email us at Bioinformaticsolutions @ gmail . com

Friday, October 28, 2011

NIHR / NOCRI Information Systems Workshop - Part 6 - Model Driven Engineering for Translational Research

Professor Jim Davies, Director of the Software Engineering Programme at Oxford University, is always an engaging speaker (we were lucky enough to catch him at the NCRI-caBIG joint conference earlier this year), on this occasion delivering an inspiring overview of model-driven engineering for translational research.

Prof. Davies talked about ‘disposability’ as a key consideration in systems engineering – we loved his lo-tech and low cost approach – he asserted that development project costs (as charged by third party organisations) are, in his opinion, out by at least one order of magnitude (did he say two?) – i.e. 50K projects could be completed for 5K, especially using model-driven techniques. 

Modelling languages are becoming programming languages, he noted and he also numbered amongst those advocating the use of open source solutions, mentioning OpenClinica software for clinical trials management. He covered metamodels and stated that “all our models are semantic metadata” before issuing the following grand challenge: “to make model-driven management of data standard practice in healthcare research and to ensure that every item of data is linked to a computable semantic representation!”

This is the abstract for Prof. Davies’ talk:

“Using model-driven techniques, we can generate or configure parts of an information system based upon abstract descriptions of structure and processes. These descriptions serve two purposes: they are "source code" for the system, but they are also documentation - they are items of metadata. By managing the production and evolution of this metadata, we can increase the quality, value, and interoperability of our data without increasing the costs of data collection, management, or linkage. In this talk, I will explore the use of model-driven technology in CancerGrid and other projects, and discuss some of the services needed to support metadata production and evolution.”

No comments:

Post a Comment