AP UX Week '07: Semantic Technologies
[Live blogged from Adaptive Path UX Week 2007. Semantic Technologies by Cameron Hunt]
Dynamic context to search, display, and workflow
What are Venn diagrams? Relationships, overlap. What is context? Relationship.
Background
Why do people cut their hair the way that they do? Why do they wear rings, etc.?
As a missionary in Venezuela he learned that community is the single most important influence on behavior. Think about yawning and the cascade of power.
Social narrative driving sustainable change.
Combining experience in anthropology, technology, and military intelligence to think about these issues.
So what is context? Concentrations of influence interacting in time and space. Boundaries, interactions, causation, influence.
So what are semantics? Aspects of meaning expressed as language. Axioms, concepts, rules.
All of our concepts and relationship are based on concepts that define how we categorize and organize.
So what are semantic technologies? Mathematicians have formal logic and we want computers to have this to. These are ontologies. This is expression in a standard format. We then want inference engines that discover implicit connections between concepts enabling dynamic context. This is where emegence, self-organization at a meta level, starts to occur.
Vietnam = Impedance Mismatch
The narrative did not fit reality. Successful corrective action was bottom-up. Terms were redefined, new ones were introduced. New "characters" were added, people assumed new roles. Technology was adapted to fit the new narrative and reality.
Iraq = Impedance Mismatch
Micromanaged narrative, changing reality. Corrective action reached political/military boundaries. There are some successes: CIDNE and Human Terrain System.
These solutions also emerge from the bottom up and emphasize the importance of local political considerations.
So what do we learn from this?
We want more expressiveness and new capabilities.
We need to separate data from concepts models. We want multiple hierarchy types ("is a" "part of"). Powerful relation properties emerge from this that are transitive and symmetric. We can do inference based operations (search, display, workflow), control concept access, and evolve concepts.
So do we let end users create new entities? We are talking about taking data and letting people create new structures, new combinations. This let's us connect diverse communities (federation, integration, association) with flexible/sustainable data mapping. Map Service Oriented Architectures to
Check out ILOG, Thetus, and Liferay.
So let's define some terms.
A Personality is a collection of ontologies and data. It provides foundational context for search, display, and workflow.
Think of the data structure in a simple structure. Map relationship properties and data type properties. These properties have domains and can be dependent.
Constraints can be concepts, properties, and instances. Individual search result items return dynamic property lists.
[Basic Search Demo]
This structure let's you put data anywhere and map concepts across data sets.
Let's look at domain specific ontologies. Unique ontologies provide concept model evolution, allow for protected access to concepts/data, and can be merged with other concept models. You can hide concepts, and assertions made on those concepts, until they are properly evolved. You can define behaviors or low-level relationship characteristics as sub-sets of other relationship characteristics. An abstract concept model allows you to map sub-concepts.
-1 TrackBacks
Listed below are links to blogs that reference this entry: AP UX Week '07: Semantic Technologies.
TrackBack URL for this entry: http://www.samfelder.com/mt/mt-tb.cgi/313
Who is this guy?
Leave a comment