In 2011, Xavier Rival received the very selective ERC (European Research Council) grant. As a result, he benefited from five years and 1.5 million Euros to work on the automatic verification of critical software.
Our objective was to work on data structures within - to begin with - the relatively classic framework of operating systems or embedded programs
The emergence of a new opportunity
At the heart of this research, one encounter was to open up new possibilities. When Tie Cheng suggested coming to do an internship in the Antique team, Xavier Rival was immediately intrigued by his atypical profile - combining both experience in the private sector and research. “As he was specifically interested in spreadsheets, I suggested he work on their verification”, he continues.“Everybody knows Excel, and the prejudice would be to consider it as office software with no scientific interest.And yet, a single error can have serious consequences. For example, I think of the paper "Growth in a Time of Debt" written by two Harvard economists, Carmen Reinhart and Kenneth Rogoff, which establishes a strong link between public debt and growth. Moreover, their conclusions were able to be used to justify the austerity policies during the crisis of 2008. However, their proof was based on programming errors in the spreadsheets which, once corrected, come to a quite different conclusion...”
And so, thanks to this encounter, the research framework of the ERC grant was enriched with this new dimension on the verification of spreadsheets which, as it progressed, gave a glimpse of transfer opportunities. After defending his thesis in 2015, Tie Cheng created a start-up, Matrix Lead, and the project received the ERC Proof of Concept grant of 150,000 Euros to develop its structure.
Being part of the long term
It is now time to recruit cutting-edge engineers to make this tool - originally very complex - the most easy to use. The challenge is also to start working with the first potential clients and refine its properties and functionalities. However, as Xavier Rival reminds us: “The verification of programs is complex and takes a lot of time. The transition from science to transfer cannot be done in a year. For example, the transfer of the Astrée analyser, designed and developed within the team since 2001, could only be envisaged in 2007, and the project found its cruising speed with a tool used industrially in 2010. In our fields, we therefore need to have this long-term option. That is precisely what a grant like the ERC Proof of Concept enables us to do.”
Do you want to apply for an ERC Proof of Concept grant?
The advice of Xavier Rival
As it is not possible to apply for the ERC Proof of Concept grant without the prerequisite ERC grant, Xavier Rival recommends first and foremost to:“prepare your project well and plan it over the long term. You really need to be ready to do it because it involves very extensive work over five years.”Something which is notably made possible within the Inria research teams: “In fact, the way our research is organised enables us to move towards long-term projects, to take on a subject that is dear to us, to develop it and to feel ready to apply when the time is right.”
Xavier Rival adds a second piece of advice, inspired by his work with Tie Cheng: “once your main research strategy has been defined, don't be scared of making a few detours. Some opportunities that are found slightly 'off-piste' can become very promising later on. There again, a structure like Inria enables us to remain open to these sometimes unexpected avenues.”