broke the logs: Fatal error: Uncaught exception 'Exception' with message 'DateTimeZone::__construct(): Unknown or bad timezone (Asia/Mumbai)' in /web/sites/indiewebcamp.com/www/irc/inc.php:118 Stack trace: #0 /web/sites/indiewebcamp.com/www/irc/inc.php(118): DateTimeZone->__construct('Asia/Mumbai') #1 /web/sites/indiewebcamp.com/www/irc/logs.php(226): formatLine(Array) #2 {main} thrown in /
#
Loqi
Ok, I'll tell them that when I see them next
#
pfefferle
kind of
#
sandeepshetty
but you were right about it being an issue with me not encoding..
#
pfefferle
we were very noisy...
#
sandeepshetty
!tell aaronpk: Fixed it. Might want to handle invalid timezones to avoid killing the logs
#
Loqi
Ok, I'll tell them that when I see them next
#
sandeepshetty
though some of these things need to be documented now..
#
sandeepshetty
will do it after I fix my implementation
#
pfefferle
but what about if you encode your stuff in p-summary? should the parser encode also the "real" tags?
#
sandeepshetty
I shouldn't have "real" tags in p-summary and if anyone encoded it then they are right in doing so.
#
pfefferle
"shouldn't" mean they have to be ignored and that is exactly what the parser does ;)
#
sandeepshetty
I think in both cases the problem is with the publisher
#
sandeepshetty
but I don't think the parser ignores it right?
#
pfefferle
it does
#
sandeepshetty
can you add more content after the <br />
#
sandeepshetty
actually that is already in my test page
#
sandeepshetty
yeah was also there in my test page..
#
sandeepshetty
I'm thinking it's not right to just strip it. encoding it might be better... but again I think that should be left to the client.
#
sandeepshetty
Need to figure out why it does that in the first place to saying anything about it.
#
pfefferle
because p- should be parsed as plain text
#
pfefferle
and that means no tags
#
sandeepshetty
pfefferle: encoded but now it affects your auto-linking code :) http://notizblog.org/replies/converspace-activity-syntax/
#
pfefferle
yes, crappy wordpress ;)
#
sandeepshetty
pfefferle: just realised this might have been the same thing that affected <3
#
sandeepshetty
so the summary of our noisy discussion is that we must encode p-summary to account for things like html code snippets and emoticons <3 in it?
#
pfefferle
and that tags should be ignored
#
sandeepshetty
that part I'm not convinced about yet :) I still prefer leaving it up to the client
erikmaarten joined the channel
#
sandeepshetty
pfefferle: testing the <3 post again by re-sending that webmetnion
#
pfefferle
seems to work
#
sandeepshetty
haven't tested it yet.. did you?
#
aaronpk
sandeepshetty: ooops fixed
#
Loqi
aaronpk: sandeepshetty left you a message 3 hours, 37 minutes ago: Whoa! (RE http://indiewebcamp.com/irc/2013-06-28/line/1372404587)
#
Loqi
aaronpk: sandeepshetty left you a message 36 minutes ago: It looks like I broke the logs: Fatal error: Uncaught exception 'Exception' with message 'DateTimeZone::__construct(): Unknown or bad timezone (Asia/Mumbai)' in /web/sites/indiewebcamp.com/www/irc/inc.php:118 Stack trace: #0 /web/sites/indiewebcamp.com/www/irc/inc.php(118): DateTimeZone->__construct('Asia/Mumbai') #1 /web/sites/indiewebcamp.com/www/irc/logs.php(226): formatLine(Array) #2 {main} thrown in /
#
Loqi
aaronpk: sandeepshetty left you a message 33 minutes ago: Fixed it. Might want to handle invalid timezones to avoid killing the logs
friedcell joined the channel
andreypopp joined the channel
erikmaarten, caseorganic and tantek joined the channel
#
Loqi
tantek: sandeepshetty left you a message 4 hours, 36 minutes ago: Will try to login later tonight (your morning) and discuss 404. I'm all for constraints and starting small... just coming from the angle that 404 is a valid response code for delete.
#
tantek
sandeepshetty the problem is that 404 could mean so many other things, that you cannot interpret it *specifically* as a delete. It's too vague.
caseorganic joined the channel
#
tantek
and vagueness is bad when it comes to handling a destructive action like delete
#
tantek
whereas 410 permanent removal pretty much means a delete as specified, so that's all we need.
#
tantek
simpler design = we have 410 which works, therefore no need to confuse anyone with 404.
#
@nd_kane
@alicaurusrex @iambettywood Write good, quality content and people will find it! #contentout #indieweb #ihateseoandsoshouldyou
#
tantek
is summarizing the arguments against 404 as delete, since it still seems to be confusing.
#
tommorris
410 argument sounds reasonable
erinjo joined the channel
#
tantek.com
created /deleted-404-discussion (+6130) "move lengthy discussion of why 404 must not be considered a delete to separate page"
(view diff)
#
tantek.com
edited /deleted () "(-4166) summarize reasons why not to treat 404 as a delete, and move lengthy back/forth to a separate page"
(view diff)
#
tantek
tommorris, yeah, keeping protocols simpler seems like protocol design 101 to me.
#
sandeepshetty
tantek: converspace doesn't have delete yet, both for it's own post and for webmentions. It's next in my queue (which is why I've been thinking about this)... I don't have plans to store deleted content or mentions which makes it difficult for me to send a 410
#
tantek
doing a good job at protocols is not necessarily "easy"
#
tantek
"I don't have plans to implement" is not an excuse to screw up a protocol
#
tantek
consider keeping tombstones for deleted URLs as a part of cool URIs don't change.
bret joined the channel
#
tantek.com
edited /deleted (+284) "/* Issues */ record No plans to implement and why that's a non-issue"
(view diff)
#
tantek.com
edited /deleted (+458) "/* Issues */ additional techniques for implementation"
(view diff)
andreypopp joined the channel
#
aaronpk
all you have to do is store the URL somewhere, you don't need to store the actual deleted content
#
aaronpk
could even be done in an .htaccess file that sets the response code to 410, lol
#
tantek
aaronpk - yup
#
erikmaarten
Does anybody want to try my webmention "implementation"?
#
sandeepshetty
tantek: I guess I come from the worse is better school... for me implementation simplicity is more important than correctness
#
tantek.com
edited /deleted (+84) "add subheads to be able to reference individual issues"
(view diff)
#
sandeepshetty
storing URLs also means another query everywhere
#
tantek
will add your .htaccess comment
#
tantek.com
edited /deleted (+81) "/* No Plans To Implement */ htaccess even"
(view diff)
#
tantek
sandeepshetty - format simplicity is more important than protocol simplicity is more important than implementation simplicity
#
sandeepshetty
erikmaarten: sure, what do you need?
#
tantek
"worse is better" is never a good design principle. it's just a laziness principle.
#
tantek
formats last longer than protocols last longer than implementations. hence the priority
#
sandeepshetty
I guess we can choose to disagree on that :)
#
tantek
"simpler is better" is good. "worse is better" is not.
#
erikmaarten
sandeepshetty: Send a regular webmention to me, please. Nothing fancy, very basic support so far (works with my own posts, but that's not much of an assurance). address: http://erikmaarten.com/articles/2013/06/28/testing-webmention
#
sandeepshetty
and since everyone else seems to prefer this I will give in but note my disagreement
#
tantek
sandeep, xmlrpc is worse but that doesn't make pingback better
#
tantek
"worse is better" is a really crappy way to reason about anything
#
sandeepshetty
tantek: you can't take worse is better literally.. you have to understand what it prioritizes: http://en.wikipedia.org/wiki/Worse_is_better
#
sandeepshetty
My motivation to come up with webmention was to make something simpler to implement..
#
tantek
right, so say "simpler is better", not "worse is better"
#
aaronpk
erikmaarten: I sent you one too :)
#
sandeepshetty
but it already means something..
#
tantek
the WP article even starts with essentially saying simpler
#
tantek
"less functionality" = simpler
#
tantek
sandeep - no, people don't read it like that in summary
#
sandeepshetty
I used worse is better as a label .. I was guessing that you know the details
#
tantek
in summary, "worse is better" is treated as "lazier is better"
#
tantek
especially in conversations
#
tantek
in the WP article, "worse" has a specific meaning, meaning, less
#
sandeepshetty
yeah but it's not meant to be in quotes.. it's a label for something
#
tantek
so cut out the middlemand
#
tantek
and just say simpler
#
tantek
"simpler is better" is better than "worse is better"
#
tantek
yeah, it's a crappy label
#
sandeepshetty
but it's not just about simpler is better.. it also about the other prioritities
#
sandeepshetty
simpler over correct
#
sandeepshetty
(thought but just a little)
#
sandeepshetty
*(though only slightly)
#
sandeepshetty
erikmaarten: can't get to that url..
#
erikmaarten
aaronpk: great, thanks! Sitting in the database now.
#
erikmaarten
sandeepshetty: No? How do you mean, can't get to?
#
sandeepshetty
Oops! Google Chrome could not connect to erikmaarten.com
#
sandeepshetty
tantek: so from that perpective, I would have preferred 404 but since everyone else thinks otherwise I'll stop now on this issue and just add my objections to the page.
#
erikmaarten
sandeepshetty: that's strange. Just a temporary glitch, perhaps?
#
tantek
sandeepshetty - ironically, from that perspective, 410 is preferable since it is simpler than 404 (which is complicated by the fact that it can be returned for many different reasons)
#
tantek
I don't think you're making an actual legitimate "worse is better" argument per the WP page.
#
tommorris
"accurate is better" ;)
#
tantek
Your argument seems to amount to "I don't want to implement"
#
sandeepshetty
from an implementation perspective.. 404 is very simple.. if it's not in your db just throw a 404... when you delete just delete the db row, don't maintain ulrs, etc
#
tantek
might as well tell people that you only want CRU not CRUD.
#
tantek
would you consider CRU better than CRUD?
#
tommorris
also, the correct layer for specifying whether or not resources are available is... HTTP
#
tommorris
basic WebArch
#
tantek
sandeepshetty - and implementation is 3rd priority after formats and protocols
#
tantek
because implementations have the greatest turnover = least priority
#
tantek
and have the least in common
#
tommorris
I'e been meaning to support some publishing workflow stuff - draft posts, deleted posts, scheduled posts.
#
tommorris
returning sensible HTTP codes would be part of that
#
tantek
much much much more important to keep the formats and protocols themselves simpler
#
tantek
sandeep a lot of the "keep implementations simpler" misconception is responsible for all the XML complexity crap that occurred in the late 1990s / early 2000s.
#
tantek
and why the XML crowd objected to and/or laughed at microformats
#
sandeepshetty
tantek: how so?
#
tantek
why would you want to deal with the complexity and mess of implementing HTML? XML is so much simpler!
#
tantek
so you're just repeating a 10+ year old mistake
#
sandeepshetty
I thought that had to do more with well-formed documents than anything else
#
sandeepshetty
HTML was very forgiving in those days
#
sandeepshetty
BTW, HTML and browsers are also good examples of the worse is better philoposhy
#
tantek
nothing is such an example in a vacuum
#
tantek
that's a huge misconception
#
tantek
"worse" is relative. without providing A and B, one of which is worse, you don't have an example.
#
sandeepshetty
I thought that was obvious:worse than well formed documents, previous hypertext systems where you could not have broken links, etc...
#
tantek
another principle in design that you might be missing is forgiveness - it's part of robustness
#
aaronpk
tommorris: for scheduled posts, would you return an HTTP code that says "come back in 2 days"? lol
#
tantek
aaronpk - code 288 - ok in the future
#
sandeepshetty
erikmaarten: every time.. seems to be a DNS issue.. will look into it in a bit
#
tantek
the "don't treat 404 as a delete" is particularly important from a forgiveness perspective
#
aaronpk
288? i haven't heard of that one before
#
sandeepshetty
tantek: by forgiveness, I'm guessing you mean be liberal in what you accept.. which is what I'm also proposing
#
tantek
since server error / misconception can cause it
#
tantek
"liberal in what you accept" is not generally applicable
#
tommorris
aaronpk: might do. ;)
#
sandeepshetty
I agree.... (looks for the post by tbray on that)
#
tantek
because deletes are destructive, you have to be conservative in what you accept for them
#
tantek
forgiveness is tied to user forgiveness
#
tantek
which is very much action specific
#
aaronpk
here's a somewhat common use case where 404=delete would cause a lot of harm -- your site is in front of a lod balancer or a proxy, and a misconfiguration in the proxy causes everything to return 404s. do you really want every site downstream to delete all your content?
#
sandeepshetty
but on the indieweb deletes are not really that destructive.. you can just resend the webmention
#
sandeepshetty
(that was for tantek)
#
tantek.com
edited /deleted (+26) "/* 404 Discussion */ or proxy"
(view diff)
#
aaronpk
"you can just re-send" is actually going to require more implementation logic than just returning 410 in the first place
#
tantek
sandeepshetty - deletes are destructive in that the destination has removed the content (and potentially the proof that they got the webmention back in time when they did)
#
tantek
the time information (when they received the original webmention) may be particularly important
#
tantek
e.g. for ordering comments
#
tantek
threading
#
tantek
you can trust the time of when you received the webmention but you can't trust the asserted time the comment was written (dt-published)
#
sandeepshetty
aaronpk: I don't. (re load balancer) but that would be a rare case and not always right after you sent the webmention
#
sandeepshetty
tantek: On the indieweb, I trust the published time...
#
aaronpk
tantek: that's actually an interesting point. right now i'm using dt-published when ordering and displaying content
#
sandeepshetty
you can see my recent mentions where mentions are sent afterwards
#
tantek
sandeepshetty - you must be new here :)
#
sandeepshetty
I've done that myself also when the server wasn't responding when I wanted to send it
#
sandeepshetty
tantek: I am :)
#
tantek
it won't be long til we scale to the point where dt-published errors are occuring (both accidentally e.g. timezone, and intentionally, wanting to comment in the past)
#
aaronpk
timezone errors will be the first thing that bites us
#
tantek
already has :)
#
aaronpk
you know... i should write up something about proper timezone handling
#
tantek
Will Norris had it happen at the camp
#
aaronpk
pretty sure not many of us are doing it right
#
tantek
same thing happened with Technorati indexing posts by claimed date vs. discovered/pinged date. fixed that pretty early on
#
sandeepshetty
I spent some time trying to do it right
#
tantek
the datetime of post thing will become an issue as we go from 10s of indieweb sites to 100s.
#
tantek
it's coming soon
#
tantek.com
edited /deleted-404-discussion (+244) "proxy misconfig too"
(view diff)
#
sandeepshetty
I have the other date as well but when I'm presenting it on my site I prefer published when it's available
#
tantek
I'm going to skip trusting dt-published altogether when I start displaying others' comments
#
tantek
in terms of ordering
#
aaronpk
p.s. you may note that in the mf version of the IRC logs, timestamps are localized to the timezone of the person who spoke :) http://pin13.net/mf2/?url=http%3A%2F%2Findiewebcamp.com%2Firc%2F2013-06-28%2Fline%2F1372418432
#
tantek
it will be by discovery (possible crawl in response to http referer), and webmention datetime
#
tantek
aaronpk - WHOA
#
aaronpk
I spend a lot of time thinking about timezones :)
#
aaronpk
it makes it so when you display the reply context to a line in IRC, the timezone shows up properly: http://aaronparecki.com/replies/2013/06/28/2/indieweb
#
tantek
aaronpk - those aren't offsets!
#
tantek
you asked yesterday about h-card and timezones, I documented p-tz with an example and you chose to use it differently?
#
aaronpk
ah yea... tz is required to be a numeric offset?
#
tantek
because you data can't depend on politics
#
aaronpk
turns out it's actually way easier to deal with (and more accurate) as a named zone since there's the whole tzdata
#
tantek
all the named timezone stuff is political
#
sandeepshetty
I just normalize them all to one timezone so that they timestamps are comparable on the post page
#
tantek
there's NEVER the "whole" tzdata
spinnerin joined the channel
#
tantek
that's the problem
#
tantek
it changes in the future
#
aaronpk
given "-0700/-0800" how do I know which to use? my summer time rules are different from other zones
#
tantek
so from a storage format perspective, the only thing you can trust over time is offsets
#
aaronpk
that's true, but the other thing you can trust from a storage format perspective is the unix timestamp
#
tantek
summertime rules change in general - that's part of the problem
#
aaronpk
yes but they're all documented in tzdata
#
tantek
until politicians change them
#
aaronpk
and then they get documented again
#
tantek
they're not "all", just the laws passed so far
#
tantek
so that doesn't work for stable data / formats
#
aaronpk
so, the best solution would be to store the current numeric timezone offset for the user's IRC line at the time it is created
#
aaronpk
rather than figuring it out later
#
aaronpk
right now i'm storing unix time, then looking at the irc-people page for the offset/zone and localizing it when displayed
spinnerin joined the channel
#
tantek
yes - best is to store the offset of the user's timezone at the time the IRC line is created
#
tantek
definitely not at display time!
#
tantek
the only TZ to use at display time is the user who is reading it
#
tantek
is *for the user
#
sandeepshetty
I do it while storing... convert everything to one timestamp... could there be problems with this approach?
#
aaronpk
are you storing the timezone offset?
#
aaronpk
or storing as unix time
#
www.sandeep.io
edited /deleted (+164) "/* Three Cases */ Answered "how is this an issue?""
(view diff)
#
sandeepshetty
mysql datetime
#
aaronpk
heh. what timezone is your server in?
#
aaronpk
mysql datetime is not very useful unless your server time and database time are set to UTC
#
aaronpk
otherwise you don't really know what it represents
#
aaronpk
postgres stores the timezone offset in the datetime column, it's nice
#
erikmaarten
the mysql timestamp converts time to UTC, in contrast to datetime, may be more useful
#
sandeepshetty
aaronpk: IIRC UTC
#
aaronpk
sandeepshetty: as long as you know for sure that both your system clock and the database clock are set to UTC you can safely use datetime in mysql
#
aaronpk
but then you should also store the timezone offset in a separate column so you can re-localize the timestamp when displaying it, or you can just display everything in UTC or your local time as long as it's denoted as such
#
sandeepshetty
added "should add timezone info to timestamps" on this todo
#
aaronpk
sandeepshetty: if you add your local timezone offset to your dates, it will show up properly in my reply context http://aaronparecki.com/replies/2013/06/24/2/indieweb
#
aaronpk
right now you're showing up as +0000
#
sandeepshetty
yeah I should look into this at some point... added a note
#
tantek
aaronpk - sounds like you have material for a follow-up blog post to your one about relative times
#
sandeepshetty
erikmaarten: can't ping erikmaarten.com and traceroute says "traceroute: unknown host erikmaarten.com "
#
sandeepshetty
erikmaarten: and I can see the A record "erikmaarten.com. 14367 IN A 143.95.33.86"
#
EHLOVader
looks good from here, and there sandeepshetty
#
sandeepshetty
yeah it looks like it's just me.. just trying to figure out why and why only this site...
#
erikmaarten
sandeepshetty: well that's quite curious. It's fine from here, and aaronpk was able to send a webmention just a while ago ...
#
aaronpk
internet storms
#
erikmaarten
are you using some particular dns server?
#
sandeepshetty
I could try sending the webmention which will be from my server as opposed to my local network.... but I'm curious to find out why it's down here
#
erikmaarten
perhaps PRISM is gobbling up the network packets without sending them on
#
sandeepshetty
any ideas for what I could to find out what's wrong?
#
erikmaarten
Well, you could try changing your dns server, just to make sure there's no problem there ...
caseorganic joined the channel
#
erikmaarten
can you ping 143.95.33.86?
#
sandeepshetty
nope. can't ping that ip
#
erikmaarten
oh, perhaps you've already tried that
friedcell1 joined the channel
#
sandeepshetty
erikmaarten: tarceroute shows me that it doesn't go beyond my ISP, must be a problem with them
#
erikmaarten
sandeepshetty: strange though, if everything else works
andreypopp joined the channel
#
sandeepshetty
erikmaarten: sent you a webmention
#
erikmaarten
sandeepshetty: thanks, got it! By the way, how do you handle info about the commenter? Store url, profile photo url, name etc in separate table?
#
sandeepshetty
for receiving I have a webmentions table that has cols for author_name, author_url and author_photo along with published date, type of mention (in-reply-to|repost|like|mention) and some other fields
#
erikmaarten
right, almost like my setup, except I put the author information in one table and the mention info in another
#
aaronpk
posting events on my site is making me want to be able to create private events and selectively give access to people if they've logged in with indieauth
#
sandeepshetty
erikmaarten: you have a better design because it allows you to update the h-cards for all mentions from the same author
#
tantek
aaronpk - cool
#
sandeepshetty
erikmaarten: I'll revisit this when it becomes a problem for me... right now I want to ship and see what works and what doesn't.
#
aaronpk
also wondering how much DRY to violate by having a copy of events on my own site
#
aaronpk
for example, here's my RSVP to benwerd's event: http://aaronparecki.com/replies/2013/06/25/3/indiewebcamp but should I also post that as an event on my site?
#
erikmaarten
sandeepshetty: Yeah, that's probably about the only thing with my code that is good so far, I tend to want to test things out before setting things in stone too ...
#
tantek
aaronpk - I think it's fine to have the event embedded as part of the RSVP
#
aaronpk
and what about events like this? https://events.geoloqi.com/events/59 http://esridevelopersocial.eventbrite.com/ I consider the eventbrite links to be ephemeral. Should I make an event on my site for that which I then RSVP to?
#
tantek
and then maybe even have an "upcoming events" page
#
tantek
I've got the little module (now out of date) at the bottom of my home page for this
#
aaronpk
tantek: yea that's just the reply context, handled exactly the same way. I'm happy with the way my current RSVPs to werd.io are handled.
#
tantek
they have hCalendar, but it's not the canonical event page
#
aaronpk
I suppose I could always point back to the canonical event page with u-url even if I do have my own URL for the event
#
tantek
eventbrite links are for ticketing, they're not really permalinks
#
tantek
aaronpk - you're quickly going to run into the event consolidation problem: http://microformats.org/wiki/event-consolidation
#
aaronpk
that actually soundslike hosting my own h-event with u-url pointing to each other version of the event may help solve it
#
aaronpk
because then sending webmentions from my own RSVP url would go ping the other u-urls
#
tantek
yes, something worth exploring further now that we're starting to do federated events
#
sandeepshetty
aaronpk: DRY allows for duplication as long as there is only one authoritative source that lets you know when it's changed
#
tantek
sandeepshetty - not sure I agree with that. that's been used as an excuse by folks who put the one authoritative source in a database and then claim it's ok to have both HTML and RSS versions.
#
tantek
whereas in practice, their RSS versions break or become invalid or show bad content up to 2/3 of the time.
#
tantek
but I think I get at what you're saying - for this example. for visible events, where the "copies" visibly link to the original, it helps reduce DRY problems.
#
sandeepshetty
http://en.wikipedia.org/wiki/Don't_repeat_yourself (code generators are big with DRY enforces)
#
tantek
sandeepshetty - that's an older notion of DRY
#
tantek
the notion of DRY as we discuss it here (and have in microformats for years) is about DRY at the publishing level
#
tantek
the WP definition is good but insufficient for today's purposes
#
tantek
nope, we're still talking about content/data, not code
#
sandeepshetty
is ashamed to say that he was into Agile (with a capital A) at one point
#
tantek
OnceAndOnlyOnce is more about code
#
tantek
agile has a lot of hype - it's natural for folks to try it out ;)
#
tantek
and there are some aspects of it that can be repurposed
#
sandeepshetty
I'm all for agile with a small a :)
#
sandeepshetty
tantek: this leads into the next questions perfectly.. I've been working on a uf2 based reader because I can't keep up with all the indiewebsites now
#
sandeepshetty
but after looking at the markup differences 4 sites makes me think that either h-feed is not clearly specified or no one follows it as specified..
#
Loqi
sandeepshetty meant to say: but after looking at the markup differences of 4 sites makes me think that either h-feed is not clearly specified or no one follows it as specified..
#
tantek
sandeepshetty - it's not clear what the exact use-cases are for h-feed
#
tantek
so basically, just grab the top level h-entry objects from a page, and you've got a feed
#
sandeepshetty
I think only aaronpk has a top level h-entry
caseorganic joined the channel
#
tantek
oh interesting, right, I have mine inside an h-feed
#
tantek
I think just because I was *trying* to make h-feed work
#
tantek
in the old hAtom spec, we had the specified as the two possibilities
#
tantek
either look for a top level hfeed, and use its hentry's
#
tantek
or if there's no hfeed, then just use top level hentry's on the page
#
tantek
so you could do that with h-feed and h-entry instead
#
tantek
that should work
#
sandeepshetty
also I want to generate atom from my uf2 but h-feed seems to be missing channel name (or whatever it's called in atom)
#
sandeepshetty
(feed title)
b0bg0d joined the channel
#
tantek
here you go sandeep - my braindump notes on h-feed, and my example in wild: http://microformats.org/wiki/h-feed
#
tantek
the "feed title" you can get from the p-name
#
tantek
see mine for an example
#
tantek
I'm just using my name as the name of the feed, but you can specify something different if you want
#
sandeepshetty
this is perfect.. I will have a feed per tag (channel) so I want to name them.
#
sandeepshetty
tantek: also I couldn't find dt-published for your h-entry
#
tantek
it's there
#
tantek
it uses the value class pattern
#
tantek
rather every h-entry has a dt-published
#
tantek
I have no dt-published for my h-feed. that would be redundant (DRY violating) with the latest h-entry dt-published
#
tantek
sounds like an issue to file :)
#
tantek
<a href="2013/178/t1/surreal-meeting-dpdpdp-trondisc" rel="bookmark" class="dt-published published dt-updated updated u-url u-uid"><time class="value">10:17</time> on <time class="value">2013-06-27</time></a>
#
sandeepshetty
I was just looking at the mf2 output because that's what I'll be using and couldn't see it there.
#
tantek
so there you go, that's enough info to file the issue
#
sandeepshetty
Don't thing the h-feed requires a dt-published :)
#
tantek
thanks sandeep!
#
sandeepshetty
at this point it looks easier to use simplepie and just read atom/rss from everyone :)
#
tantek
assuming the feeds are correct :)
#
aaronpk
hfeed->rss converter for backwards compat with rss readers :)
#
sandeepshetty
that's the plan which is why I don't have rss/atom yet .. waiting on getting the h-feed right so that it can be converted to atom.. now with http://microformats.org/wiki/h-feed it looks like I should have it ready by tomorrow :)
#
tantek
aaronpk, sandeepshetty - I noted your use-cases: http://microformats.org/wiki/h-feed#Use_Cases
#
tantek
feel free to edit/clarify/add
#
aaronpk
ah cool
#
tantek
aaronpk is your original-post-discovery code separated out as a nice neat function?
#
tantek
(PHP)
#
tantek
you might have the only implementation, and it would be great if it was modularly shareable so others could re-use it
#
sandeepshetty
tantek: does the h-feed also need a fallback to rel="author" ?
#
aaronpk
hmm it isn't, but it wouldn't be too difficult for me to turn it into a module
#
tantek
sandeepshetty - already does by supporting /authorship
#
sandeepshetty
ah saw that.
#
sandeepshetty
!tell pfefferle: comments appearing on your site don't point to the original which makes it difficult to follow the thread: https://notizblog.org/replies/converspace-activity-syntax/#comments
#
Loqi
Ok, I'll tell them that when I see them next
#
tantek.com
edited /authorship (+45) "/* Determining */ use #s for steps, clarify rel-author use to go use the representative hCard of the destination page"
(view diff)
#
sandeepshetty
tantek: thanks for that clarification...
#
sandeepshetty
it didn't make sense before that to me.. had asked about it a week ago
#
tantek
yes - sorry about that - I definitely braindumped that one quickly without all the details :)
eschnou joined the channel
#
Loqi
WITHOUT ALL THE DETAILS http://loqi.me/7Z1
#
tantek
barnaby's improvements seem interesting - I'll have to analyze those later to see if they work in the general case
#
tantek
so I've lost track of everyone who is receiving/displaying indieweb comments
#
tantek
automatically that is
#
tantek
so far there's only three on the wiki
#
tantek
that's code vs. sites
#
aaronpk
how so? each project is really just someone's site
#
tantek
because the emphasis on site leads to more examples being cited
#
tantek
whereas being project-centric deemphasizes the real world examples
#
aaronpk
my goal with the implementations page was to have an "at a glance" view of the state of everyone's site
#
aaronpk
it doesn't have to be project centric
#
tantek
e.g. the Wordpress row. just "yes" is useless there
#
sandeepshetty
aaronpk: wordpress is not someones site
#
aaronpk
take off wordpress then, I didn't add that
#
sandeepshetty
there are at least two people using WP
#
tantek
WP proves the problem with this focus
#
tantek
because there is no one WP
#
aaronpk
i'd be happy with switching the first column to each person's domain name
#
tantek
there's WP with various plugins
#
aaronpk
clearly listing the projects in the first column led everyone to the wrong conclusion about the page
#
sandeepshetty
we could as pfefferle to change that WP row to talk about his site instead
#
aaronpk
i was just getting tired of digging around all the various pages looking to see who implemented what
#
Loqi
AROUND ALL THE VARIOUS http://loqi.me/7Z2
#
sandeepshetty
*ask.. oh yeah we all have project names there :p
#
aaronpk
cause it's scattered around the project pages (/p3k /Falcon etc) and also the building block pages (/comment-presentation etc)
#
tantek
aaronpk - aha ok
#
tantek
aaronpk - yes, please switch the first column to site
#
tantek
and it's fine to have a second column with "what it's running" where project(s) can be listed/linked
#
tantek.com
edited /indieweb-implementations (+52) "link displaying comments to the specific section on indieweb implementations of comment presentation"
(view diff)
#
tantek
and each column header in that table can link to the specific list of indieweb support for that thing (like what I just did with the "Displaying comments" column header)
#
tantek
also note that the table grows too wide and gets clipped by the wiki styling
#
tantek
our wiki is not really setup for wide tables like that
#
tantek
might need multiple tables
#
tantek
or minimizing the columns
#
aaronpk
hrm... side-scrolling should be a thing
#
tantek
like I'm not sure how useful "receiving comments" is as a separate column
#
aaronpk
it's also meant as a sort of checklist for individual implementors
#
tantek
well then it's too many purposes
#
tantek
checklist vs. summary/overview
#
tantek
that's quickly going to become unmanageable for one or the other
#
aaronpk
ok then let's figure it out
#
aaronpk
someone asks "what indieweb feature should I build next to interoperate better?"
#
tantek
your initial impulse of wanting a table for a usable overview is good
#
tantek
aaronpk - that's the point of IndieMark
#
aaronpk
let's tie IndieMark in with openbadges then :)
#
tantek
I don't think we need openbadges for this (though I know you meant :) )
#
tantek.com
edited /IndieMark (+17) "subheads in ideas section for easier expansion"
(view diff)
#
tantek
feel free to add to the subsections in http://indiewebcamp.com/IndieMark#idea_capture
#
@benwerd
Mozilla's released its social API to developers. I'll definitely be playing with this: http://techcrunch.com/2013/06/28/mozilla-opens-its-firefox-social-api-to-developers/ #indieweb
#
aaronparecki.com
edited /indieweb-implementations (+402) "domains instead of projects"
(view diff)
#
sandeepshetty
aaronpk: now we need a new sort order for that page :)
#
aaronpk
lol true
#
tantek.com
edited /IndieMark (+440) "/* Level 1 */ some guiding ideas for Level 1"
(view diff)
#
sandeepshetty
tantek: Was the and here meant to be an or? "if the h-entry has a p-author, use that and the representative h-card at the rel-author destination to determine the author of the post"
melvster joined the channel
#
tantek.com
edited /authorship (+19) "oops, move rel-author details to proper step. fix and Undo revision 3820 by [[Special:Contributions/Tantek.com|Tantek.com]] ([[User talk:Tantek.com|talk]])"
(view diff)
#
tantek
good catch sandeepshetty. thank you.
#
tantek.com
edited /IndieMark (+1464) "expand a bit more, inspired by all of the past week's developments"
(view diff)
#
tantek.com
edited /Falcon (+409) "/* Working On */ after indieweb replies with webmention, next is rsvp posts, inspired by benwerd's progress with federated events"
(view diff)
#
sandeepshetty
tantek: so it's basically 1) check for p-author in h-entry else 2) look for h-card with url+uid else 3) look for h-card with url + rel-me else 4) follow rel-author and repeat from step 2 there.
#
tantek
that's a reasonable summary
#
tantek.com
edited /Falcon (-1) "/* Working On */"
(view diff)
#
tantek
the url+uid step has to be to the page itself - so that's unlikely to happen
#
tantek
in the p-author case
#
tantek
rel-me is also unlikely
#
tantek.com
edited /authorship (+79) "/* Determining */ noting unlikeliness of representative hCard on a post permalink"
(view diff)
andreypopp joined the channel
#
sandeepshetty
tantek: what does "whose value is the url of the page (source)" mean here: "The first hCard found which has a "url" property whose value is the url of the page (source) and is also a "uid" property for the hCard, is the representative hCard for the page."
caseorganic joined the channel
#
tantek
the url of the page that you're parsing
#
sandeepshetty
that I'm lost about ths whole uri_uid step
#
tantek.com
edited /IndieMark (+42) "/* Level 2 */ reply context"
(view diff)
#
tantek
sandeep you're thinking too hard about it - the page itself has URL
#
tantek
"url" + "uid" are properties of the hCard
#
tantek
if all three match - then that step works
#
sandeepshetty
match what?
#
sandeepshetty
you mean they all need to be the same?
#
tantek.com
edited /IndieMark (+48) "/* Level 3 */ CRUD"
(view diff)
#
sandeepshetty
how is the page url the same as the uid?
#
sandeepshetty
Maybe I should read it again a couple of times :)
#
tantek.com
edited /IndieMark (+143) "/* Level 4 */ require all previous level optional stuff"
(view diff)
#
tantek
yes, match = all three the same
#
tantek
sorry I should probably rewrite that prose for representative hCard
bizybee and eschnou joined the channel
#
tantek.com
edited /IndieMark (+42) "/* Level 2 */ other types of replies"
(view diff)
#
tantek.com
edited /IndieMark (+78) "/* Level 3 */ more types of replies to receive and display"
(view diff)
#
tantek.com
edited /IndieMark (+58) "/* Level 4 */ receiving/displaying multiple types of comments/replies"
(view diff)
#
sandeepshetty
tantek: so it's basically this then: 1) check for p-author in h-entry else 2) follow rel-author, 2a) look for h-card with url+uid that points to the rel-author link else 2b) look for h-card with url + rel-me
#
tantek.com
edited /IndieMark (+386) "/* Level 4 */ perhaps full comment posting/receiving CRUD is the key to level 4"
(view diff)
#
aaronpk
why are there levels 1, 2, 3, etc if it's not supposed to be a rank?
#
tantek
aaronpk - it's not *only* a rank
#
tantek
think in 2+ dimensions. rank is only 1 dimension
#
tantek
arronpk - like this: http://rng.io/
#
tantek.com
edited /IndieMark (+144) "/* FAQ */ add subhead permalink, link, A2"
(view diff)
#
tantek.com
edited /IndieMark (+1) "/* Not just a rank */"
(view diff)
fmarier joined the channel
#
tantek
ok I'm off for a bit. hoping to somehow get indieweb replies with webmentions working today. we'll see.
KevinMarks joined the channel
#
sandeepshetty
!tell tantek: Let me know if this authorship algo makes sense (I'm looking at fixing my current authorship code and it helps me understand how to markup my site): 1) check for p-author in h-entry else 2) follow rel-author, 2a) look for h-card with url+uid that points to the rel-author link else 2b) look for h-card with url + rel-me
#
Loqi
Ok, I'll tell him that when I see him next
#
www.marcus-povey.co.uk
edited /2013/UK (+346) "/* RSVP */"
(view diff)
friedcell, caseorganic and pfefferle joined the channel
#
Loqi
pfefferle: sandeepshetty left you a message 1 hour, 43 minutes ago: comments appearing on your site don't point to the original which makes it difficult to follow the thread: https://notizblog.org/replies/converspace-activity-syntax/#comments
#
pfefferle
sandeepshetty yes, removed because i thought is was a bug
#
pfefferle
will re-add it
#
@Superglufon
RT @t: federated #indieweb comments, likes, event + RSVPs demo by @benwerd:
#
@nchenga
is there a POSSE (Publish on own site, syndicate everywhere) plugin for Twitter and a self-hosted #Wordpress? http://indiewebcamp.com/POSSE
caseorganic joined the channel
#
@wavis
RT @datacosmos: #dataecosystems #indiewebcamp Mycelial networks of autonomous nodes will replace the skyscraper models of today.
poppy joined the channel
#
sandeepshetty
pfefferle: replies are not showing up on the comment permalink page: https://notizblog.org/replies/converspace-activity-syntax/?replytocom=214583 (it's there on the main thread: http://notizblog.org/replies/converspace-activity-syntax/)
#
sandeepshetty
pfefferle: should I file a bug?
#
pfefferle
sandeepshetty: yes, if you open a comment, then the context is missing
#
pfefferle
yes, it would be nice if you could also report your ideas or mentions or critic!
#
@pfefferle
@nchenga btw. you can find a list of wordpress plugins here http://indiewebcamp.com/WordPress
erinjo, bnvk and tantek joined the channel
#
Loqi
tantek: sandeepshetty left you a message 2 hours, 11 minutes ago: Let me know if this authorship algo makes sense (I'm looking at fixing my current authorship code and it helps me understand how to markup my site): 1) check for p-author in h-entry else 2) follow rel-author, 2a) look for h-card with url+uid that points to the rel-author link else 2b) look for h-card with url + rel-me
bret and poppy joined the channel
#
lionzan.me
edited /2013/UK (+528) "/* RSVP */"
(view diff)
#
pfefferle
sandeepshetty: thanks :)
scor and tantek joined the channel
#
@xtof_fr
#repos et #sommeil. Prochaine fête #indieweb 2013-07-10
#
sandeepshetty
wow, it's interesting that PuSH 0.4 does not specify the way to ping a hub about new content.... changing a hub now might involve re-integration.
#
tantek
sandeepshetty - that sounds like a good iteration/simplification
#
sandeepshetty
tantek: If I understand it correctly you can take out the representative card step for the post permalink page (like you've noted)
#
tantek
editing now
#
ozten
spent the day steeping in camlistore
#
tantek.com
edited /authorship (+591) "/* Determining */ simplify authorship algorithm based on suggestions from sandeepshetty, incorporate inline the representative h-card steps. note limitations and encouragement of implementation heuristics documentation"
(view diff)
scor and b0bg0d joined the channel
#
@sandeepshetty
#webmention forces implementations to treat #comments as first-class content by giving them permalinks... http://www.sandeep.io/96 #indieweb