(Interface language should have been autodetected, if not: English | Polski)

Carrot is a clustering engine for Web search results. It constitutes a part of a larger project, which is Dawid Weiss' master thesis. Despite all the efforts put to ensure reliability and quality of Carrot, it is still a research project and as such lacks strict testing, reviews and engineering process of a commercial product. You may expect failures and exceptions - in such case an error report should come up and you are kindly asked to send the feedback for Carrot's further improvement. Author takes no liability in how Carrot works or what its results are like and are used for. Carrot has only been tested with Microsoft Internet Explorer 5.0+. Netscape browsers are not supported at this moment (and are said not to work properly). All suggestions are welcome through the feedback form.

Some of the resources, utilized during the writing of Carrot and my master thesis come from online Rzeczpospolita archives. Used by permission. Polish lematisation engine is based on the work of prof. Jan Tokarski - Schematyczny Indeks a tergo polskich form wyrazowych, edited by prof. Zygmunt Saloni. The electronic version of that book as well as the permission to use it was granted by prof. Saloni.

Query results are obtained from Google^(tm). As for now it is the only datasource implemented. I would like to ask Google to contact with me if it is illegal in any way. I have tried to obtained legal permission, but never received any answer. It is pretty obvious that graphical design of Carrot mimics that of Google's. I believe that excellent examples should be followed, and Google is the most ergonomical and aesthetic user interface I have seen so far. I made a number of small improvements, like the status window, which I hope make Carrot a little bit different from Google (at least in the sense of legal rights :).

Carrot was written in Java. Grammars and parsers written using JLex and JCup by Elliot Berk and Scott Hudson. The system was written using JBuilder from Borland, configuration management using CVS, build scripts in ANT, Tomcat from Apache Software Foundation used as an HTTP / JSP server.

STC algorithm based on the work of "Grouper" Team, especially on the PhD thesis by Mr. Oren Zamir. Grouper was the first clustering engine, developed at University of Washington.

I would like to express my gratitude to these respectable individuals for their help: Jerzy Stefanowski, Irmina Masłowska, Krzysztof Szafran, Zygmunt Saloni, Grzegorz Jagodzinski, Tomasz Obrebski, Oren Zamir, Bogdan Dorohonceanu, Andrew MacFarlane, Marek Kopyt, Jon Knight, Luis Gravano.

Don't show this page for 2 weeks