#social 2015-05-14

2015-05-14 UTC
bengo and jasnell joined the channel
#
ben_thatmustbeme
elf-pavlik: obviously its better to use https for security reasons. People have mostly not wanted to require https because new users find it difficult to set up their own ssl cert, or just don't want to spend any money
jasnell, bblfish, bengo, tilgovi, KevinMarks, the_frey, bret, kylewm, shepazu, rhiaro, raucao, the_frey_, melvster1, elf-pavlik, wilkie, rhiaro_, jaywink and stevenroose joined the channel
#
elf-pavlik
ben_thatmustbeme, re: "new users find it difficult to set up their own ssl cert, or just don't want to spend any money", as I said by the time specs we work on here will reach status of W3C TR, https://letsencrypt.org should work and more similar solutions start to appear
#
elf-pavlik
action-64
#
trackbot
is looking up action-64.
#
trackbot
action-64 -- Pavlik elf to Document reasoning for requiring https (possibly coordinate it with harry) -- due 2015-05-20 -- OPEN
#
elf-pavlik
oshepherd, http://foo.me != https://foo.me and if people start using http: URIs for their identity, switching them to https: later might come with some challenges
#
oshepherd
That's another radon
#
oshepherd
Reason***
the_frey joined the channel
#
ben_thatmustbeme
we've found in practice its better to treat https:// == http:// or at least provide a notification of the change and upgrade their account on said system
#
ben_thatmustbeme
technically facebook.com/me and facebook.com/ME are different, but always treated the same. most web servers force case incensitivity. Don't get me wrong i would prefer https:// everywhere
#
elf-pavlik
IMO using https://foo.name/bar and http://foo.name/bar for different resources also doesn't really make much sense
#
ben_thatmustbeme
it doesn't. its technically possible, but seems rather foolish, far to many applications assume them to be the same
gx joined the channel
#
ben_thatmustbeme
we have started to suggest people warn those logging in without https and recommend they upgrade, but we don't require it
#
ben_thatmustbeme
I could see requiring it for private messaging though.
#
elf-pavlik
using HTTP with IndieAuth sounds like pretty bad idea, especially currently with rather centralized indieauth.com, i think on hot spots someone could rather easily explit redirects and spoof ...
#
elf-pavlik
better leaves it to security experts to deal with it..
#
elf-pavlik
afk for ~2h
#
ben_thatmustbeme
tries to think through indieauth for a sec. well there are 2 layers to it. I doubt anyone would be running their server on a local wifi. if you are trying to log in to a site with HTTPS via indie-auth a person would have to be sniffing traffic between the servers (you never load your own site directly)
#
ben_thatmustbeme
if you are logging in to a system without https.... all you post is your URL, that redirects you to https on indieauth, where you have actually auth, but the redirect links back to your site are able to be sniffed and thus gain access
jasnell joined the channel
#
ben_thatmustbeme
its not giving away ability to log in generically, but you could steal a single log in, which depending on how much access is give, that could be bad, yeah
#
elf-pavlik
ben_thatmustbeme, can you elaborate on "how much access" part?
#
elf-pavlik
will need to run in 15min but will check logs later
#
elf-pavlik
if i run micropub client, can i have any restrictions on what i can post to people's micropub endpoints of for how long i can write to them?
#
elf-pavlik
action-62
#
trackbot
is looking up action-62.
#
trackbot
action-62 -- Pavlik elf to Document possible danger of malicious apps when moving more responsibilities to clients -- due 2015-05-12 -- OPEN
#
ben_thatmustbeme
elf-pavlik, that depends entirely on their implementation. I expire tokens for my micropub endpoint after a month going unused. I was talking more about auth vs getting a token
#
ben_thatmustbeme
for micropub i would definitely strongly recommend https
#
elf-pavlik
needs to catch up on authentication vs. authorization in IndieAuth
#
ben_thatmustbeme
although its possible even when its only auth to just rewrite their original request to ask for post access
#
ben_thatmustbeme
elf-pavlik: its pretty simple, if there is a scope defined after authentication, then it goes on to the token endpoint to get an auth token for that scope
#
ben_thatmustbeme
you see the scope requested on indieauth when logging in, it clearly says what scopes it is asking for. but I'm sure many people will ignore that
#
elf-pavlik
ben_thatmustbeme, where i can sign in to see request of access for certain scope?
#
ben_thatmustbeme
any micropub endpoint
#
ben_thatmustbeme
s/endpoint/client
#
elf-pavlik
"No Auth Endpoint Found"
#
ben_thatmustbeme
just type in my URL
#
elf-pavlik
https://monocle.p3k.io/ logs me in but i don't see it asking for scope
#
ben_thatmustbeme
ahh, yeah, i need to recheck that. i believe i limit to no dirs
#
elf-pavlik
remoteStorage does scopes per 'module'
#
elf-pavlik
contacts, todos etc.
#
ben_thatmustbeme
i added edit and delete. i believe its something that would have to be standardized
#
ben_thatmustbeme
as your site has to know what the hell 'edit' means
#
elf-pavlik
so app for events don't messes with contacts etc.
#
elf-pavlik
ben_thatmustbeme++
#
Loqi
ben_thatmustbeme has 73 karma
#
elf-pavlik
okie doke, got to run :)
#
ben_thatmustbeme
i think that seems pretty reasonable post, contacts, etc
#
elf-pavlik
calendar
#
elf-pavlik
wishlist
#
ben_thatmustbeme
definitely something that would have to be thought through. I don't think i'd want one for every post type 'post', 'checkin', etc
#
elf-pavlik
needs to check how Android defines access categories
#
elf-pavlik
IMO IndieWeb seems very blogger centric while i see social going much broader than posting notes
#
ben_thatmustbeme
i'd say thats a fair evaluation. Its mostly where the simplest starting point is.
aaronpk, bengo, tilgovi, elf-pavlik_, the_frey and LCyrin joined the channel
#
wilkie
LCyrin: hi
#
LCyrin
hey wilkie
#
LCyrin
wilkie, what do you use for image uploads on rstatus
bengo joined the channel
#
wilkie
what I've been experimenting with is putting them into the DB actually
#
wilkie
because mongo dbs, for instance, have a free tier, you can just put them in the store. what mongo calls the "Grid"
#
wilkie
rstatus doesn't have image upload, this is the bigger stuff
#
elf-pavlik
raucao, have you looked at https://demo.cozycloud.cc/#home ?
#
elf-pavlik
it uses vCard and iCalendar (also CardDAV and CalDAV), both vCard and iCalendasr have RDF definitions published by W3C
#
elf-pavlik
we could focus on getting those working across remoteStorage, CozyCloud and Social Data we work with here
#
elf-pavlik
SoLiD also has contact manager and meeting scheduler apps https://github.com/linkeddata/SoLiD#applications
#
raucao
sounds good
#
bengo
nice share
#
bengo
What is the point of the as:Article objectType that just extends as:Content but doesn't encourage any new properties?
#
bengo
The only qualifier for that subset of as:Content is 'multi-paragraph'. Does that mean if I'm accepting user input, I should call one-paragraph Content as:Content and everything else as:Article (or extension)?
#
bengo
(same for as:Note, which is non-normatively ""A Short note, typically less than a single paragraph. A \"tweet\" is an example, or a \"status update\""@en ."
bengo joined the channel
#
elf-pavlik
jasnell, maybe you can answer bengo's questions? ^
KevinMarks joined the channel
#
bengo
kinda seems like a 'no right answers' sort of thing. I'm implementing now and may just make everything Content
#
bengo
or our own extensions
#
jasnell
bengo: as:Content is the top level, as:Article is intended for longer multi-paragraph, while as:Note is for one paragraph or less
#
jasnell
note that other things derive from as:Content also
#
bengo
Does that mean if I'm accepting user input, I should call one-paragraph Content as:Content and everything else as:Article (or extension)?
#
jasnell
as:Document for instance
#
jasnell
the general rule has always been the prefer the most specific type. so instead of using as:Content, use as:Note for anything one paragraph or less
bblfish joined the channel
#
bengo
Got it, thanks for the context :)
#
bengo
@jasnell It'd be awesome if someday the utility of this lib was eventually baked into browsers. https://github.com/jasnell/linkeddata-vocabs
#
bengo
That way every lib that needs to say "Is this LD object a subclass of as:Content" wouldn't need to bundle it itself
#
bengo
*browser lib
#
bengo
e.g. yesterday I tried to browserify/uglify/gz https://www.npmjs.com/package/activitystrea.ms and it was 88k gripped https://cloudup.com/cI5DiuvvCea
#
bengo
not that that was necessarily because of linkeddata-vocabs, but every byte saved is nice!
almereyda joined the channel
#
elf-pavlik
issue-12
#
trackbot
is looking up issue-12.
#
trackbot
issue-12 -- Action Types Structure and Processing Model -- closed
#
bengo
thanks @elf-pavlik
#
bengo
Or maybe a good browser feature is a json-ld expansion feature that automatically knows about type inferences from w3c-published ontologies like AS/Annotation
LCyrin and bblfish joined the channel
#
elf-pavlik
action-43
#
trackbot
is looking up action-43.
#
trackbot
action-43 -- Pavlik elf to propose *lightweight* inference based on RDFa Vocabulary Expansion -- due 2015-04-14 -- OPEN
#
elf-pavlik
going to sleep :)
#
aaronpk
does activitypump have a way to search things? or is that out of scope?
Arnaud joined the channel
#
shepazu
reads "lightweight" and "RDFa Vocabulary Expansion" and chuckles :)
#
bengo
aaronpk not tight now afaict
#
bengo
http://linkeddatafragments.org/ would be a good pattern to support for it
#
bengo
Anyone ever implemented AS2 for 'flag as spam' or 'flag as inappropriate'?
#
bengo
How do you model the difference?
#
bengo
Discrete subclassOf as:Flag?
#
bengo
or target: property of as:Flag instance?
bengo joined the channel
#
bengo
nvm the wiki answers "The Flag activity is generalized here without specifying detail about *why* the content was flagged. Subclasses of the Flag activity can be used to provide more detail... e.g.
#
bengo
:FlagAsAbusive a as:Activity ;
#
bengo
rdfs:subClassOf as:Flag ."
#
jasnell
bengo: we've implemented AS1 for flag as spam, not AS2 yet tho
#
bengo
For now I will do
#
bengo
* as:Flag
#
bengo
- lf:Flag
#
bengo
- lf:FlagSpam
#
bengo
- lf:FlagOffensive
#
bengo
- lf:Disagree
#
bengo
- lf:OffTopic
#
jasnell
in our impl we use an extension to identify the difference. it's not ideal tho
#
jasnell
that's a good approach. Or, just lf:Spam, then do "@type": ["as:Flag", "lf:Spam"]
#
bengo
ooh
#
bengo
interesting
#
bengo
hard for me to think about what the independent semantics of lf:Spam would be there
#
bengo
e.g. how to interpret [as:Content, lf:Spam]
#
bengo
According to http://www.w3.org/TR/2015/WD-activitystreams-vocabulary-20150129/#dfn-mention, the as:Mention ObjectType is owl:subclassOf as:Link
#
bengo
however: as:Link a owl:Class ;
#
bengo
owl:disjointWith as:Object ;
#
bengo
Wouldn't that mean that as:Mention is not an as:Object. And then it's weird that it's listed under "Object Types"?
#
bengo
it is the only "Object Type" that extends as:Link
#
jasnell
right... Mention is not an Object, it's a Link
#
jasnell
yeah, that part is a bit weird
#
jasnell
may need to restructure that
#
jasnell
Mention is a link... so you'd expect something like {"@type": "Mention", "href": "http://example.org/foo"}
#
jasnell
the intent is that Mentin is a kind of Link that you're calling special attention to
#
bengo
thx
#
oshepherd
bengo: Take as:Note vs as:Article as presentational suggestions
#
oshepherd
e.g. as:Article might be rendered with serif fonts in a different style, while as:Note would be more compact
#
oshepherd
bengo: Basically think of as:Note as more short form (e.g. tweet, "status update") and as:Article as more long form (an article, blog post, etc)
melvster1 joined the channel
#
melvster1
elf-pavlik: so does your webid work?
bengo, tilgovi and Arnaud1 joined the channel