Fabien Benetou's PIM | Person / Person

Principle

Many tools from research in social networking have emerged recently we must benefit from them, especially since our networks have grown a lot and if we want to be scalable and have quality relationships we won't be able to go further without coping tools (again).

Applications

improve caller-ID
- display contextual information about the caller
- provide a VoIP version
use the "Contact" section of wiki pages to automatically include the potential persons interested
- bottom and up limit
- privatization system
  - from the more specialized (perfectly matching tag) to the more general (anybody)
go pass beyond the Dunbar's number while maintaining high quality relationships
- The Dunbar Number as a Limit to Group Sizes by Christopher Allen, Life With Alacrity 2004
  - see also its related posts
- his Chapter 2 of The Prehistory Of Language
- his research group EPBERG @ Liverpool Evolutionary Psychology and Behavioural Ecology Research Group
- Robin Dunbar: How Many Friends Does One Person Need? RSA London February 2010
- DIYh+ thread initiated by Bryan Bishop
leveraging social models (e.g. Six degrees of separation) and mediated digitalized conversations (SNS, Google Wave, VoIP, ...)
- explore the notion of "social routing"
  - Jon Kleinberg's Homepage at Cornell University and his research on the algorithmic aspects of the "small world" experiments
applying and visualization models of sociology
- through own personal data
  - as content and as meta-data e.g. dates and frequency of updates based on categories
- e.g. Wikipedia:Computational sociology or more specifically Wikipedia:Actor-Network Theory or Wikipedia:Social epistemology
- inspired by discussion with Martin in Madrid

Note that being social has a cost (energy, resources, ...) and that maintaining ones network might also be seen with a (non financial) return on investment mindset in order to avoid extremes in both direction (asocial or ultra-social/eusocial).

Potential solution

Person.Template used to add another person
generated network
- PageLists to list persons
  - (:pagelist group=Person list=normal name=-Person,-*Template*:)
- PageListTemplates to make GraphViz code
  - Person.VizTemplate
- PmGraphViz to generate the image and display it

Note that the wiki format provide natively support for history of each page, potentially way to study the "history of a relationship"

Result

(:pmgraphviz -- [=

   digraph {
      me [label="me" style=filled fillcolor=blue];
      Person.Test1 [label="Person.Test1" style=filled fillcolor=orange];
      me -> Person.Test1;
   }

=] :)

See actually the local implementation (more secure since we are dealing with actual private information). Check Processing for improved visualization.

Remarks

Could family fatigue, meeting family members not for pleasure but because they are related, be a consequence of predictability? Could the low information exchange lead to gradually unpleasant encounters?
- note that this also happens with long-term close acquaintances but to a lesser degree
- risks
  - increasing predictability
  - social pressure
  - conflict avoidance
- advantages
  - no small talk required
  - known domain of agreements
- solutions
  - find personal topics in which the other person has specific knowledge that is interesting to you
    - unique trips for geography, family heritage and culture, youth and history, ...
  - introduce new topics and activities
  - leverage own skills the other person believe are unique to himself or herself for your own problem
    - offer to do the opposite but never force it, let the person take your offer later on if preferred
- see also
Teamwork as offloaded cognition for the initiator of the project
- is it not just delegating by laziness or for efficiency but a solution to solve complexity through scalability
- also, without delegating, could simply the presence within a specialized environment, for example a research laboratory, very efficiently facilitate the conduct of complex work by asking questions to other members and letting them play with those questions in the background of their mind lower the stress on each participant?
- inspired by N is a Number on Paul Erdos, The Mechanical Mind in History and Les Prodigieux Theoremes De Monsieur Nash in which each brilliant work was closer to vision and orchestration of the work to reach it rather than solely tedious attempts
  - also Chapter 5 The lone inventor of The Myths Of Innovation
- see also http://127.0.0.1/wiki/Collaborations/Collaborations
topology of the current social group, dynamics and their consequences
- 1 to 1 conversation
  - exchange, dialog, absence of the audience, no possibility to hide, ...
- many to many conversation
  - about "winning", having an audience, demonstrating alpha-type qualities, ...

Done

SocialAnnotating
- GreaseMonkey script that links back to Person:X for every website and profiles about X
- so far, only done with 1 person
filled http://127.0.0.1/wiki/index.php?n=Person.Person and gradually improving it
added News from social networks with Cookbook:PmFeed
- Facebook
- LinkedIn
lousily defined Format
added to localmap.txt and InterMap
- Person: http://127.0.0.1/wiki/index.php?n=Person.Person#
- email: https://mail.google.com/mail/?shva=1#

To do

use Seedea:Oimp/Socialannotating
define default network (email, IM, IRC, Twitter, ...) to transmit a message
- per user
- per type of message (texte, picture, ...)
- per content (personnal, professional, ...)
- as a composition of the 3
consider adding a password (or .htaccess limited to 127.0.0.1)
find an API for importing data from social networks
- embed/update data on pages
stats from irssi logs (IRC, IM through bitlbee/minbif, Twitter through twirssi, VoIP through Asterisk), Facebook, Twitter, ...
- people you are connected to
- frequency of msg exchanged
- date of last message sent
- qdos and its applications
  - schema that shows what counts are made for the various supported sites
  - rdf
    - turtle format
    - ext format
output a visual through PmGraphViz
- check The Friend of a Friend (FOAF) project "creating a Web of machine-readable pages describing people, the links between them and the things they create and do" and its potential usage
  - my own FOAF thanks to identi.ca
    - http://identi.ca/utopiah/foaf
  - see also the active wiki
    - its ApplicationIdeas
see the related discussion at PIM (2009-10-09T16:12:18)

PmGraphViz

distance represents frequencies of messages exchange
color type or last message sent

link to the social aspect of Cookbook.Cognition#Thinking
- add Content.Needs and use categories to automatically select the person most likely to be able to help
- example House.House#Contacts
- imagine this as a "distributed/social cognition API"
  - consider a resume (formalized as XML resume) an equivalent of API definition or WSDL)
integrate my irssi /note
- /alias note trigger add -once -privmsgs -masks '$0*' -command 'echo -level hilight -window &bitlbee $0 -> $1-'
- make note read/write in the wiki instead, letting me modify information outside of irssi
trust systems
- Reputation management according to Wikipedia
  - "the process of tracking an entity's actions and other entities' opinions about those actions; reporting on those actions and opinions; and reacting to that report creating a feedback loop."
- existing commercially inspired
  - eBay scores-reputation and feedback
  - HackerNews karma points
  - Amazon sellers feedback
- apply the same to each interactions with an equivalent FOSS framework/library
  - examples in related communities
    - Slashdot Comments and Moderation FAQ
- research in multi-agent system (MAS) and contract negotiation
  - brief work in 2007 on MAS and collaboration by contract E:\Work\Co-Evolution\Prototype\architecture.ppt
- On Social Reasoning in Multi-Agent Systems by Jaime Simao Sichman and Yves Demazeau, 2001
  - "In our context, such a mechanism entails that agents must explicitly represent, exploit and revise certain properties concerning the other agents [in order to infer some new beliefs from the current ones]."(p1)
- especially in regard to Seedea:Seedea/Onlineoutsourcing
- On Karma: Top-line Lessons on User Reputation Design by randy, Building Web Reputation Systems: The Blog February 2010
- Computational Trust trustcomp.org
- Reputation Patterns in YDN Design Pattern Library
- Reputation Systems for Open Collaboration, Google Research October 2010
- The Monkeysphere Project extend OpenPGP's web of trust to new areas of the Internet to help us securely identify servers we connect to, as well as each other while we work online.
Semantic Signup allowing to populate a user page with semantic data at signup time
a possible visualization could also be a timeline with the social network as a series of circles (in order to give an "entourage" perspective) representing persons (eventually with significant part of their own contacts) in order to get an overview of the evolution of the personal social network
- highlighting periods in the timeline, for example transition from a school to another, a place to another and thus displaying radical reconfigurations
updated contacts section of the different pages
- House#Contacts
- Energy#Contacts
tag cloud per person based on conversation logs
- generate them periodically to see an evolution over time
  - different color for "new" tags
  - use cron and a command line encoder (like mencoder) to generate automatically a "movie" per person
    - if the person will be present in an upcoming event, review his or her "movie" to refresh your memory
- consider Wordle
find a tool generating social network from recorded voices
- train and test through movies
  - use for meeting recordings
- mapping a voice to a person
- linking persons based on the the lapse between each voice caught
- positioning nodes based on frequency, length of reply, ...
- resources
  - auto-captions in YouTube Help
  - PieSpy Inferring and Visualizing Social Network on IRC
    - uses a simple set of heuristics to infer relationships between pairs of users.
    - Shakespeare Social Networks Social Networks based on the plays by Shakespeare (YouTube video)
trombinoscope test
- faces of people tagged E (previous presence at the type of Events) that should be present at the event Im heading at (thanks to the calendar recipe)
- random walk displaying the profile picture of each then asking for a name, upon click display the name under the picture then upon click move to the next
- objective being to avoid awkward "hey... mister!... what's your name already? Sorry I forgot..." moments
display shared links from Irssi#Note in order to keep an history
- useful to links that friends lost
store rendez-vous waiting time statistics
- produce a small visual (using for example Sparklines)
Information context
- use Google Trends (Insights) to check the context the person you are interested in lives in
  - according to the local situation, from the broadest (world) to the smallest (his or her city, even the neighborhood)
  - in time, from years ago up to now
embed good practices or solutions like FamilyFatigueSolutions
tech-savy (or better security-savy) trust scale
- principle
  - filter sensitive content (especially URLs) from contacts
- motivated by IM drive-by-downloads hacks and Social Engineering Toolkit (SET)
- see also trust systems
Cultural Intersection
- find what connects you and other the other person you are talking too thanks to
  - the current location
  - geographical origon of persons present
  - historical background of persons present
- locate where their cross
- limit to what should be the most interesting to develop
- cf notes from the paper notebook written in Italy
socially shaping algorithms
- list the algorithms that impact your own social life through their usage by you or others
  - Facebook EdgeRank
  - Google PageRank
- avoid listing implicit algorithms or models of how we behave
  - e.g. no Wikipedia:Condorcet method despite its fundamental impact should not be listed here
- guest proposing Le livre des algorithmes maudits by David Guez, poptronics
  - during the show Le transhumanisme in Place de la toile on France Culture, June 2010
- Seedea:Seedea/CollaborativeManagementTools as some deals with reward (financial or social)
- TunkRank based on the TunkRank algorithm by Daniel Tunkelang
- Stack Overflow reputation, Hacker News karma, Quora reputation
- see also Eli Pariser's TED talk
- we are in a society in which we dont pick who we collaborate with based on the shape of their car but rather on the quality of the algorithm they use (and consciously decide to avoid too) and build
- Quora Analytics: Who on Quora have the highest q-indices?
  - q-indices, q-index, particularly relevant now that credits are used to request answers
social notification on updates
- when Events, Museum, Trips or other activities where done with contacts, and pages are updated, prove a simple eventually automated way to notify them
- example
  - going with Paola to the Quai Branly museum in Paris to see the Teotihuacan exhibit then watching months later The Decline of Classic Maya Civilization: A Systems Perspective talk by Jeremy Sabloff
trigger on names during conversation to display link to profiles
- ability to deactivate specific profiles that are too common
use the system for the family tree
automated counter-Shibboleth
- briding the gap between what you know and what they say
- A = stats on your vocabuary based on email or chat logs
- B = stats on the tribe you want to be part of based on email or chat logs)
- B - A = what you must learn
Real-time meta-data analysis of chat
- per channel
  - social network
    http://www.jibble.org/piespy/
  - activity
    http://pisg.sourceforge.net/
  - history
    links never previously shared
- per sentence
  - link information
    - visual preview
    - already visited
      - present in own wiki
- per person
  - profile
    - location
      - geoip
    - most used words
      - word cloud
      - Wikipedia:Statistically Improbable Phrases
    - most mentionned other person
- See also
- Real-time and caching
  - costly computation wth low impact changes should be postponed to a low activity moment (usually night) rather than done in real-time
    - e.g. example statistics

Person {Person}