For context see Problems with the ANR Bigben project.
The presidency of the french Agence Nationale de la Recherche kindly answered to my demand of reexamination of the awarded ANR Bigben project. [added: … after several previous interesting mail exchanges with higher and higher ANR representatives; one argument impressed me a lot, perhaps it deserves a full discussion because after all we want Open Science to win.]
Here are the two parts of the answer: avis (pdf) and letter (pdf).
Here is my reply to ANR answer (links added and [text added between brackets here]):
Thank you for the precise response and for the time spent by ANR concerning this subject. For the scientific part there is an article in preparation.
Here are some short remarks.
- There are no “generalized bipotentials”. The name is invented by the project leader to fit with his competences. The source of the theoretical foundation which gives the name to the project is arXiv:1902.04598, [On the information content of the difference from hamiltonian evolution] where in proposition 1.3 is explained the appearance of what the project leader now calls “generalized bipotentials”. This is not referenced in the project.
- For the experts: a bipotential is always relative to the duality chosen. What matters is the difference between the bipotential and the duality. Change the duality and you obtain a change of bipotential, by a substraction of the old duality and an addition of the new duality. Is this a theoretical advance towards “generalized bipotentials”?
- The subject of my hamiltonian inclusions is old (2008) [Hamiltonian inclusions with convex dissipation]. It was turned into “symplectic BEN” during a collaboration with the project leader, where the connection with Brezis-Ekeland and Nayroles principles was made, by neglecting the inertial terms. But the reduction of hamiltonian inclusions, or SBEN if you like, to BEN is misleading. Back in 2014, when Djimedo Kondo, member of the bigben team, was introduced to the subject, [slides of d’Alembert seminary from 2014], [figshare], he immediately remarked that the reduction of SBEN to BEN has problems. Indeed, what the leader de Saxce claims is that all reduces to the minimization of a cost functional over all evolution curves. But the cost, as easily remarked by Kondo, is infinite for most of the trajectories, unless one already satisfies the dynamic equations. By neglecting the inertial terms, one does not solve the problem, because the same kind of infinities force to consider the satisfiability of the trajectory (as in contact or some plasticity or damage problems). Even since 2014 it was clear that the “SBEN is BEN” idea of de Saxce, cannot work in practice, except for some trivial examples. In this project the leader de Saxce wants to pursue the same. I claim that other ideas are needed (some of them I have, but am I willing to make the public again, without attribution?).
- I am of course willing to see my work being developed and moreover enhanced by meaningful collaboration. It does not seem the case until now. There is also the ethical aspect which I refrained to mention because the scientific case is almost enough. That is why I refused the “visio” meetings recently, because previous ones, where I protested against “SBEN is BEN” or where I was assured about the details of collaboration, amounted to nothing. I am now accused, in messages leaked by inadvertence by de Saxce, to have strange financial demands, when in reality I asked for a schedule and details just like the other members of the project. This is a lack of collegiality which I take very seriously.
- Therefore, I take as very positive the interest into hamiltonian inclusion, or SBEN, or by any other name. It is also positive that some work on arXiv (not any work, only the useful one) is appropriated, thus recognizing the value.
- I take as negative the lack of attribution. I challenged my career on these Open Science ideas, do I need to see my work being used without proper attribution? But I admit though that from the point of view of ANR, or any other management organization, it would be risky to accept as valuable any arXiv, say, “preprint” (and probably it would mean the death of arXiv itself, due to low quality submissions). But in this case, clearly this work is valuable.
With best regards to all the members of this discussion,
After the response, a final comment: it would be nice if the reputed ANR takes a step towards acknowledging more Open Science, just like more than 100 years ago the french society accepted impressionism in art.
The history of that art movement is an inspiration since a long time, see Boring mathematics, artistes pompiers and impressionists.
2 thoughts on “Answer from ANR concerning the ANR Bigben project”