Autonomy's Enterprise Search with Clustering

Search tool analyzes context, user profiles By: Kathleen Lau, Computer World Canada (March 31)

Autonomy added new features to its enterprise search platform, IDOL, that provides for "deep video indexing and quantum clustering".

"The goal of IDOL, or Intelligent Data Operating Layer, is to provide a single unified and vendor-neutral platform for searching all file formats and media-types for legal and business purposes, said the San Francisco-based company’s CEO, Mike Lynch."

Features:

+ "intent-based ranking bases search results on a user’s profile and other contextual factors, instead of solely depending on popularity or keyword matches".

+ "the quantum clustering feature applies quantum mathematics to calculate concepts within data so that users can better identify conceptual information"

+ "enhancements to existing media support for deep video indexing and analytics."

It's another advance in meaning extraction and meaning-based search. Could advances with this technology obviate the need for taxonomies or will the controlled vocabulary of the taxonomy help in expressing meaning?

Reports on Sharepoint

The March 2008 issue of the Montague Institute Review has two articles on Sharepoint.

http://www.montague.com/le/alerts/le0308.html

From the newsletter:

Sharepoint conference report: Report on what's new at the first public Sharepoint conference in Seattle on March 3 - 6. Includes information on the FAST acquisition and a new federated search function.


Sharepoint search: An enterprise contender? (free but requires registration)
Montague Institute founder Jean Graef thinks that MOSS 2007 will suffice
in some situations -- but by no means all.


Jean Graef commented on MOS capabilities for using existng metadata and thesaurus relationships.

Color and Information Design

Color Chart: Reinventing Color from 1950 to Today
[Macromedia Flash Player]
http://www.moma.org/exhibitions/2008/colorchart/flashsite/

This online exhibit on color from the Museum of Modern Art is a fascinating for the way it has organized information and used multiple points of views. Visitors can browse by artist name, medium (photographs, paintings, sculptures, drawings and prints, and other), and timeline by year starting in 1918 to the present. The commercial color chart is the organizing premise or starting point.




The site "handles" very well in navigation and has a good mix of image, audio, and text. When viewing the art work, watch for links to Full Caption | Extended Text.

There are also four videos of artworks being installed at MOMA.

Postscript (Mar 10, 2009) New Link - http://www.moma.org/visit/calendar/exhibitions/30 - page about the exhibition with link to the exhibition site and a link to an audio tour.

Checklist for enterprise search

Delivering on the Promise of Enterprise Search -- 10 Issues to Consider, by BRIAN DIRKING for the Enterprise Search Center (Mar19)

Checklists are always helpful. This one is about enterprise search.

"When an organization identifies the need for enterprise search, there are myriad questions it must answer. It is vital to identify the most important criteria for your organization, as they will guide your evaluation and eventual implementation of enterprise search. As a starting point, here are 10 issues every organization must consider to help ensure that its investment in a search solution delivers on its promise."

Taxonomy isn't mentioned directly, but it might come into play when considering scale and scope -- "As you consider what needs to be searched, you will also find new uses and new users for your search technology. With new users comes a question: Should you choose separate search tools for different uses, or can you use a single search solution to address these different applications?"

Types of Taxonomies

Taxonomy Design Types by
Barbara Blackburn, AIIM (May 31, 2006)

Good introductory article on types of taxonomies with examples.

"Taxonomies are usually hierarchical. Categories (nodes) in the hierarchy progress from general to specific. Each subsequent node is a subset of the higher level node. There are three basic types of hierarchical taxonomies: subject, business-unit, and functional."

Information is Overwhelming

National Workplace Survey Reveals American Professionals Overwhelmed, Headed for “Breaking Point” LexisNexis press release (Feb 26)

LexisNexis survey of 650 US white-collar and knowledge workers points to information overload

+ more than 70% of 10 U.S. white-collar workers feel inundated by information at work

+ more than two in five say "they're headed for an information "breaking point"

+ "62 per cent said they spend a lot of time sifting through irrelevant information to find what they need"

+ "68 per cent wish they could spend less time organizing information and more time using it"

+ white-collar workers spend on average 7.89 hours each day "conducting research, attending meetings and searching for previously created documents".

+ 68% would like to spend less time "finding and organizing information and more time using it".

Also see "The High Cost of Not Finding Information", Stephen Abram, Stephen's Lighthouse (Mar 4)

Leveraging Knowledge

Wonder how knowledge management will change through the move to social networks and collaboration? Ross Dawson answers the question in The Future of Knowledge Management (2004)

"In the course of my travels and speaking and consulting engagements around the world, I have found that there are five key frames for leveraging knowledge in organisations that are emerging as the successors to knowledge management, and that executives find relevant, compelling, and actionable."

Document Clustering Software

New document clustering software from HotNeuron.

Hot Neuron Introduces Document Clustering Software Newsbreaks, Information Today (Mar 3)



Clustify (www.cluster-text.com) is "aimed at helping corporations and law firms explore, organize, and tag large document sets. Clustify uses a proprietary, agglomerative algorithm designed to provide excellent cluster quality and scalability. Clustify can help corporations organize their internal documents. It can also enhance search engines by identifying related documents that may be of interest even if they don’t match the search query exactly."

Enterprise 2.0 with Open Text

Two press releases today (March 3) from Open Text show how important "web 2.0" concepts have become in companies and the ever widening adoption collaboration technologies for knowledge management.

Open Text Unveils Enterprise 2.0 Strategy -- Comprehensive Strategy Will Help Organizations Accelerate Social Computing, Embrace Web 2.0, Advance Collaboration and Ensure Compliance; Company Announces New Solutions As Part of Broader 2.0 Strategy.

Some bits:

+ Experience Optimization: Optimize the end user experience through personalization, language preferences, analytics, voting, tagging, blogs and moderation.

+ Meta-Data to Underpin Experience Optimization: Deeper levels of meta-data handling will enrich the user experience. This includes richer meta-data handling for content, but with particular emphasis on richer meta-data for people and processes as well. It also includes blending the best of both top-down taxonomies with bottom-up tags.

+ Classification: Improved intuition to auto-classify content, particularly drawing on past user behaviors and current context.


Open Text Introduces Solution to Accelerate Social Computing and Collaboration in Organizations -- New Livelink ECM - Extended Collaboration Connects People, Processes and Content in Real-Time

"The solution combines a robust knowledge repository with project workspaces, polls, news channels, tasks and milestones. A set of community applications brings specialized, enterprise-ready tools, such as forums, blogs, wikis, and real-time collaboration; along with newsletter views, FAQs, and event calendars to promote shared expertise and best practices."

Book: Tagging: People-powered Metadata for the Social Web

Many enterprise search companies are incorporating elements of collaborative tagging into their systems. The book Tagging: People-Powered Metadata for the Social Web is intended to help people understand the tagging and folksonomies and design for it in their systems.

Peter Morville wrote, "In Tagging, Gene Smith has written the definitive book on designing applications for the social web."

The author, Gene Smith is an information architect, blogger, designer, consultant and "tagging aficionado". He is principal at nForm User Experience in Edmonton, Alberta and has advised clients like Comcast, Ancestry.com and the Canadian Patient Safety Institute. [From bio at Northern Voice.]

Some notes and materials are at the companion web site - Tagging. This has a couple of interviews and some footnotes for the chapters. Potential readers would most like to see the table of contents, but this is not at either the web site or the Amazon page.

Mr Smith also maintains the blog Atomiq which is largely about information architecture and tagging.

For an introduction to the principles of tagging see Tagging 101, his presentation to Northern Voice available through Slideshare.