Re: Why not AS2? was Re: Getting the group back on track

On 19 October 2015 at 05:27, Randall Leeds <randall@bleeds.info> wrote:

>
>
> On Sun, Oct 18, 2015 at 3:11 PM Melvin Carvalho <melvincarvalho@gmail.com>
> wrote:
>
>> On 18 October 2015 at 22:05, Randall Leeds <randall@bleeds.info> wrote:
>>
>>> I would hope that maybe the vocabulary can just be aggressively culled.
>>>
>>
>> Im curious as to the advantage of culling the vocabulary.  Might it make
>> more sense to say, just use the terms that your system needs, and ignore
>> the others?
>>
>>
>
> Probably doesn't bother me. Some things always turn out to be more useful
> than others, but falling short of specifying enough for useful interop
> would be a worse result. Thanks.
>
>
>> I believe that it's a bad idea to try to specify all the kinds of
>>> activities and relationships that people engage in and have.
>>>
>>> But, I thought that at least the framing, the most abstract vocabulary
>>> pieces, might find broad agreement.
>>>
>>> I'm mostly just confused, maybe because I haven't been following closely
>>> enough, about what, if anything, the group has consensus about wanting to
>>> do (regardless of the specifics of implementation).
>>>
>> Not really sure the group has strong consensus on very much, at this
>> point.
>>
>> There doesnt seem to be consensus on how to do profiles, relationships or
>> reuse of web standards.  IMHO this is a fundamental problem, but it's been
>> challenging to even discuss these items, let alone find common ground.
>>
>> The items that seems closest to consensus IMHO
>>
>> 1 Reuse HTTP to form a cross origin decentralized social web
>> 2 Create a JSON based syntax for exchanging messages
>> 3 Use Activity Streams 2 as a candidate for that syntax
>>
>>
> Thanks, that's really helpful.
>
> I hope my e-mail wasn't too rude.
>
> I see the idea that something that doesn't rely on AS2, or any other
> activity vocabulary, could provide value. I think that's (1) in your
> numbering. I would be happy to see any that specifies discovery,
> subscription, or notification.
>

AS2 has an extensibility mechanism that enables you to use any other
vocabulary as part of the mix.  That's a key feature.

In a general sense, the web is a discovery system.  In that you supply an
HTTP URI and you are going to get back documents and/or data.  In the last
15 years, the W3C has done quite a bit of innovation in this front to
create linked data, a mechanism to return structured data from an HTTP URI
in the form of JSON, XML, HTML or turtle.  For the purposes of getting
things done, JSON was selected for this group, however there are equivalent
forms with a one to one mapping in other formats.  What this means is that
a distributed social graph can be formed by adding data fields to various
entities and following links.  One of the implementations of this group
illustrates an existing social graph several million profiles deep using
this technology.

Notifications I agree are a key part of the modern user experience.  In
this group we've done work on push notifications via websockets (eg in
Solid) and pull notifications eg using AS2.  As you say, subscriptions tie
into this.

One final aspect that we found that was needed over the last 10 years
(predating this group) is that privacy and access control appear to be an
important feature for users, so that they can share the right content with
the right people.  That's one of the foundational aspects of Solid that was
inspired by UNIX to make the web a bit more like an operating system in
that sense.


>
> I see this list here and think it's great:
> http://www.w3.org/wiki/Socialwg/Social_API/Requirements#Requirements_shortlist
>
> And now I think I understand the conversation.
>
> I also think AS2 probably has value. I support this idea of loose coupling
> of these layers, and regret my exaggerated statements about this falling
> short of useful.
>

Thanks for the input! :)