ANN: Online Presence Ontology

[Milan Stankovic]

Hi all,

I would like to draw your attention to a new project that we have started at the Good Old AI (http://goodoldai.org.yu/) research group. The project addresses the issue of integrating and exchanging the data related to users' presence in the online world.

The core part of the project is the Online Presence Ontology (OPO) that can be used to represent instant messaging statuses, status messages, avatars and other elements that form the image of a user's presence in the online world. As a difference from FOAF that models more static and persistent user profile data, our goal is to model the dynamic and frequently changing aspects of user profiles.

All those dynamic aspects of online presence are currently published on different services (social networks, instant messaging platforms, Twitter-like and lifestreaming services) and the aim of OPO is to facilitate their exchange across those services.

For more information on our current and future work please visit the OPO website (http://www.milanstankovic.org/opo/) and take a look at the Nodalities blog post related to OPO (http://blogs.talis.com/nodalities/2008/08/opo-modelling-dynamic-online-presence.php).

All your comments, suggestions and ideas are welcome, as we are trying to further develop our ontology into a highly usable community-shaped vocabulary.

Cheers

Milan Stankovic

[Loïc DIAS DA SILVA]

It seems good stuff but what's the adhesion with XMPP ?

2008/8/20 Milan Stankovic <mil...@gmail.com>

--
--
Loïc DIAS DA SILVA
mglcel[at]mglcel[dot](fr/com/net)
The future is bright.

[Kevin Marks]

looks massively overcomplicated to me. Here's what we came up with for OpenSocial, modelled on XMPP:

http://code.google.com/apis/opensocial/docs/0.8/reference/#opensocial.Enum.Presence

[Christian Scholz (mrtopf.de)]

Hi!

I wanted to chime in here because when working on the Open Grid
Protocol (sort of the HTTP for virtual worlds) we will sooner or later
also have to work on the idea of presence. I need to think more about
this topic and how it's related to what's needed on the web but it
would be great of course if only one presence model would be used.

One thing which directly comes to mind though is one difference in
virtual worlds: You also want the location of somebody (as long as
that user agrees to share it with you). This might even be interesting
for RL situations to tell people where on the world you are right now.
The other difference is that with OpenSocial we probably only talk
about one server/service on which you are online. In the Open Grid
Protocol it shouldn't matter where you are, on which Agent Domain
(this is the part which stores your account and profile and groups and
friends list) or on which region (the actual 3D environment in case
you are connected to one). So this deeply dives into what data
portability is about.

I am not sure what else we would need but I also mailed to the Open
Grid list and people interested might also join the conversation
there: https://lists.secondlife.com/cgi-bin/mailman/listinfo/gridnauts

cheers,

Christian

On Aug 21, 8:24 am, "Kevin Marks" <kevinma...@gmail.com> wrote:
> looks massively overcomplicated to me. Here's what we came up with for
> OpenSocial, modelled on XMPP:
>

> http://code.google.com/apis/opensocial/docs/0.8/reference/#opensocial...

>
> On Wed, Aug 20, 2008 at 11:21 AM, Loïc DIAS DA SILVA <mgl...@gmail.com>wrote:
>
> > It seems good stuff but what's the adhesion with XMPP ?
>

> > 2008/8/20 Milan Stankovic <mils...@gmail.com>

>
> >> Hi all,
>
> >> I would like to draw your attention to a new project that we have started
> >> at the Good Old AI (http://goodoldai.org.yu/) research group. The project
> >> addresses the issue of integrating and exchanging the data related to users'
> >> presence in the online world.
>
> >> The core part of the project is the Online Presence Ontology (OPO) that
> >> can be used to represent instant messaging statuses, status messages,
> >> avatars and other elements that form the image of a user's presence in the
> >> online world. As a difference from FOAF that models more static and
> >> persistent user profile data, our goal is to model the dynamic and
> >> frequently changing aspects of user profiles.
>
> >> All those dynamic aspects of online presence are currently published on
> >> different services (social networks, instant messaging platforms,
> >> Twitter-like and lifestreaming services) and the aim of OPO is to facilitate
> >> their exchange across those services.
>
> >> For more information on our current and future work please visit the OPO
> >> website (http://www.milanstankovic.org/opo/) and take a look at the
> >> Nodalities blog post related to OPO (

> >>http://blogs.talis.com/nodalities/2008/08/opo-modelling-dynamic-onlin...
> >> ).

[Kevin Marks]

On Thu, Aug 21, 2008 at 2:29 AM, Christian Scholz (mrtopf.de) <tao.t...@googlemail.com> wrote:

Hi!

I wanted to chime in here because when working on the Open Grid
Protocol (sort of the HTTP for virtual worlds) we will sooner or later
also have to work on the idea of presence. I need to think more about
this topic and how it's related to what's needed on the web but it
would be great of course if only one presence model would be used.

One thing which directly comes to mind though is one difference in
virtual worlds: You also want the location of somebody (as long as
that user agrees to share it with you).

we have a current location field:

http://code.google.com/apis/opensocial/docs/0.8/reference/#opensocial.Person.Field.CURRENT_LOCATION

specified as an Address:
http://code.google.com/apis/opensocial/docs/0.8/reference/#opensocial.Address.Field_field_detail
 

This might even be interesting
for RL situations to tell people where on the world you are right now.
The other difference is that with OpenSocial we probably only talk
about one server/service on which you are online.

Wiht the 0.7 release there is usually one container that provides the social data for the running applications, but with the REST APIs in 0.8 you can federate this server to server.

 

[Christian Scholz / Tao Takashi (SL)]

On Thu, Aug 21, 2008 at 12:09 PM, Kevin Marks <kevin...@gmail.com> wrote:
>
>
> On Thu, Aug 21, 2008 at 2:29 AM, Christian Scholz (mrtopf.de)
> <tao.t...@googlemail.com> wrote:
>>
>> Hi!
>>
>> I wanted to chime in here because when working on the Open Grid
>> Protocol (sort of the HTTP for virtual worlds) we will sooner or later
>> also have to work on the idea of presence. I need to think more about
>> this topic and how it's related to what's needed on the web but it
>> would be great of course if only one presence model would be used.
>>
>> One thing which directly comes to mind though is one difference in
>> virtual worlds: You also want the location of somebody (as long as
>> that user agrees to share it with you).
>
> we have a current location field:
>
> http://code.google.com/apis/opensocial/docs/0.8/reference/#opensocial.Person.Field.CURRENT_LOCATION
>
> specified as an Address:
> http://code.google.com/apis/opensocial/docs/0.8/reference/#opensocial.Address.Field_field_detail

Ok, that's good for RL locations but not so much for one in a virtual
world. And you might want to have multiple locations then.

>
>>
>> This might even be interesting
>> for RL situations to tell people where on the world you are right now.
>> The other difference is that with OpenSocial we probably only talk
>> about one server/service on which you are online.
>
> Wiht the 0.7 release there is usually one container that provides the social
> data for the running applications, but with the REST APIs in 0.8 you can
> federate this server to server.

I wanted to look into that anyway so I hope to have time for this soon
to check how this might fit into the Open Grid Protocol. Are there any
controls for specifying what can go from server to server? Like
permissions attached to data etc.? (which is also something we are
discussing already for OGP, as it's a very important topic there and
relates very much to which services you trust and which you don't).


Thanks for the info!

-- Christian

--
Christian Scholz
Tao Takashi (Second Life name)
taota...@gmail.com
Blog/Podcast: http://mrtopf.de/blog
Planet: http://worldofsl.com

Company: http://comlounge.net
Tech Video Blog: http://comlounge.tv
IRC: MrTopf/Tao_T

[Milan Stankovic]

Hi Loïc, Hi Kevin,

Thanks for your interest in OPO.

XMPP is more focused on delivering the inter-platform chat functionality than on semantically representing IM statuses. The exchange of IM statuses there is done by mapping them all to a simple scale of 6 statuses, and there are IM platforms that use many more. The semantics of those statuses that differ from predefined XMPP statuses is often lost when mapping them to XMPP status scale. (We can take the example of skype statuses Online and SkypeMe, and suppose we wanted to map them to that scale. In that hypothetical case we would have to map them both to the XMPP status CHAT, and there is a way to represent their different meanings in OPO)

We think of OPO's expressivity as of an useful property that might justify its complexity.

On the other hand, we find XMPP to be very useful, and having in mind its wide adoption, we are thinking to suggest its extension to support the exchange of OPO data.

I hope this answers your questions. Please feel free to ask for further clarifications if necessary.

Best regards,

Milan