Interview: Three questions put to …

... Professor Dr. Ralph Ewerth, Head of the “Visual Analytics” Research Group at TIB, about the 2nd Conference on Non-Textual Information

The interview, also published in the latest issue of TIB News, is about software and the role it plays in science.

TIB will be hosting the 2nd Conference on Non-Textual Information on 10 and 11 May 2017. The title of the conference is “Software and Services for Science (S3)” − what topics can participants expect to hear about at the two-day event?

There will be an exciting and varied programme. Experts have been invited to deliver a wide range of lectures in the sessions, including topics such as the sustainability of software infrastructures and software archiving, the collaborative use of newer web applications such as Jupyter Notebooks in science, persistent software citation, as well as legal and practical issues that are of great relevance for scientific software.

Portrait Prof. Dr. Ralph Ewerth
Head of the Visual Analytics Research Group // Photo: TIB/EUROMEDIAHOUSE

Luckily, we have managed to gain Edzer Pebesma, Professor of Geoinformatics at the University of Münster, as a keynote speaker. In his lecture, he will address the incentives and reward systems for scientists involved in the development and provision of software: why do researchers make their software programs available, what effect does this – or should this at least – have on their reputation? In another session, participants will discuss the use of distributed databases such as Blockchain. So far, Blockchain has become notorious as a basic technology for digital crypto currencies – bringing to mind Bitcoin. The question of how such technologies can be used in science is an exciting matter.

In the evening of Day One, participants will have the opportunity to exchange views and get to know new people in a relaxed atmosphere in a pretty setting – the Old Town Hall in Hannover.

How important is scientific software in science today, and how has it developed in recent years?

One development is definitely that more and more software is being provided as open source. Many external funding bodies now expect the researchers they fund to make available the software programs they develop in such projects as open source. Another aspect is that the reproducibility of research results can be improved on the basis of freely accessible software (and research data). In some subareas of computer science, for example, the reproducibility of results was somewhat neglected in the past.

In information retrieval or pattern recognition (computer vision, etc.), for instance, results are usually dependent on algorithmic parameters, some of which are often not even mentioned in the publication or cannot be portrayed in a “traditional” publication at the level of detail required. For this reason, recent information retrieval conferences have concentrated on topics that address these aspects. Examples include the workshop on “Reproducibility, Inexplicability, and Generalizability of Results”, held at ACM SIGIR (Conference on Research and Development in Information Retrieval) in 2015, and the “Reproducible IR Research Track” at ECIR (European Conference on Information Retrieval) in 2015.

From the past to the future: where is the development of scientific software heading, and which role can infrastructure facilities such as TIB play?

The importance of open scientific software is likely to continue to increase. If it becomes more common to cite software properly, it may gain in relevance for the reputation of scientists. Here, infrastructure facilities such as TIB can help design technologies and standards, as well as develop and offer relevant services. These could include search technologies for scientific software, the linking of relevant publications and research data, and the provision of solutions for matters such as the persistent referencing of scientific software.

More informaion about the 2nd Conference on Non-Textual Information