threshold of confluence
DESCRIPTION
My presentation at the Media Ecologies Workshop 2009, at Salford University in Manchester. A short presentation where I attempt to explain briefly a practical approach to interoperability, based on gradual adoption of supplementary standards which do not assume or attempt to "predict the unpredictable", but rather acknowledge the internet software landscape as a complex system where the reasonable course of action is to "expect the unexpected".TRANSCRIPT
on the thresholdof confluence
Steinn Eldjárn SigurðarsonCELSTEC2009
Evolution (1/3)
usersSYSTEMS >
Evolution (1/3)
few usersfewer systemslow interoperability
not very connected at all
limiting factors: basic connections
Evolution (2/3)
usersSYSTEMS >
Evolution (2/3)
more users
more systems
data flowing a little bit better between systems:
search engines, e-mail, basic groupware
limiting factors: systems (software)
Evolution (3/3)
<SYSTEMSSYSTEMSSYSTEMS
SYSTEMSSYSTEMS usersusers
usersusersusers
usersusersusersusers
users
Evolution (3/3)
today.
a lot of users
a lot of systems
proportionally, every user engaging in several different systems
limiting factors: time, usefulness
That's a lot of data..
multiple e-mail addresses
multiple blogs
multiple social networks
wiki's
google docs
collaborative softwares
where do I spend my time?
So what changed?
(since web 1.0)
What changed..
New limiting factors: user's time and preference
Old limiting factors: systems and their possibilities
As opposed to systems being few & users using whatever they can have, now users have choice, and a system without users is inherently worthless (Metcalfe's law).
Systems compete for users
Users migrate
Hundreds of types of systems
Thousands of instances
Is this some kind of problem?
Does this mean anything? Is this some kind of problem? It sounds pretty great to have all this choice
actually...?
Yes it is a “problem”
How do I access all my data?
How do I make sure the people I work with see the data I want them to see?
What happens when I want to move all my data somewhere else?
How have we attempted to solve this problem?
Mostly from a system perspective:
Ontologies and standards(SCORM, LOM, IMS-* in education)
SOA and web-services(SOAP, XML-RPC, REST)
But what has been most effective?
Good, old, underrated RSS! (and HTTP)
Why?
Because it's SIMPLE!
What do the following things have in common?
blogs
forums
instant messages
status updates
...
(some) Shared characteristics
created by
intended for
title and/ordescription
time of creation/modification
usually part of a stream/feed/list
How are we trying to exploit this?
Three dimensions:
Target (person/group)
Content size
Item mean time
Dynamically typing feeds?
FeedBack (spec)
Simple specification for push-based feed updates (& advertising).
Specifies:
feedback.offer(sub_uri, targ_uri)
feedback.request(sub_uri, token)
feedback.notify(data/uri, token)
FeedBack (cont.)
Implementations ready for
WordPress, Moodle, Scuttle
Interested in writing one? We have some docs and a validator :-)
Thanks..
Thanks to:
Fridolin Wild (Open University UK)
Zuzana Bizonova (Žilinská univerzita)
(and others.. sorry)