Software Design for Empowering Scientists

Abstract

Scientific research is increasingly digital. Some activities, like data analysis, search and simulation, can be accelerated by enabling scientists to write workflows and scripts that automate routine activities. These capture pieces of scientific method that can be shared with others. The Taverna Workbench, a widely deployed scientific workflow management system, together with the myExperiment social web site for the sharing of scientific experiments, have been devised according to six principles of designing software for adoption by scientists and six principles of user engagement.Science is becoming increasingly digital, and the scientist's tools are not just the experimental apparatus of the laboratory but also the software apparatus which they use to conduct their researchto analyse data, to search databases, to run simulations and to record their scientific process. New experimental methods -from DNA microarrays to sensor networks in the environment -are generating volumes of data that, without software assistance, just wouldn't get processed.Watch a researcher at work and we see lots of activity at the computer, using applications, services and data that might be local to the laboratory, the enterprise, or out on the Web. These new research tools and methods are evident across a very broad spectrum of disciplines. Some, like bioinformaticians working with protein sequences, might conduct some of their research entirely in silico. Meanwhile, chemists in the laboratory are looking up procedures, designing their experiments, conducting and recording their work, and running simulations, searches and analyses. Climate change researchers are capturing the latest data from radar images and sensor networks, and running and testing their models. Musicologists are extracting features from recorded music as part of musical analysis. Archaeologists are exploring 3D visualisations of their digs.Some of the processing is highly repetitive -repeating processes with new data, varying parameters, and making small changes to the process. It clearly helps speed things up and makes them more repeatable if some of this scientific activity at the computer can be automated. This in turn lets scientists concentrate on the science rather than the repetitive handling of data through multiple tools and applications. So, increasingly, we see these scientific activities at the computer being 'programmed'. Some disciplines have created scripting solutions, perhaps adopting their favourite scripting languages and creating libraries to handle their discipline-specific data, while others adopt workflow systems which automate the processing of data through a series of processing stages.Taverna is one of many scientific workflow management systems (see Workflow Management Systems on page 2) and supports what we might describe as 'application-level workflows', as opposed to some other systems which focus on scheduling tasks across computing resources [1]. Emerging from the UK's e-Science programme, Taverna is used extensively by scientists a...

Open Science: One Term, Five Schools of Thought

“…Autonomous computers are interconnected in order to achieve high throughput research goals. The Open Science Grid provides a collaborative research environment for communities of scientists and researchers to work together on distributed computing problems [24].…”

“…He further emphasizes the potential to enable large-scale sharing of resources within distributed, often loosely coordinated and virtual groups-an idea that according to the author is not all new. He refers to a case from 1968, when designers of the Multics operating system envisioned a computer facility operating as a utility [24]. What is new though, according to Foster is the performance of such network utilities in the light of the technological progress [24].…”

“…He refers to a case from 1968, when designers of the Multics operating system envisioned a computer facility operating as a utility [24]. What is new though, according to Foster is the performance of such network utilities in the light of the technological progress [24].…”

Machines, methods and music: On the evolution of e-Research

“…This is because each one comes prepackaged to solve particular problems for particular research communities, familiar when they open the box. This is another example of co-evolution in action, and demonstrates that adoption of a technology comes by focusing on the specific before the generic [8].…”

Some Challenges Facing Scientific Software Developers: The Case of Molecular Biology

“…The successful development of the Taverna workbench and myExperiment web-site (Goble and De Roure, 2007) is somewhat at odds with this tradition. De Roure and Goble, both computer scientists, describe how they first of all implemented software to meet the specific needs of specific groups of biologists, and then reflected on how these needs and hence the software might be generalised (De Roure and Goble, 2009). This bottom-up approach, as opposed to the more traditional top-down approach of software engineering, led to the production of software which proved very successful in supporting biologists.…”