spacer
Services Methods Customers Network Presentations Bio/Contact

Applications

Click here for Applications of the Data Projection Model.

Demos

Click here for Demos of the Data Projection Model.

The Data Projection Model was presented:

2008-04-10. The Data Projection Model (PDF), SGML/XML Users' Group, Washington DC, April 16, 2008

2007-07-24. The Data Projection Model. Making Information Auditable (PDF Version), (Powerpoint), July 24, 2007, Bobst Library, New York University, New York.

2006-10-11. The Melting Pot of Information: Enabling Multiple Perspectives, including Auditing (PDF version), October 10-11, 2006, Fifth Semantic Interoperability Conference, MITRE, McLean, Virginia. Powerpoint version., HTML version.

2006-03-14. Multiple Perspectives: From Wishful Thinking to Implementation, Introduction to the Data Projection Model (PDF version), Collaborative Expedition Workshop #49, March 14, 2006, National Institute of Standards and Technology (NIST), Gaithersburg, Maryland. Workshop Agenda and Powerpoint Version.

Accountable Information: The Data Projection Model

Michel Biezunski, The Data Projection Model: Toward Auditable Information System Through Unified Declarations of Operations on Data, November 14, 2006. ©2006, Michel Biezunski

Requirements

The Data Projection Model is designed to facilitate accountability of information, by enabling detailed customized views of processes that occur within information systems. It is used within the following contexts:

Information is messy, and it is likely that this will always be the case. Rather than pretending that we are going to eliminate the mess, we can use it as a fact of life, a starting point.

There are many possible ways to say the same thing. Whether we are using different languages, different expressions within the same language, different styles of presentation, words belonging to different vocabularies, or different classification schemas, we want to get our point across. And with the exception of isolated and closed communities, where every one is supposed to speak with one voice, we won't be able to get rid of the diversity.

In spite of the preceding, information needs to be trusted and to be accounted for. Organizations, such as governments, have to be accountable for the information they provide.

Proposed Solution

Every information item is a step in a process.

All processes can be expressed as a set of one-to-one relationships.

(This is the same idea which is at the core of RDF, Topic Maps, and other graph-driven information models.)

From the elementary components, views can be rebuilt according to multiple perspectives.

Usages

Integration of taxonomies and ontologies, emanating from various sources, various people, at various times.

Creation of specific views for specific purposes, including maintenance, quality control, reporting, auditing.

The Integration Challenge

Integration in 2 steps

spacer

Provide a hospitable environment for items to coexist and connect with each other, even from different origins.

Enable various perspectives on the same information objects, i.e. freedom of speech.

Provide accountability, transparency, so that clear policies can be established.

In other words, integrating information systems entails similar challenges to those of running a democratic and open society.

1st step: decompose a system into its elementary processes expressed as binary relationships.

2nd step: rebuild views integrating the components: Each view conforms to a given perspective. Multiple perspectives are possible.

Projection and Perspective

Perspective is a set of methods to represent three-dimensional space on a flat surface. Geometrical laws of perspective express what is invariant given various points of view.

Perpective expresses projection from 3D reality to a 2D flat surface. Similarly, information models are often multidimensional. Each of them can be decomposed into a number of binary relations. These binary relations form a 2D-like space. Therefore, perspectives appear when viewing projected data expressed as binary relations. Hence, the name of "Data Projection" chosen for the current model.

spacer

More on the Data Projection Model

Applications, demos, relations with existing standards and technologies, and other technical details: click here.

 

Infoloom - 402 85th Street 5C Brooklyn NY 11209. Voice: +1 (718) 921-0901. Email:info@infoloom.com  Web: www.infoloom.com
Page created October 16, 2006. Last updated September 17, 2008.

gipoco.com is neither affiliated with the authors of this page nor responsible for its contents. This is a safe-cache copy of the original web site.