IBM Unstructured Information Modeler   

Hi There!

I just read this post, interesting stuff.
http://alphaworks.ibm.com/tech/uimodeler/

I would like to add my 5 cents about unstructured information:

Talking about Unstructured Information (Ontology, Search and Taxonomy), the only software I know of where the software really works is InfoCodex. InfoCodex comes with a linguistical database with 3 Mio words and terms in German, French, English, Italian and Spanish. What does this mean? This means IC can actually do a cross-language search and find similar documents in another language. This is true-cross-language-search. Autonomy, Fast, OmniFind, none of those can do that. InfoCodex will also produce a Graphical Heat-Map of all the contents in any given collection. Also see: InfoCodex Procedure

Its true this version of the software is somewhat limited in terms of only working on ASCII text. There is an upgrade that will allow it to work on UNICODE:

First you need to download the package "icu4j.jar" which contains the unicode parser:

http://www.icu-project.org/download /

Then start up UIM using the following command line call:

java -cp uimodeler.jar;icu4j.jar -Dunicode=yes com.ibm.cv.text.EAdvisor

This should cause it to parse the text input as UNICODE instead of ASCII.