Where we are and how we got here

Since about 1994, work has been underway in the University of Pennsylvania's Database Group on mediators - systems which facilitate the querying of distributed, heterogeneous data. The first fruit of this effort was a system called Kleisli. Described by Dr. Limsoon Wong and others, Kleisli is an easily extensible system with a powerful query language called CPL. It is named after Professor Heinrich Kleisli, the mathematician who discovered the characterization of monads that plays a central role in the manipulation of collections in the Kleisli system.

Through a close collaboration with the Penn Center for BioInformatics, Kleisli was expanded and refined. For more information about its use at PCBI, see the paper titled BioKleisli: A Digital Library for Biomedical Researchers.

In 1996 the Database Group began a collaboration with SmithKline Beecham (now part of GlaxoSmithKline) to try out Kleisli in a production environment. Through this work, Kleisli has been connected to over a dozen different data sources, and several parameterized queries have been developed and deployed to hundreds of research scientists. If you're reading this from behind the GSK firewall, you can go directly to the Kleisli homepage for more information.

K2 began life in 1997, when Val Tannen and Jonathan Crabtree wanted to incorporate some new research results into a more portable, flexible implementation. Scott Harker was added to the project shortly after, and is the current maintainer of K2.

[K2 Front Page]