Fonenantsoa Maurica

2019: Optimization engineer, Eurodecision
2018: Postdoctoral researcher in Software Verification, French Commission for Atomic Energy
2017: PhD in Software Verification, University of Reunion Island
2014: Engineer's degree in Systems and Networking, ESIROI

Contact me
Page last modified on January 7, 2019


Defense of my PhD thesis: termination analysis of floating-point computations. This is how I introduced the matter: ''Mark spent three hours taking a selfie. Now he wants to post it on Facebook. In order to get as much Likes as possible, he decides to edit it. He chooses to use ImageMagick which is a command-line tool for image processing. He tries to modify the hue of the image and then, disaster! ImageMagick stops responding. The culprit is a software bug that I discovered: the computation of the new floating-point value of the hue may fail to terminate. To the fictional character Mark, that bug is actually not very harmful. His Facebook fans just need to wait a little bit longer until he manages to retouch his selfie. However, there were cases in history where termination bugs and incorrect floating-point computations lead to dramatic events. On one hand, a non-expected infinite loop caused the Microsoft Azure Storage Service to be interrupted for more than ten hours in 2014. On the other hand, a faulty implementation of the floating-point division operation ultimately cost $475 million to Intel in 1994.''
I was invited by UPV to talk about Software Verification, a discipline whose goal is to assure that programs are bug free. I took this opportunity to present two code analyzers: Frama-C for C, on which I have already worked, and Julia for Java. They are great tools: they can guarantee absolute invulnerability to, say, runtime errors or code injection. Do not hesitate to try them!
Believe it or not: those are high school teenagers who voluntarily participated to a summer school on mathematics! #mathc2+
This is a school in Madagascar, my country. Most Malagasy kids attend classes in horrible conditions. My parents were of such kids. Yet, they persevered and managed to escape poverty. That story is the reason why I value education so much. If you want to help, the Akamasoa association might be the one who has done the most for educating homeless people in Madagascar.
Before entering university, I wanted to be a pianist. Fortunately, I ended up doing computer science: I would have been a terrible musician! Still, I keep playing from time to time as it comforts me during those days when science completely defeats me. I even had the pleasure to be invited to talk about my view on baroque music at RLI.