#social 2015-07-16

2015-07-16 UTC
bigbluehat, KevinMarks, jasnell, tessierashpool_, cwebber2, ben_thatmustbeme, tommorris_, bret, mattl, JakeHart, dwhly, ElijahLynn, Arnaud, shepazu_, the_frey and bblfish joined the channel
#
melvster1
raucao: what we learnt working on webid after the first few years is was to cleanly separate the concept of identity with the concept of verification ... as the web has grown it's become less and less practical to prescribe a one auth system fits all approach to security ... so webid itself is now about how to identify, name and represent a user ... it is authentication agnostic ... other large systems like the postal service, email, and telephone often
#
melvster1
also take this approach
#
melvster1
PS : I think that's an important lesson for this WG to learn from too if the goal is an interoperable social API
bblfish_, jasnell, bblfish and jaywink joined the channel
#
raucao
it does't change the fact that you want to know who sends you something, though
bblfish, jasnell and the_frey joined the channel
#
melvster1
regarding the IG call today, i've updated https://github.com/w3c-social/social-ucr
#
melvster1
i believe the WG has approved some of the stories officially now
#
melvster1
from the page we currently (from 13 user stories)
#
melvster1
proposed: SoLiD 5. AP 4. IWC 4.
#
melvster1
implementations: SoLiD 5. AP 0. IWC 1.
#
melvster1
i believe there was some advancement on one of the use cases tho, which will be added soon
jasnell joined the channel
#
melvster1
some of us working with SoLiD are now looking at implementing inbox and direct messaging
#
melvster1
so hopefully getting to the half way point in user stories ... which is not bad imho given SoLiD is only 3 months old as a spec!
#
cwebber2
hi everyone
jasnell, Arnaud and tilgovi joined the channel
#
aaronpk
melvster1: where are those numbers coming from? (5,4,4 and 5,0,1)
#
ben_thatmustbeme
melvster1: where do you get 13 user stories?
#
ben_thatmustbeme
we only have 8 approved
#
ben_thatmustbeme
ahh, i see, adding in 4 that have been revised on github and one that is just made up because there is an implementation
#
ben_thatmustbeme
which was pretty much blah on all votes and likely to not get accepted as a user story
#
rhiaro
I'm confused by that first grid, 'approved user stories' - what do the ticks mean?
#
rhiaro
oh, the descriptions according to each protocol
#
aaronpk
man that Inbox story is horribly named
#
aaronpk
also wow there is a lot in it
#
aaronpk
posting a video to a group, posting a sound file to a specific person
#
aaronpk
and seeing that a friend has just added a new friend
#
rhiaro
melvster1: https://github.com/w3c-social/activitypump/blob/master/userstories/inbox.md gets a tick in the first table, but doesn't actually contain any content
#
rhiaro
is looking through the others
#
aaronpk
i know we re-visited this on the last call, and realized that the wiki is out of date w.r.t. the decision to approve some more user stories
#
aaronpk
where is that list of ones that should be approved? the main list doesn't have them under "approved" yet
#
rhiaro
http://ods.openlinksw.com/dataspace/doc/dav/wiki/ODS/ODSSWATOTutorial (linked under SoLiD-SWAT0) is a tutorial for how to use the GUI of a piece of software. Is that relevant?
#
rhiaro
It is very specific to that software, rather than a protocol. And is older than SoLiD, so I don't see how it could be implementing SoLiD.
#
aaronpk
also it doesn't count as passing SWAT0 unless there are different implementations
#
rhiaro
aaronpk: It's listed as a walkthrough (first table) rather than implementation
#
rhiaro
The implementations table doesn't have SWAT0 listed at all
#
rhiaro
I assume that's what the first table is, anyway, even though the SoLiD links are a paper and a tutorial
#
aaronpk
i actually don't understand the organization of this page at all https://github.com/w3c-social/social-ucr
#
rhiaro
yeah, me neither
#
aaronpk
for example, why is "existing implementations" a subheader of "proposed implementations"
the_frey and KevinMarks joined the channel
#
rhiaro
!tell melvster1: I think I recall you asking for a list of micropub implementations to be on the WG wiki at one point? There are links to lists already here https://www.w3.org/wiki/Socialwg/Social_API/Candidates#Micropub_implementations
#
Loqi
Ok, I'll tell them that when I see them next
the_frey joined the channel
#
KevinMarks
jasnell: where does "inbox" come from instead of replies? That's an email analogy, not anything any social network I have seen calls it
#
jasnell
pump.io does
#
jasnell
has for some time
#
aaronpk
most social networks call it "notifications" or "timeline" at the very least
#
jasnell
"inbox" is simply a more generic name.
#
rhiaro
or stream
#
aaronpk
reading the "inbox" user story again, I'm pretty sure everything falls under "timeline" as facebook would describe it
#
jasnell
"inbox" is semantically clearer than "timeline"
#
jasnell
it's the place you go to see the stuff that's been directed towards the object
#
rhiaro
inbox has the connotation of being able to mark things as read, which seems misleading to me
#
KevinMarks
if I search fro "replies" on google, twitter and facebook pages are top hits; I think that implies a generic understanding
#
jasnell
why? in any reasonable user interface that displays these things, I would fully expect the ability to mark things as being read
#
rhiaro
maybe just 'in' and 'out'
#
aaronpk
I disagree with the framing of that, since 90% of the stuff in my "home timeline" or whatever on twitter I never read even though it was "directed" at me as described in this context
#
ben_thatmustbeme
i think of inbox as being able to insert things from external sources, whereas a timeline is reading only
#
rhiaro
The facebook timeline doesn't have read and unread
#
aaronpk
do you mark things as read on twitter?
#
KevinMarks
not having to ark thinsg as read is a key distinction of social networks
#
rhiaro
And I'm not sure that pump.io even does actually, I'll check
#
aaronpk
right. otherwise it's just email
#
ben_thatmustbeme
the "inbox" (generic) referenced in user stories gets really confused with "inbox" (techinical item) in pump.io
#
ben_thatmustbeme
feel like we need to get that all sorted out
#
rhiaro
I can't see any indication in pump UI that I've read something or not
#
jasnell
aside from the name of the property, is there any issue with what I'm proposing?
#
ben_thatmustbeme
what is store for?
#
KevinMarks
not being an inbox is a key distinction of a social network from email
#
ben_thatmustbeme
was that added for soem user story?
#
rhiaro
It's a name for the endpoint to send stuff, right?
#
rhiaro
Sorry, a name for the relationship between a user (Profile?) and the endpoint to send stuff?
tilgovi_ joined the channel
#
jasnell
the store is to point to the collection of content type things. it's where you'd expect to find your photo albums, or documents, etc. Whereas "outbox" would be used for outbound notifications of activities, the "store" is where the actual content is managed
#
jasnell
so as a user I may upload a photo into an album represented in the "store" collection, and the system would drop an "Update" activity into the "outbox"
#
rhiaro
(sorry, I was talking about 'inbox')
#
rhiaro
What's the origin of store?
#
jasnell
origin?
#
rhiaro
Is it derived from an existing spec or implementation, or is this new in AS2? Seems unfamiliar, that's all
#
ben_thatmustbeme
so, do you need store if you have collection? is it really just implementation details at that point?
#
melvster1
i didnt make that page, I was just pointed it at, and asked to fill it in ... suggest adding your improvement suggestions to the issues list on github
#
jasnell
consider facebook. facebook let's you upload photos and videos, right? while those are displayed directly in the timeline, they are not actually part of your timeline. You can go to your separate "photos" area to see those
#
melvster1
jasnell: on facebook, photos are contained on your timeline, they are also contained in a gallery
#
rhiaro
doesn't a facebook gallery map onto a collection?
#
jasnell
"store" is essentially a generic form of that separate "photos" area. Consider the collection that it points to as a root in a file system. It can contain the content items directly, or it can contain pointers to other collections (albums, folders, etc) are organized from that point
#
melvster1
is implementing inbox as we speak ...
#
jasnell
when I upload a photo, to the service, an activity would be dropped into "outbox", and the photo itself would be found under "store" somewhere
#
jasnell
"outbox" is essentially your own personal timeline, "store" is your gallery
#
melvster1
for me the hardest part has been to GET a list of *part* of the inbox, as opposed to, *the whole* of the inbox stream
#
jasnell
"inbox" => notifications, "outbox" => personal timeline, "store" => gallery, "relationships" => friends, connections, blocks, etc
#
ben_thatmustbeme
actually it FB photos are pretty heavily linked to the post of them
#
ben_thatmustbeme
comments are shared
#
ben_thatmustbeme
also trying to find a User Story that requires it
#
aaronpk
jasnell: why is the distinction between "store" and "relationships" important?
#
jasnell
relationships can, conceivably, be just another content object to be found under store. However, they are distinct enough in terms of use that it warrants separating them
#
melvster1
aaronpk: because you cant have a social web without friends (ie relationships)
#
aaronpk
melvster1: lol that's not what i said
#
melvster1
jasnell++
#
Loqi
jasnell has 25 karma
#
jasnell
only 25? damn
#
ben_thatmustbeme
jasnell++ for dealing with a million questions right now
#
Loqi
jasnell has 26 karma
#
ben_thatmustbeme
you asked for feedback though :P
#
aaronpk
jasnell: i guess I don't see why a named/typed collection couldn't be used for both photos and friends
#
jasnell
aaronpk: it could, but that increases the overhead when all you want is a list of your friends.
#
jasnell
in other words, I want to pull up your list of friends so I can see who our shared contacts are
#
aaronpk
or at least pull up the list of friends I am willing to publish
#
melvster1
jasnell++
#
Loqi
jasnell has 27 karma
#
rhiaro
Why does 'store' need to be in the vocabulary? Isn't where to store things something that implementations would just decide for themselves? What's a use case for needing to discover this relationship?
#
jasnell
Discoverability *is* the use case
#
melvster1
this ^^ (im not going to +! *every* thing he says) :)
#
jasnell
awww. I need that karma boost
#
Loqi
cute!
#
jasnell
for instance, let's say we didn't have "relationships" as a separate property and just lump my friends into my "store"
#
rhiaro
Hmm. But can't you discover where a photo (or anything embedded) is stored from a post by looking at the src in the post content?
#
aaronpk
wait is this just a single "store"?
#
aaronpk
like, there's only one of them?
#
jasnell
consider "store" as being the root of a virtual fs
#
aaronpk
so I would make a "friends" folder and put them there
#
jasnell
it's a collection whose items represent the things that are accessible via that root
#
jasnell
right. so then, if I want to know how to find your friends folder consistently, how do I do that?
#
aaronpk
i could point to it with some vocab from my home page
#
jasnell
which vocab?
#
aaronpk
rel=friends for example (probably a bad exmaple)
#
aaronpk
as:friends? i dunno
#
jasnell
ok, that's exactly what "relationships" does
#
aaronpk
but why is it different from store?
#
jasnell
it shortcuts the discoverability
#
melvster1
jasnell: there's a user story for browsing the friendship graph ... SoLID have actually implemented this already ... the idea is that profiles have links to other profiles which become relationships (of a certain type) ... I believe indieweb have looked at "blogrolls" in the past, but I dont believe it's yet implemented anywhere
#
jasnell
let's say I have store with {"@type": "Collection", "items": [{"@type": "Collection, "displayName": "My Friends"}]}
#
jasnell
but there can be any number of other items in that "store" collection also
#
jasnell
to find the friends collection, I have to either dig through the items, or try to find a rel=friends kind of pointer
#
jasnell
locating the friends list is a common enough use case that I ought to be able to short cut that process, and simply put "relationships" as a pointer to the "My Friends" collection directly
#
jasnell
if an implementation wishes to manage relationships as content items in the store, then "relationships" is a kind of symbolic link that makes it easier to find
#
aaronpk
does this handle "friends" vs "following" vs "followed by" distinctions somehow?
#
jasnell
no, "relationships" only shortcuts it so far. Under "relationships" you'd expect to see all the relationships that involve that object. friends, follows, followed-by, blocks, etc. The API would need to provide some reasonable means of filtering that collection
#
jasnell
in other words, this gives us the entry point for discoverability but does not solve the full filtering problem
#
jasnell
it's conceivable that extension properties can be used to point to the more specific sub-types of relationships, but I'd rather not add too much extra to the core vocabulary
#
rhiaro
+1 extensions for that
#
aaronpk
okay, "relationships" makes sense. but what is "store" for then?
#
aaronpk
it seems that it's an implementation detail at that point
#
jasnell
also, keep in mind that these properties are not functional... they can have more than one value. for instance "relationships": [{"@type":"Collection", "displayName": "Friends", ...}, {"@type":"Collection", "displayName": "Blocks"}]
#
melvster1
in SoLiD we call stores workspaces, it's where you stort stuff that you or people in your relationships want to find
#
melvster1
s/stort/store
#
melvster1
ie your social data
#
rhiaro
workspaces in solid are more for access control, right?
#
melvster1
no
#
melvster1
just for storing stuff
#
melvster1
you can access control them ... but that is an advanced use case
#
rhiaro
so then, they're like collections?
#
melvster1
yes
#
jasnell
aaronpk: the "store" is where all of the various content artifacts are. So, if you're talking about a blog, that's where the actual blog posts live. If you're talking about a photo album, that's where the photos live.
#
melvster1
"containers" in LDP language
#
aaronpk
jasnell: I guess I don't see why that's part of the API, that just sounds like how you store things after the API request is made
#
aaronpk
this is supposed to be an API rather than a blueprint for building a system, right?
#
jasnell
again, the use case here is discoverability
#
jasnell
for instance, "I don't want to see James' timeline, I want to see his photo gallery. Where do I find that?"
#
melvster1
aaronpk: an API *is* a blueprint for building a system ... it exposes data for another party to interact with in order to build certain types of system
#
aaronpk
melvster1: an API should not make any assumptions about *how* the API is implemented, and should not dictate how to do things behind the API layer, so no
#
rhiaro
can I see an example of how 'store' helps to discover a photo gallery?
#
jasnell
ok.. assume that we're talking about an implementation that *only* does photo storage for a user
#
melvster1
if doing it like facebook ... you would put photos in posts, and then put posts in a gallery
#
jasnell
{"@type": "Person", "displayName": "James", "store": {"@type":"Album", "displayName": "James' Photos", "url": "http://james.example.org/album", "items": [...]}
#
melvster1
*i think*
#
rhiaro
So anything can be a store, and this property just points to it. What if there's more than one?
#
melvster1
jasnell? was album supposed to have a trialling / there?
#
jasnell
more than one is no problem
#
rhiaro
so then it's up to whatever is retreiving them to sift through?
#
jasnell
you can either have multiple values for "store" or "store" point at a collection whose items are sub-collections
#
jasnell
yes, like I said with relationships, this only gets you so far
#
jasnell
it provides the entry point for discoverability, it does not do all the filtering for you
#
jasnell
extensions and API details can provide the rest
#
rhiaro
So it's technically the same as relationships (ie. a relationship to a Collection) but from the semantics you'd expect to find different types of things there?
#
rhiaro
But the different types of things are not specified
#
jasnell
well, "relationships" would imply that you're only looking at as:Relationship objects
#
jasnell
"store" would imply that you're primarily looking at content objects
#
rhiaro
Okay. Did that ought to be explicit?
#
jasnell
but, as has been pointed out, an implementation could lump as:Relationship objects in there also
#
jasnell
I could make it more explicit yes
#
jasnell
"outbox" and "inbox" would imply that you'd be seeing as:Activity objects
#
rhiaro
I'm not sure if it should be or not.
#
rhiaro
ponders
#
jasnell
we can bikeshed about whether the *names* "outbox", "inbox", "store" and "relationships" are the right names, but regardless of what you call them, the equivalent concepts do exist in plenty of existing social platforms
#
jasnell
"outbox" are your outbound notifications (your personal timeline), "inbox" are your inbound notifications (what others have done that you're being made aware of), "store" is your gallery/content store and "relationships" are the social graph connections of various types
#
ben_thatmustbeme
usually a "inbox" is only viewable internally (at least in most systems)
#
jasnell
for People, yes
#
ben_thatmustbeme
would an API ever need to read that? oh, for social, not for federation i suppose
#
jasnell
{"@type": "Article", "inbox": {"@type": "Collection", "items":[ ... ] } }
#
jasnell
here, inbox would serve the same basic role as "replies" (which this PR removes)
#
jasnell
{"@type": "Event", "inbox": {"@type": "Collection", "displayName": "RSVPs", "items": [ ... ] } }
#
ben_thatmustbeme
still not sure inbox is a good term, its confused with "inbox" as an endpoint for others to send messages to vs "inbox" as a feed of my notifications
#
ben_thatmustbeme
this would be a notification feed really
#
aaronpk
it sounds like it's more than notifications
#
jasnell
having a term that is too specific can be just as dangerous as one that is too generic
#
jasnell
in particular, it can constrain how developers make use of it and give the impression that uses that would be perfectly acceptable in practice need an extension
#
jasnell
in this case, there's really nothing wrong with viewing "inbox" as an endpoint to send messages
#
jasnell
that is, an implementation could easily you that URL as the location for sending pings across systems to notify a user of some event
#
jasnell
HTTP POST to add items to my notifications, HTTP GET to view my notifications
#
jasnell
s/you that URL/use that URL
#
ben_thatmustbeme
? thats a pretty big difference to me, an inbox as an endpoint is write, vs an activity stream which is read
#
jasnell
"An activity inbox at /api/user/<nickname>/inbox. This is where the user can read posts that were sent to him/her. Remote servers can post activities here to be delivered to the user"
#
melvster1
jasnell: looking at the user stories the term "inbox stream" seems to be used in conjunction with "inbox"
#
melvster1
just like an activity stream is also not an activity
#
jasnell
melvster1: in my experience, those user stories tend to be horribly written with very imprecise language
#
jasnell
but that's just me
#
aaronpk
i agree lol
#
melvster1
jasnell: good point, i joined this group after they were voted on, so just am trying to understand them as well as possible
#
jasnell
ben_thatmustbeme: "inbox" does not *have* to be the same place you post, but it can be
#
melvster1
swat0 imho was the most difficult to understand, it doesnt surprise me that it took 5 years for even an *attempt* to be made
#
jasnell
again, the use case here is discoverability. Where can I go to find the inbound notifications. If we feel that it is necessary to split the "where do I read my notifications" from "Where do I deliver notifications", then we can define an alternate means of discovering the "where do I deliver"
#
jasnell
but in practice, there's nothing wrong with using the same url for both
#
jasnell
so long as your authentication/authorization is properly configured :-)
#
melvster1
jasnell: that's exactly how I implemented it (in the last week) ... different use case tho (social payments)
#
melvster1
i have an inbox of where i send stuff
#
jasnell
Unfortunately I'm going to have to get going, need to get on to some other things
#
melvster1
and then an area where I read stuff that goes onto the screen
#
jasnell
I'll be back on tomorrow morning if you'd like to continue the discussion
#
jasnell
or we can continue in the github pull request
#
jasnell
bye all!
#
jasnell
is stepping away from the keyboard
jasnell joined the channel
#
ben_thatmustbeme
did you use pump.io as an example for something being not unusual, when pump.io is the only one to use it?
bblfish joined the channel
#
melvster1
ben_thatmustbeme: was that a question for me? i tried to understand pump.io ... it was pretty confusing to me at least, but I think I got to the bottom of what I wanted to do ... in my implementation inbox is a discoverable place to send things ... then the person who owns the inbox gets an inbox stream ... what I had to get my head round was that the inbox has two roles line in email, one for the owner, one for the sender
#
melvster1
but it's working really well now ... im starting to roll out decentralized payments
bblfish joined the channel
#
melvster1
excellent: known supports webmention + JSON
#
melvster1
isnt that 99% of indieweb profiles?
benwerd joined the channel
#
melvster1
benwerd: do you have any stats for how many profile known supports, and how many are in the indieweb eco system?
#
benwerd
Known powers ~8000 websites, and by definition they're all in the indieweb ecosystem.
#
Loqi
Known has 7 karma
#
melvster1
benwerd: benwerd, awesome ... sorry, I should rephrase ... do you know roughly how many profiles are in the wider indieweb ecosystem (ie outside of known)
#
benwerd
oh. no, I have no idea, but it's significant, because of brid.gy and the wordpress plugins
#
aaronpk
there is almost no way to know by definition, since there are so many different implementations and such
#
aaronpk
all any one person can say is how many indieweb sites they have come in contact with
#
ben_thatmustbeme
decentralized++
#
Loqi
decentralized has 1 karma
#
aaronpk
there's a start of a project to try to crawl and gather stats on how many sites support which features http://indiewebcamp.com/indie-stats
#
melvster1
benwerd: but do you have any ball park numbers ... e.g. c. 10, c. 100, c. 1000, c. 10000 more?
#
aaronpk
it depends on how you count
#
aaronpk
e.g. how many websites can you subscribe to in a reader? lots and lots
#
ben_thatmustbeme
well, obviously more the 10, 100, 1000
#
ben_thatmustbeme
since known is 8000
#
benwerd
melvster1: it would be irresponsible of me to even guess. we don't consider our market as a startup to be "indieweb users" as such, but here's what I know -
#
benwerd
entire universities are jumping on board for their students, not using known
#
benwerd
(some are, of course, but some aren't)
#
melvster1
ben_thatmustbeme: question was *not including* known ...
#
benwerd
what that means is that entire student bodies will have access to indieweb tools
#
benwerd
largely using bridgy and those wordpress plugins
#
ben_thatmustbeme
benwerd, you also don't have stats when its a private install do you?
#
benwerd
ben_thatmustbeme: we have very loose stats, because known pings to see if there's a newer version available
#
ben_thatmustbeme
also worth noting that known is multi-user so there soulc be any number of users on those sites
#
aaronpk
like I said, a start of a project
#
ben_thatmustbeme
wow, my typing today
#
benwerd
ben_thatmustbeme - Right, we can count sites, not users
#
ben_thatmustbeme
s/soulc/could
#
melvster1
so would it be wrong to say that majority of user profiles that implement a full indie web stack, to the extent of being able to implement WG user stories, are part of known?
tilgovi joined the channel
#
melvster1
*that we know of*
#
aaronpk
melvster1: where are you going with this line of questioning?
#
melvster1
aaronpk: just trying to understand the install base of the technology ... it looks like about 8000 profiles on known, and about up to 80 home page of varying degrees of use ... i wanted to understand what else was out there
#
melvster1
if we know
#
ben_thatmustbeme
melvster1: 8000 sites, probably a much higher number of users/profiles in that
#
melvster1
got it
#
ben_thatmustbeme
i don't know if there are any stats on the wordpress plugins
#
aaronpk
some of the plugins expose rough numbers of downloads
bblfish joined the channel
#
aaronpk
also keep in mind there are varying degrees of supporting "indieweb technology", everything from having a feed on your home page to supporting web sign-in, to using micropub, etc
#
melvster1
aaronpk: im mainly interested in the context of this group of systems that can support the user stories ...
#
ben_thatmustbeme
melvster1: there are tons of user stories and they were all written in one-shot passes, they are not gospel at all
#
melvster1
ben_thatmustbeme: well that may be the case, but they are relevant to this working group, implementations of user stories count, i am told ... so i was trying to understand what is out there, and where things are currently
#
ben_thatmustbeme
melvster1: "implementations count" does not mean "implementations of user stories"
#
melvster1
aaronpk: yes I understand, for example I support indieweb and indieauth, but have not yet found a need for webmention ... or some of the other parts, because I use SoLiD for that type of use case
#
ben_thatmustbeme
implementing means you learned things about what works and what doesn't
#
melvster1
ben_thatmustbeme: yes exactly, this is how SoLiD developed over the years ... trying different things, finding what worked, what doesnt ... when you hit a brick wall ... so many of things I see going on today, at various times in the past we've tried, and found out more about ... some things are new and we'd like to try
#
melvster1
ben_thatmustbeme: for example about 9 years ago I started out with the "your homepage is your identity" philosophy which is really popular in indieweb ... however over time I learnt some problems with that approach and had to start from scratch, that's how webid was born
#
ben_thatmustbeme
what problems did you hit with that approach?
#
ben_thatmustbeme
homepage as identity that is
#
ben_thatmustbeme
and if its just range14, just pass
bblfish joined the channel
#
melvster1
ben_thatmustbeme: ive always been interested in making payments a first class citizen of the social web ... when sending money, security is a slightly higher concern, and so is making sure money goes to exactly the right place ... the other area that ive experiences is client side caching ... where you want to separate information about HTTP documents with information about things ... over the years *through implementation experience* you pick up many m
#
melvster1
any learning experiences, which are hard to articulate in a paragraph
#
melvster1
ben_thatmustbeme: it doesnt make too much sense to send a payment to a homepage
#
melvster1
ive always considered web 1.0 and web 2.0 a prelude to a web economy
#
ben_thatmustbeme
your personal preference to doing payments systems really doesn't effect the social systems though
#
melvster1
ben_thatmustbeme: as I said, 10 years of implementation experience of social is too much to articulate in a single paragraph, there's lots of tiny details ... i was giving an example because you asked, take it of leave it
#
melvster1
it was just meant to be informational
#
melvster1
the example I brought up before was likes, which are similar to payments ... you may like someone's homepage but not like them
#
melvster1
a page about a book may have a license, but the book may have a different license ... lots of little things come up over time
#
ben_thatmustbeme
sigh, http-range14
#
ben_thatmustbeme
lets just move on
#
melvster1
for me there were enough problems to merit starting over, we did start out in the same way as you guys
#
ben_thatmustbeme
and in practice, in actual social, people don't care to like pages, so its really not been an issue
#
melvster1
ben_thatmustbeme: pointless conversation, you have your world view, and no number of examples will shake you from it ... the only way is to do what we did 5-10 years ago and that's try to build stuff and see what brick walls you it
#
melvster1
s/it/hit
bblfish joined the channel
#
ben_thatmustbeme
points to indieweb. #nonissue
#
melvster1
sure if you dont consider interop an issue ...
#
aaronpk
"oh no, if I click the Like button, people are going to think I like Nickelback, when really I just liked what they had on their Facebook profile"... said no facebook user ever
#
melvster1
which is essentially the definition of a silo
#
ben_thatmustbeme
interop is within a spec. interop between wildly different systems is not
#
melvster1
ok, fair enough, but that's what i consider "siloism"
#
ben_thatmustbeme
as you have been told many many many times, we are not trying to make solid, pump.io, and indieweb work together is some conglomerate mess
#
ben_thatmustbeme
that won't work
#
ben_thatmustbeme
thats not the goal here
#
ben_thatmustbeme
we have a pile of implementations that all interop already
#
melvster1
yes but at a snails pace ... you indieweb implemented 1 user story from 90 in 5 years
#
melvster1
solid is 3 months old and has documented 4-5 and is working on 6-7
#
ben_thatmustbeme
how many software stacks do you have?
#
ben_thatmustbeme
also, those user stories were made rather outlandishly
#
melvster1
http, linked data, websockets, SSL
#
ben_thatmustbeme
sorry, how many implementations
#
ben_thatmustbeme
that interop
#
melvster1
we dont use software stacks, we just use the web
#
melvster1
https is part of the web
#
melvster1
you could argue whether or not websockets are ... but they probably are
#
melvster1
is it a software stack?
#
melvster1
is http a software stack?
#
ben_thatmustbeme
how many websites is solid working on?
#
melvster1
we dont have stats but there are 10 different implementations for the core tech ... with #11 being worked on ... solid is only 3 months old so it's really early to say ... but actually it's just another name for "the web" ... we're basically just using w3c RECs in a logical way of putting things together ... very very little new is invented
#
melvster1
there's over a billion profiles out there
#
ben_thatmustbeme
okay, then do you have examples of those user stories happening across tech?
#
melvster1
not yet using the full SoLiD stack ... but hopefully we can build things to incentivize that
#
ben_thatmustbeme
not any of them?
#
melvster1
ben_thatmustbeme: yep, look at the browsable social graph user story ... there's 3-6 implementations ... and millions of profiles in the social graph
#
ben_thatmustbeme
not an accepted user story
#
melvster1
nope
#
ben_thatmustbeme
so you have completed 0 user stories
#
ben_thatmustbeme
s/user stories/accepted user stories/
#
melvster1
because you're on a war path to try and put another tech down ... i dont want to be part of that conversation, sorry
#
melvster1
it's not a competition
#
ben_thatmustbeme
you are the one who keeps bringing up numbers
#
ben_thatmustbeme
i'm trying to say, use real numbers
#
melvster1
i wanted to understand how known relates in indieweb as they have already implemented JSON
#
melvster1
with webmention
#
melvster1
so that's why i asked
#
melvster1
if the majority of indieweb is already using JSON that's a great case for standardization
#
ben_thatmustbeme
no, i'm talking about number of completed user stories
#
ben_thatmustbeme
not the number of users
#
melvster1
5 user stories
#
ben_thatmustbeme
facebook beats us all on that
#
melvster1
going for 6-7 which is the half way mark on that page
#
ben_thatmustbeme
every user story requires at least 2 softwares
#
ben_thatmustbeme
interoperating between them
#
melvster1
as i said we have 11 implementations passing the test suite
#
melvster1
10 soryr
#
melvster1
11 being worked on
#
melvster1
for the core
#
ben_thatmustbeme
thats not the question. how many user stories can you point to that have been completed using more than one implementation?
#
ben_thatmustbeme
thats how you cound completion of user stories
#
ben_thatmustbeme
s/cound/count/
#
melvster1
all of them
#
melvster1
we have one implementation in GO, one in PHP and one in node
#
melvster1
definitely the GO and node work together on all ... we have a test suite and it passes all ... php is there for most of it
#
melvster1
someone is working on python
#
melvster1
but i really want ruby
#
melvster1
so we can have Solid gold with gems :)
#
melvster1
someone is working on an android implementation too
#
ben_thatmustbeme
you had just said 5 minutes ago, none "okay, then do you have examples of those user stories happening across tech?"
#
ben_thatmustbeme
"not yet using the full SoLiD stack ... but hopefully we can build things to incentivize that"
#
melvster1
that was the billion user profiles
#
melvster1
main adopter being facebook ... but we know they are kind of a silo ...
#
melvster1
at least they are a baby step ahead of google
#
melvster1
and mozilla
#
ben_thatmustbeme
so you do have user stories completed? show me a link
#
melvster1
ben_thatmustbeme: we have/will put all links in the github area ... you are welcome to verify or raise any issue if you find them ... we'll fix them, or add extra info
#
ben_thatmustbeme
those look to be a description of how it will work, and how to do it in one software
#
ben_thatmustbeme
one links to a pdf
#
melvster1
well thats because there were almost no implementations when the page stared, only proposals ...
#
melvster1
ive raised an issue saying that we should link to implementations
#
melvster1
yes
#
ben_thatmustbeme
just on that main page
#
melvster1
i didnt make that page
#
melvster1
i was just invited to fill it in
#
melvster1
id like to see it better orgaized
#
ben_thatmustbeme
go for it. beauty of github
#
melvster1
not a PR, in the issue tracker
#
ben_thatmustbeme
still no links
#
melvster1
the pdf links to a full implementation http://cimba.co/
#
ben_thatmustbeme
you said you had links
#
ben_thatmustbeme
yes, 1 software
#
melvster1
actually that probably implements about 5 user stories already
#
ben_thatmustbeme
again, cannot implement user stories in one implementation
#
melvster1
yes we have 3 implementations of every user story
#
ben_thatmustbeme
time to go catch a train
#
melvster1
cu :)
#
melvster1
in fact we have 3x3
#
melvster1
because 3 on the back end
#
melvster1
3 on the cllient side
#
melvster1
in general
#
melvster1
so that's 9
#
melvster1
PHP + GO + NODE
#
melvster1
then on the client side there's apps, and a generic data pane (tabulator) ... you can also use curl I suppose
#
rhiaro
melvster1: I tried to curl public posts on cimba but it wanted auth
#
melvster1
rhiaro: they were not public then ... maybe it was a cert thing ... you can add -k to allow self signed certs
#
rhiaro
I used the cimba UI to choose 'public'
#
rhiaro
That's what I was basing it on.
#
rhiaro
I can't access it in a browser without auth either
#
rhiaro
Aside: I can't curl a public channel because it's all rendered in JavaScript
#
melvster1
rhiaro; i dont know about the UI label, but if there's no ACLs it's public and accessible by curl ... if not, it's a bug, just raise an issue
#
rhiaro
Yeah I mentioned it on gigger
#
rhiaro
gitter
#
melvster1
gitter is a chat room, not an issue tracker ... we're starting to use waffle for issues
#
melvster1
but that aside
#
melvster1
the UI is *instead* of curl
#
melvster1
you send curl *not* to the UI, but rather, to the data store
#
rhiaro
I used the UI to create it, I mean
#
melvster1
all resources can be create with curl too
#
melvster1
just a bit more low level
#
melvster1
it's like git vs a git UI
#
melvster1
git command line i mean
#
melvster1
you can use warp to look at the underlying workspace
#
melvster1
that will also enable you to change files using a file system type UI
#
melvster1
you just need to know which channel your resource is pointing to
#
melvster1
cimba has channels
#
melvster1
with names
#
melvster1
fun, work, family etc.
#
melvster1
those label may be different from the ACLs
#
melvster1
just look at the data space and then see what the ACL says ... if you want something 100% public just remove all acls, but this is like chmod 777 in unix ... be careful
#
melvster1
initially everything was public as default ... but there's a drive to make them private by default now, for security reasons
#
melvster1
as timbl said, 'we learnt that about security the hard way'
#
rhiaro
Do you have a public post on cimba you can link me to?
#
melvster1
not sure i made any public posts, but each post will have a hyperlink
#
melvster1
just click it and it will open in warp
jasnell joined the channel
#
melvster1
andrei would probably haev some
#
melvster1
btw all the posts on webid.im have the same format as cimba
#
melvster1
i think andrei is working on verson 1.2 of cimba so he'll the most up to date thing
#
melvster1
<>
#
melvster1
<http://rdfs.org/sioc/ns#content> "tryin out cimba version 1.2" ;
#
rhiaro
Thanks melvster. By the way, I searched cimba for yours and andrei's profiles, but didn't see any posts
#
melvster1
andrei would probably be the quickest to be able to tell you why that is
the_frey joined the channel
#
melvster1
maybe a bug crept in
#
melvster1
also browser tend to be a pain and change their security settings every few weeks
#
melvster1
"for your own protection" of course
#
melvster1
chrome is really hard to get anything done, and ive almost given up hope already on firefox
#
melvster1
i just logged in and got my list of posts
#
melvster1
I can see 6 channels in my channels tab
#
melvster1
you can also use https://webizen.org/ to find webizens
#
melvster1
you're in there!
#
melvster1
wow 2 amy's with webids :)
#
rhiaro
What's your profile with posts?
jasnell joined the channel