Re: An Introduction: SOCML

On 13 February 2013 02:46, Christopher A <chris.socml@gmail.com> wrote:

> Melvin, Pelle, Julian,
>
> It is nice to meet all of you. I was worried that I had stumbled into a
> dead forum -- good to see some life in here! I also want to thank you all
> for the feedback and encouragement!
>
> I have to say, I find the concept of Activity Streams to be nice and close
> to what I would like to accomplish; however, it seems to lack an emphasis
> on portability of user content, and largely seems to be "action related." I
> worry that this format, based on the backgrounds of its authors, will be
> mainly used as a method for the "big social networks" to aggregate more
> user activities outside of their service rather than fostering the growth
> of competitive alternatives.
>

Could not agree more.  AS is use case driven rather than a generic solution
where you can plug in your own data models, afaik.   However, as you state
it can be the right tool for the right job in some places.  I am concerned
about portability of AS, and am unsure how much effort has been put into
that aspect, to date.  Though it does have it's own mime type which can
help the client understand what it's getting.


>
> While, I find JSON to be meritorious in its usefulness and conciseness; I
> still believe that XML is the most widely understood and supported data
> interchange standard. Most shortcomings with XML can also be solved with
> simple compression algorithms.
>

Sort of yes.  The serialization isnt always that important, the principles
behind the serialization are though.  Things like portability, namespacing,
self describing data, non collision, extensibility can be beneficial when
trying to build systems that scale.  XML in the form of RDF/XML for example
kind of reached a local minimum in terms of adoption, and I get the
impression people like to work with JSON these days.  JSON-LD is a
promising serialization, imho ( http://json-ld.org/ ) and easily
translatable to XML forwards and backwards.

A good video on this from Douglas Crockford :
http://inkdroid.org/journal/2012/04/30/lessons-of-json/


>
> With respect to Melvin's comment about putting too much of an early
> emphasis on security and authorization; I think he's right. It's probably
> best to keep the standard focused as a data standard and leave the
> implementation of security and authentication up to the user's service.
>

Thanks :)  I do think security is very important, but also that the modular
approach is the one that scales, rather than the, 'one size fits all', that
tends to be popular.  This allows one team to work on communication and one
to work on security and the ability to fit things together.


>
> Currently, some of the main design issues I'm struggling with at the
> moment are:
>
>    1. Should the standard include "action related" structures that
>    incorporate a "push" and "retrieve" methodology for facilitating interop
>    between services?
>
> Would like to understand this better ... do you have an example?


>
>    1. How can we standardized social media objects such as messages,
>    status updates, pictures, etc. while maintaining the ability to add
>    additional extensibility to these objects?
>
> There's a few efforts in progress.  AS already mentioned.  Some work was
done here http://xmlns.notu.be/aair/ ... tent.io have a neat system,
opensocial perhaps, and facebook APIs and open graph protocol are certainly
are worth a look.


>
>    1. With respect to service interop: is there a way to uniquely
>    identify users while allowing each service to assign its own IDs to these
>    entities?
>
>
Very happy that you mention this!  imho, it's generally the most overlooked
problem in designing scalable protocols. If you have a common ID system
across heterogeneous systems scaling and interop are relatively easy, case
in point, http URLs have demonstrated massive scalability and easy
interop.  If you are unsure of your ID system or create your own (which I
think is the strong temptation) you struggle to scale to anything more than
yourself.


>
>    1. How can we create a data structure that prevents impersonation and
>    spoofing?
>
> You can never 100% prevent this, but there are ways to mitigate such
things, such as PKI.


> Most of you have probably noticed that I have been slow with updating the
> Wiki. The main reason for this is because of the aforementioned issues.
>
> Any insight and help would be greatly appreciated. In the meantime, I
> think I'll work on some of the low hanging fruit and knock out the XML
> structures for some of the more simplistic social media objects.
>

Thanks for sharing some great ideas! :)


>
> Chris
>
>
>
>
>
>
>
> On Mon, Feb 11, 2013 at 4:50 PM, Melvin Carvalho <melvincarvalho@gmail.com
> > wrote:
>
>>
>>
>> On 11 February 2013 23:38, Pelle Wessman <pelle@kodfabrik.se> wrote:
>>
>>> 11 feb 2013 kl. 23:03 skrev Melvin Carvalho <melvincarvalho@gmail.com>:
>>>
>>>
>>>
>>> On 11 February 2013 19:27, Julian Steinwachs <
>>> julian.steinwachs@googlemail.com> wrote:
>>>
>>>>  Hi,
>>>>
>>>> this protocol agnostic approach sounds very interesting. Because you
>>>> said you did not search the web thoroughly: There is a standard called
>>>> activitystreams ( http://http://activitystrea.ms/ ) that is designed
>>>> to describe social interactions (activities) in atom/xml or json. Beyond
>>>> the discovery and requesting stuff socml could be designed as an encryption
>>>> wrapper for activitystreams. Or to make it very short
>>>> socml=activitystreams+PGP ?
>>>>
>>>
>>> I like AS, but my impression was that original poster asked for
>>> something that was patent free.  If AS one day put under the IETF or W3C
>>> that would work, but I dont see this happening in the short to medium term.
>>>
>>>
>>> AS is released under the Open Web Foundation which should make it as
>>> patent free as anything put under IETF or W3C? See
>>> http://activitystrea.ms/licensing/ and http://www.openwebfoundation.org/
>>>
>>
>> I know AS is very popular with some folks on this list so I hope not to
>> offend.  It's a system I like, but I like to take a more holistic approach
>> and look at the bigger picture, which includes linked data, open graph
>> protocol, tent.io and some other promising projects.
>>
>> IANAL, but it seems to me a grey area.  I have seen lawyers involved in
>> the AS mailing list, I have not to date seen that happen in social oriented
>> specs elsewhere.
>>
>> Certainly there has been patent litigation between OWF members in recent
>> years, Yahoo vs Facebook springs to mind ... in any case, I'd advise
>> looking at all the options out there.  The legal aspect is just one of many.
>>
>>
>>>
>>> We have existing standards that are mature, well adopted and approved by
>>> standards bodies, namely the linked data family of standards.  I would
>>> suggest this is a better match to the requirements listed.
>>>
>>>
>>>>
>>>> Greetings
>>>> Julian
>>>>
>>>> Am 11.02.2013 02:03, schrieb Christopher A:
>>>>
>>>> Hello all,
>>>>
>>>>  First, I want to apologize if I have overstepped boundaries with
>>>> respect to editing the groups Wiki page. I also want to apologize if you
>>>> receive this message twice. For whatever reason, the Listserv was not
>>>> showing that it as having been sent.
>>>>
>>>>  I recently posted a proposal for a "Federated Social Network Data
>>>> Standard" on the groups Wiki. I admit, that I have not searched the
>>>> web thoroughly with respect to other initiatives like this; however, given
>>>> the superficial research I have done, I have come to the conclusion that
>>>> there are no open dialogs currently on this topic.
>>>>
>>>>  Over the next couple of days I will begin posting proposed
>>>> technical specifications for the standard. I would like for everyone to
>>>> contribute feedback and make suggestions/modifications.
>>>>
>>>>  The solution I am proposing is simple: we need to standardize social
>>>> media content such that independent developers can create their own
>>>> services that can share and aggregate data under a common standard. Much
>>>> like the RSS format, this data standard should be open and free, not
>>>> encumbered by patents, and be easy to implement while offering these
>>>> features:
>>>>
>>>>    - Complete end-to-end privacy control, with the use
>>>>    of asymmetric encryption.
>>>>    - A method for authenticating that people "are who they say they
>>>>    are."
>>>>    - True protocol agnosticism, this data standard should be
>>>>    freely interchangeable with the number of web technologies that exist.
>>>>    - Content portability, meaning the user can choose to use a number
>>>>    of providers to store pictures, comments, events, messages, etc. while
>>>>    still having their information available to those with the proper access.
>>>>
>>>>  Anyway, here is the current link to the proposal:
>>>> http://www.w3.org/2005/Incubator/federatedsocialweb/wiki/SOCML_Proposal
>>>>
>>>>  The SOCML standard overiew:
>>>> http://www.w3.org/2005/Incubator/federatedsocialweb/wiki/SOCML_Standard
>>>>
>>>>  And the Technical Specifications (in progress!):
>>>>
>>>> http://www.w3.org/2005/Incubator/federatedsocialweb/wiki/SOCML_Technical
>>>>
>>>>  Please feel free to critique or reprimand.
>>>>
>>>>  I hope everyone is doing well, and I look forward to working with
>>>> everyone.
>>>>
>>>>  Chris
>>>>
>>>>
>>>> On Sat, Feb 9, 2013 at 5:08 PM, Christopher A <chris.socml@gmail.com>wrote:
>>>>
>>>>> Hello all,
>>>>>
>>>>>  First, I want to apologize if I have overstepped boundaries with
>>>>> respect to editing the groups Wiki page.
>>>>>
>>>>>  I recently posted a proposal for a "Federated Social Network Data
>>>>> Standard" on the groups Wiki. I admit, that I have not searched the
>>>>> web thoroughly with respect to other initiatives like this; however, given
>>>>> the superficial research I have done, I have come to the conclusion that
>>>>> currently there are no open dialogs around this topic.
>>>>>
>>>>>  Over the next couple of days I will begin posting proposed
>>>>> technical specifications for the standard. I would like for everyone to
>>>>> contribute feedback and make suggestions/modifications.
>>>>>
>>>>>  The solution I am proposing is simple: we need to standardize social
>>>>> media content such that independent developers can create their own
>>>>> services that can share and aggregate data under a common standard. This
>>>>> standard, should be open and free, not encumbered by patents, and be easy
>>>>> to implement while offering these features:
>>>>>
>>>>>
>>>>>    -
>>>>>     - Complete end-to-end privacy control, with use
>>>>>    of asymmetric encryption.
>>>>>    - A method for authenticating that people are who they say they
>>>>>    are.
>>>>>    - True data agnosticism, meaning the end user can choose to use a
>>>>>    number of providers to store pictures, comments, events, messages, etc.
>>>>>    while still having their information available to those with the proper
>>>>>    access.
>>>>>    - The ability to freely host or move their information to
>>>>>    different service providers.
>>>>>
>>>>>  Anyway, here is the current link to the proposal:
>>>>> http://www.w3.org/2005/Incubator/federatedsocialweb/wiki/SOCML_Proposal
>>>>>
>>>>>  The SOCML standard overiew:
>>>>> http://www.w3.org/2005/Incubator/federatedsocialweb/wiki/SOCML_Standard
>>>>>
>>>>>  And the Technical Specifications (in progress!):
>>>>>
>>>>> http://www.w3.org/2005/Incubator/federatedsocialweb/wiki/SOCML_Technical
>>>>>
>>>>>  Please feel free to critique or reprimand.
>>>>>
>>>>>  I hope everyone is doing well, and I look forward to working with
>>>>> everyone.
>>>>>
>>>>>  Chris
>>>>>
>>>>>
>>>>
>>>>
>>>
>>
>