zKnowMan �

Information Architecture for Knowledge Management

Who: Any enterprise

What: Needs knowledge to drive their processes

Where: From internal, external (proprietary) and public content sources

When: In Internet Time (where instant gratification isn�t fast enough)

Why: To meet business objectives � growth, profit, stock value

How:� By implementing a knowledge management portal that implements:

Process	Activities
Gathering	� Data Entry � OCR and scanning � Voice input � Pulling information from various sources � Searching for content
Organizing	� Cataloging � Indexing � Filtering � Linking
Refining	� Contextualizing � Collaborating � Projecting � Mining
Disseminating	� Flow � Sharing � Alert � Push

Source: techtarget.com

Architectural components:

1. Taxonomy (organization)

2. Rule based expert system (automated analysis)

3. Portal (collaboration and data collection)

Source: KPMG

Goals:

Analyze what people want

Deliver information the way users want it

Challenge:

Poor information flow bogs down a lot of processes.

Cultural Barriers to information sharing � align compensation to sharing.

Tacit vs. Explicit Information � provide tools to capture more tacit info.

Organizing around user tasks can mitigate information overload.

�� Posting � by department

�� Searching � by type

�� Known item searching � stock price

�� Existence searching � Mutual funds for Moldovan startups?

�� Exploratory searching � High level executive bullet point results

�� Comprehensive searching (research) � IMAP implementations

Source: KPMG

Techniques:

1. Multiple organizational indexes = Multiple navigational pathways to content

�� Topic

�� Task

�� Audience

�� Company Organization Chart

�� Chronological

�� Geographical

�� Who, What, Where, When, Why, How

�� Cluster analysis

2. Automated validation of the organization of content -

Searches that come up empty give more valuable feedback to the design of taxonomy than path analysis � Information Architecture, O�Reilly

3. User centered design -

Applications shouldn�t dictate how people work

Alternating browse and search.

Tools:

1.� Use of standard and custom taxonomies:

Industry Standards

Barcodes & SKU

Phone Numbers, Yellow Pages Headings � infospace.com

�� Govt. Standards � firstgov.gov,

SIC / NAICS - Economic Census, State Info

�� DISA(Defense Info Systems) � Standard Data Dictionary

Proprietary:

��SRDS

��Gail Association Directory

�� Predicast Event Codes

�� Dow Jones Industry Group Center

�� Reuter's News Categories

��

Library Organizational Techniques:

Library of Congress Subject Headings

Thesaurus � controlled vocabulary w/ relations between terms

Citation Index

Internet Indexes and Directories:

web addresses � links & references (like citation index)

emails � analyze contents of sent mail � build user profile

�� Newsgroups (moderated, non alt)

DMOZ.org

Guru.Net

2. Automated categorization via rules:

�� Rule based expert system with fuzzy logic

�� Personalized agents for individual users

3. Portal technologies:

�� Distributed access with synchronization

�� Web

�� Laptop

�� PDA

�� Cellphone � WAP, SMS, Phone number synchronization

�� Distributed content creation process

�� Workflow

�� Cascading Portals

�� Versions (like Oracle Versions)

�� Undo all data and actions (doubles as audit trail)

�� User definable role-based security

�� Collaboration

�� Groups

�� Whiteboards

�� Chats � slashdot, instant messaging, wiki

�� Personalization

�� Bookmarks

�� Agents � automated searches and tasks

�� Calendaring

�� Email

�� To do

�� File sharing

�� Proxy based web surfing

�� Web based, no browser changes

�� Automatic ad removal

�� Storage and retrieval of visited sites

- search for a site with vitamins visited last week

Search, Sort, & Categorize using taxonomies and ontologies

- Dmoz, SIC, Organization, time, project, group-visited.

- find magazine articles - http://www.findarticles.com/

Handicap enable on the fly

Seamless translation on the fly

Content filtering � adult, violence, etc.

Graphical overview of content � WebTOC, Market Matrix.

Readability analysis on the fly � Flesch, Fog Index

�� User programmable agents and automation

- refresh frequently visited sites

- optimistic read ahead

- spidering of a site

- compression of text and graphics to enable faster surfing.

Security � freedom.net, zeroknowledge, Conceal private desktop

GUI configuration � Javascript based Menus, Tabs (DynAPI)

�� Integration paths to existing IT infrastructure

�� MS Exchange

�� Lotus Notes

�� Electronic Document Management

�� SQL Databases

�� Legacy Systems

�� CRM � Pega, Vantive, Seibel

�� ERP, Back Office

�� Open

�� Cross platform (Unix, Windows)

Source code availability for high customization

Third party add-ons

Internet RFC standards compliance

�� IMAP

�� LDAP

�� Calendaring

�� Workflow

�� RDF

�� HTTP

Dublin Core Metadata and Z39.50

Ontoknowledge.org

XML schema support � www.oasis.org

�� Industry specific data and functions

Reference Implementation Prototype:

For Southern Company, a reference implementation in CLIPS / Zope / WorldPilot was prototyped.� The reference implementation serves to validate the architectural components, and is not meant to be an inflexible implementation standard.

References:

Web sites:

You think Tomaytoes, I think Tomahtoes � Information Architecture paper

http://www.cio.com/archive/webbusiness/040199_nort_content.html

�

Implementing Knowledge Management � Good overall KM paper

http://www.uis.edu/~khaund/ntemp4.htm

The Facts about Knowledge � European viewpoint

http://www.info-strategy.com/knowsur1/

Academic Presentation � Good diagrams

http://www.mbs.umd.edu/is/malavi/icis-97-KMS/sld001.htm

KM World - Magazine

http://www.kmworld.com

Books:

Information Architecture

Louse Rosenfeld, O�Reilly

Knowledge Management Toolkit

Amrit Tiwana (GSU Professor), Prentice Hall