In 1658, Richard Franck wrote in his Northern Memoirs: "Art imitates Nature, and Necessity is the Mother of Invention."

Hermes was a product of necessity, born at the Massachusetts Institute of Technology in 2002. At this time the Department of Biology at MIT enlisted the help of the Laboratory for Computational Physiology (LCP) for its expertise in physiological signal processing and interpretation. The expected task was that the LCP would analyze existing ECG recordings from genetically engineered mice in order to elucidate the disease process in these mice as it unfolded over the course of 6-8 weeks (at which point the poor critters invariably died). Upon learning about the systems employed by the Department of Biology, the LCP realized that interpretation of the data would be nearly impossible. The systems in existence were not intended for long-term collection. The work flow for collection of long time series of data required that a lab technician visit the animal facilities (a relatively inaccessible clean room environment in the basement of MIT) every 12 hours. With each visit the technician would stop the laptop used for data collection, save the contents of the hard drive to a CD-ROM, then re-start collection. By the end of an experiment the technician then had a rather large stack of CDs from which he/she would use software to format the data into a simple ASCII format; the technician would then import the data into Microsoft Excel and create a chart of the data. Given that the data was sampled at 250 samples/second and MS Excel has memory limitations, the tech needed to perform multiple iterations of importing consecutive data segments and printing a chart. After several arduous weeks of this task the Biology Laboratory had a very thick ream of paper representing a continuous time series of ECG for a single subject.

Unfortunately for both the LCP and the Biology Department, the paper reams were an unacceptable data format for software attempting to automate ECG interpretation. Obviously the lengthy process of formatting and collating records also precluded any real-time analysis for interventional studies. Attempts at stringing together multiple consecutive 12-hour digital data segments proved to be extremely arduous and only partially successful. At this point the LCP decided to design an entirely new architecture. Project Hermes had begun.

The design requirements could be summarized in a simple sentence: build a system that is easy to use, requires no user intervention, and provides automated ECG data interpretation for extremely long periods of time. The implementation of the system was not so simple, but after two years of intense hardware and software design at the LCP, the Biology Department was collecting and analyzing data at unprecedented rates with Hermes as its infrastructure. The requirements that served as the impetus for Hermes' invention - simplicity and automation - should remain the mantra of any designer implementing changes to Hermes today.