Back around the year 2000, Gunter Zeilinger wrote the popular JDicom utility suite using the Softlink (now TriSpark) Java DICOM Toolkit (JDT). After this experience with the JDT, he decided that he should write his own DICOM toolkit. Thus was dcm4che (pronounced d-c-m-for-chay) born. Gunter, inspired by the revolutionary nature of open source, decided to name the toolkit after a famous revolutionary. It just so happened that the name worked out to sound something like d-c-m-for-jay, which is a pretty common way to name a project in the Java world.
The original version of the toolkit was designed with the intention of submmitting it to Sun as a standard Java-based DICOM API. With that in mind, the toolkit was separated into an interface layer and an implementation layer. The JSR never happened, but the project started gaining in popularity among the Java DICOM development crowd. Shortly after, the dcm4jboss archive was born. This sub-project implemented the IHE Image Manager and Image Archive actors. dcm4jboss has evolved into a robust, stable platform for managing imaging studies and reports. It has participated (and succeeded) in the IHE Connectathon several times and is deployed in production at many sites.
In the recent past Gunter felt the need to re-architect the DICOM toolkit. Performance, memory usage, flexibility, and implicity (removing the JSR-motivated interface layer) were all driving factors in this decision. This new version of the toolkit has been labeled dcm4che2 (or dcm4che-2.x). With this version of the toolkit, Gunter has truly raised the bar for Java DICOM toolkits.
The major missing factor in the project however has always been documentation. Developers and users had to resort to reading the code and tribal knowledge in order to be able to benefit from the software. That's all changing now...... http://www.dcm4che.org