Below is our submissision to the W3C Workshop on the Future of Social Networking in Barcelona in January 2009.   A list of all the papers submitted to the conference can be here:  http://www.w3.org/2008/09/msnws/papers/.

The Relationship Layer and the Secretary

Copyright and Property of Minggl 2008 – www.minggl.com

Abstract

Multi-context relationship and social-interaction data should be shared and architecturally centric
to the social web.  This will allow many wonderful services which, acting as your virtual
secretary, can execute all manner of automated decision making and personal boundary
enforcement.

By:  Dewey Gaedcke, CEO, Minggl

Contributions by:
Marcus Irven, VP of Development, Minggl
Juan F. Sequeda, Founder – Semantic Web Austin, University of Texas at Austin

Introduction

A high-end Executive Assistant (my secretary) will know of each lunch, phone call, business
deal and tennis match that you and I have ever had together.  From that knowledge, she has a
sense of our relationship strength, and can intuitively and accurately prioritizes information,
requests and visibility between us.   If she’s really good, she’ll do it in a way that aligns with my
intentions and agenda regarding you, without ever bothering to ask me about it.    Our position is
that effective social automation will require technology that can prioritize and sort relationship
context in a manner similar to off-line human behavior.  Support for these capabilities will
require cross-domain data visibility and (at minimum):

1. A global identity for each person
2. Translation service between global and domain specific user identifiers (“domain”
   signifies user identity at each social site)
3. Historical knowledge of my relational activity across domains (i.e. all social services)
4. Data exchange format and retrieval API for dispersed individual “interaction history”
   (called “actionstory” from this point forward)

A Real World Example

In the absence of a central “actionstory” repository, each social service will fetch recent social
contact data from each of my other services.  They will then use this data to calculate several
dimensions of my “social tie strength” to everyone in my localized friend list.   From there, many
interesting conveniences and automations will become possible.  Before we discuss the future
potential, let’s look at problems with the current model.
For example, what happens when…

1. My accountant sends an emergency reminder about my late IRS tax-return deadline (due
   today), but he uses a 3rd tier email address….an inbox that I check only once every 10
   days?
2. A close friend posts important personal news on a social site that I rarely visit?
3. I post pictures of my children and the wrong people have access to them?

The answer:  not what I intended….I’m missing key information, getting it late, or experiencing
privacy consequences.  So what do we mean by “close friend”?

In the real world, our brains automatically (without conscious effort) prioritize relational events
and content based on context, agenda and social-proximity (how well we know someone).  You
can witness your brain doing this every time you walk into a public place and connect naturally
with people in varied ways.  Your brain’s “prioritization process” may dictate that you hug an
ex-girlfriend, shake hands with an acquaintance and only introduce yourselves to strangers under
certain circumstances.

The three anti-examples above illustrate how familiar technical abstractions, such as the
communication channel (i.e wrong email address), the social venue (i.e. land-locked information
in different websites) and specific features (i.e. inflexible privacy settings) have gotten
fundamentally in the way of reasonable human intention.  The brains’ natural prioritization
process has been thwarted because the sender and the technology did not consider our full
relationship and intentions….e.g. the site did not know that we were “close friends”

Technology Centric vs. Relationship Centric Social Infrastructure

Technical abstractions such as features, tools, and social venues, are currently how the social
web is organized, and have therefore (unfortunately) become it’s de-facto center.  Just imagine if
the supermarket (social venue) and the cell-phone (communication channel) dictated who, how
and where you could socialize with other people.  It’s just not how we humans operate….we (or
our secretaries) carry stateful context across all domains and apply this former meaning to each
new relationship situation.

“Social” is defined by the relationships and context between people. The personal meaning that
flows from our collective history is among our most valuable assets.  Our brains use it to
automate all manner of things in the real world, but this asset is almost completely missing on
today’s Social Web.

Facebook, et al, mistakenly acted as if the social venue (the website) was equivalent to the
community.  A community is a collection of people who share history, context, common values
and group norms.  A social site is only a location, no more a community than your local
nightclub or supermarket.  People, relationships and context create “community” and the social
web of today has this reality completely upside down.  We argue that the relationship (social
interaction) data should be architecturally centric to the social web.  This will allow the
technology (your virtual secretary) to support automated decision making, which could include
things like:

1. Prioritizing or demoting a message depending upon who sent it
2. Rerouting a message to other recipients or through a faster/optimal delivery track
3. Automating Privacy (context appropriate disclosure)
4. Delivering birthday flowers with minimal intervention

Relationships as the Central Driver of the Social Web

For these automation capabilities to become a functional reality, the Social Web needs an“Interaction History” (“actionstory”) data format and an API centered on a universal human identifier (UHI).  This data format might well be an appropriate extension to FOAF.  The various social services will (optionally) extend an option to log all “actionstory” activity between me andthose with whom I interact.  Only by storing robust knowledge of my entire social “actionstory”, can the cloud begin to function as my very effective executive assistant.  Participating social services will track (and share) details of every interaction or social encounter including:

1. Interaction type
2. Frequency
3. Duration (or persistent)
4. Who initiated
5. Whether responded/attended/punted/referred
6. Who else included
7. Subject matter/context (standard taxonomy based if possible)
8. Relative connectedness to other touch points with the same person (i.e. we work together)

This data will then be used in all manner of automated-assistant services.  As another example, if
LinkedIn knows that we work together, Facebook will be able to (automatically) use this meta-
data about our relationship, and keep certain data private, while giving you access to work
related content and personas.

Conclusion

Charlene Li of Forrester predicted that the future of social networks will be “like air”.  We
believe she means functional and ubiquitous, yet widely unnoticed.  Just as the supermarket and
the cell-phone do not control socialization options in the real world, online services that base
user control upon “actionstory” data, will move the technology out of the way and allow our
socializing to really be, as thin as air.