Translation of stranded IP to commercialisable software applications using workspace

Abstract

For Australia and many other countries, which have large gaps between their research inputs and economic outcomes from their national innovation systems, understanding where and in what form the IP generated resides represents a significant opportunity. Aside from publication outputs (which are usually the primary measure of research excellence) and registerable IP (patents etc) which form only a small amount of the IP generated, there is significant development of enabling tools and data. However, such IP (termed "stranded IP") can often have value to third parties if these opportunities can be identified, the IP put in a suitable form and then provided to them for use.Usage of research outcomes and research software can be highly valuable to the right customers if they can be provided in a usable form that is fit for their usage purpose in a cost effective and timely manner. Research IP is typically developed for specific research purposes and projects (most commonly with the aim of publication) with opportunities for translation to third parties either not considered or identified significantly after the form of the IP has been determined.. The form of the IP is usually decided by individual researchers or teams based on what suits their current needs rather than strategically or with the intention of translation or later commercialisation. It is also common for enabling IP to be part of a larger process or workflow which usually then also involves multiple manual or semi-manual steps. In the modelling and simulation space, this is particularly common in relation to the preparation of model inputs and the analysis, preparation or presentation of outputs. These represent substantial barriers to translation, as potential users judge the cost of expert knowledge acquisition to be too high and the distortions to their internal workflow and decision making processes to be too costly and/or risky. It is useful to consider research activities as workflow processes, being composed of a series or network of unit operations linked with input-output relationships and execution order dependencies. Typically some or many of these operations or steps will use software components (often including proprietary or open third party sources). The intervening manual steps present strong hurdles to the automation of these semi-manual workflows which inhibit reproducibility and transfer to third parties. It also allows errors in the use of the IP to be introduced. An effective solution is to convert all steps into workflow unit operations that can be manifested in a scientific workflow system. This can facilitate high levels of component re-use, improved collaboration, interoperability of the software components in an extensible and customisable way. This provides the ability to adapt to new and emerging requirements and to customise IP on a per customer basis at comparatively low cost. This paper explores the nature of stranded IP and the obstacles that limit its exploitation. It then explores how the adoption of a software workflow platform, Workspace in this instance, can overcome a sufficient subset of these obstacles so to provide a low cost pathway from internal research to exploitable software applications and products. A methodology is proposed, based on the capabilities of Workspace, for taking legacy, enabling or stranded IP, particularly in the modelling and simulation space, and building translatable and commercialisable products.