Re: streaming/push "out of scope", was Re: on API Requirements

I think we cmay need a 'Works With' section (I'm sure there's a better 
name) for things that we need to ensure we are not blocking, but whose 
definition are not core to the project.

A defined list would make it easier to refer to during discussion and 
allow us to put stakes in the ground as to the degree to which we are 
willing to include support.

The Authentication system mentioned would be another example as in a 
federated system especially, identification, authorization and 
authentication may need clear support statements.
_________________________________________
Bill Looby
Software Architect, Dublin Software Lab, IBM Ireland
_________________________________________




From:   ☮ elf Pavlik ☮ <perpetual-tripper@wwelves.org>
To:     Sandro Hawke <sandro@w3.org>, Evan Prodromou <evan@e14n.com>
Cc:     public-socialweb@w3.org
Date:   23/02/2015 08:54
Subject:        Re: streaming/push "out of scope", was Re: on API 
Requirements



Sandro, what do you think about keeping it on a back burner for now?

At the same time whenever we see that some choices in architecture may
make streaming/push harder to add later, we will clearly raise such 
concern.

On 01/31/2015 03:47 AM, Sandro Hawke wrote:
> On 01/29/2015 04:30 PM, Evan Prodromou wrote:
>> Everyone is blown away by the size of these requirements already.
>>
>> A streaming protocol for streams would be a great addition later, but
>> trying to jam it in here will literally sink this project.
>>
>> Please accept this as being out of scope.
> 
> I might be misunderstanding what you mean by "out of scope".  To me, in
> a WG, "out of scope" means "we're not even going to talk about this
> issue, because it's not the kind of problem our charter says we're
> supposed to talk about".  In general, it's up to the chairs in guiding
> to conversation to steer it away from things that are out of scope (in
> this sense) given their reading of the charter.
> 
> What I think you're saying above is that you don't think streams/push
> should be one of the requirements for the API.   That's plausible, but
> maybe we can label that as "Not a requirement", instead of "out of
> scope"?    That is, it's up to the group to come to consensus on what
> the requirements for the API are, and you're arguing this should not be
> one.  I'm sympathetic to your argument, but I'd also be interested in
> hearing whether likely vendors of this stuff think they can sell systems
> without streaming/push.
> 
> The alternative interpretation is that maybe you think our charter
> doesn't allow us to even consider this as a possible requirement, that
> it's out-of-scope for the group.   Like, an authentication system would
> surely be out-of-scope, and Harry was just arguing that WebFinger is out
> of scope.
> 
>         -- Sandro
> 
>>
>> Evan Prodromou
>>
>>> On Jan 29, 2015, at 15:00, ☮ elf Pavlik ☮
>>> <perpetual-tripper@wwelves.org> wrote:
>>>
>>>> On 01/29/2015 02:21 AM, Sandro Hawke wrote:
>>>>> On 01/28/2015 06:13 PM, Evan Prodromou wrote:
>>>>>
>>>>>> - Can you justify most of the out-of-scope stuff?   Without 
streaming
>>>>>> or push, I don't see how this system could catch on.
>>>>> I think that server-to-server stuff is going to be more pertinent 
when
>>>>> we discuss the federation protocol.
>>>> Clients need streaming/push, too, don't they?
>>> +1
>>>
>>> in my experiments a while ago i used HTTP + JSON based pub/sub 
protocol
>>> Bayeux: http://svn.cometd.org/trunk/bayeux/bayeux.html

>>>
>>> using one of its implementations: http://faye.jcoglan.com/

>>>
>>> mentioned decentralized prototype with real time geolocation map 
layers:
>>> https://github.com/dspace-ng/dspace-app-action-slim

>>>
>>>
> 
>