Foto: Kurt Mueller, 2005, Click for larger picture Frank's letzte Ruhestätte ist auf dem Friedhof in Marl-Polsum, im Grab seiner Eltern. |
I'm sad to hear that Frank Josellis (born 13. August 1959) passed away on October 5th 2022).
Frank got his mathematics diploma (BS)in Bochum in Germany.
He wrote his thesis on Global periodic orbits for Hamiltonian systems on T
in 1991 under the guidance of Eduard Zehnder at ETH. The co-referee for that thesis had been Jürgen Moser.
Frank Josellis was also mathematically close to Michael Struwe.
As one can see in this acknowledgement of his thesis, he also conferred with mathematicians from the
Forschungsinstitute like Chris Golé, Augustin Banyaga, Fredy Künzle and Maciej Wojtkowski or
other postdocs at ETH like Wilhelm Klingenberg. ^{n} x R^{n}Frank Josellis used topological methods in Lusternik-Schnirelman category in Hamiltonian dynamics. It is a subject, where Andreas Floer (an other student of Edi Zehnder) has made huge progress. Floer's death in 1991 had then of course been a big shock for the entire community. I had been sharing offices with graduate students working in that field and seen the impact which Floer has made. Floer was the tsunami. His new ideas had the effect of a bomb sucking oxygen from other approaches and so also from Ljusternik-Schnirelmann theory. I must say that I'm personally grateful of not having been in that area of mathematics at that time since the Floer methods placed temporarily everything else in the shadow. Morse theory is closely related to LS-Category theory. They both allow to count critical points of a functional and so count periodic solutions. History indicates that since Morse theory has since been much more successful than Category theory. Ljusternik Schnirelmann theory (from 1929) still is a powerful sibling of Morse theory and remains so. One can see this for example in Michael Struwe's work on the Plateau problem. Frank was also a great computer scientist. He was a linux, BSD and computer hardware expert, involved in free software development (I host over more than 20 years 3 smaller projects of him: ipwwdshield, Kal, Moonphase). But this was only the tip of an iceberg. He was familiar with essentially all the open source software and programmed fluently in many different languages. I myself learned a lot of C programming from him. Already more than 20 years ago, Frank preferred AMD CPUs over Intel and built his own specialized machines, comparing benchmarks with others. He also, like me, was a busy Math book collector. As a graduate student or postdoc, I would even photocopy first books in the library, later I started to photograph or buy them, then scan them. Frank's digitizing techniques were much more sophisticated than mine. He automated the process, used tesseract to OCR the books, and edited and improved the text even by hand, where needed and have perfectly small and searchable DJVU documents. As a graduate student, I was grateful to share some years with Frank. We had a small clan of regulars who would eat almost daily together at the ETH Mensa (dining hall): Frank Josellis, Andy Stirnemann, Leila Schneps (a math postdoc), me and my wife Ruth Knill (then a postdoc at the ETH astronomy department). Many times, other graduate students would join like Moritz Adelmeyer, Thomas Wurms or Wilhelm Klingenberg. Discussions were often about mathematics, about how our work on the thesis went, but also about politics (like the first gulf war, which had already then been then a matter of discussion: is it ok for an outside country to intervene into a conflict? I remember that we mostly agreed that it was justified. Today, we might argue differently: it has lowered the threshold to intervene in other affairs with less clear motivations: the second Irak war, Afghanistan, Syria or Libya showed this.) Or then we talked about more philosophical topics like whether it is feasible to have computers verify proofs automatically. Frank, Andy, Leila, Ruth and me had established a little ``culture club" which forced us to get a bit out from all the mathematics and ETH world. All few weeks, somebody had to suggest a cultural event and we would all have to go there. Like to the ``Cats musical", a concert, to the movies, on a trip to the Uetliberg or to eat in a fancy restaurant. Frank was a wonderful person. He was selfless, modest and always willing to help. He was always kind, constantly extremely calm, never appeared to be stressed out, even in difficult times like in 1992, when he was looking for a postdoc position (which he was never offered). He liked already as a graduate student a cigar and so looked so a bit like Norbert Wiener, even so Frank was tall and better looking than Wiener. Frank was not an athlete himself, but a big fan of soccer, saw almost all games in the Bundesliga and English premier league. He was what the English would call a "Gentleman" [the affection of English gentlemen to sports without playing it themselves is portrayed nicely by Charters and Caldicott in Hitchcock's mystery thriller "The Lady Vanishes", where the two characters always talk about cricket]. For sports, Frank liked to hike and bike, at some point even had a mountain bike. The picture to the left, I obtained from Kurt Mueller an old friend of Frank was done in 2005 while hiking (according to the map, this must have been in the Odenthal). Frank was never in a hurry, liked to think carefully about things. He liked more to listen than to talk himself. But what he said was always proofread. Frank's selflessness is maybe best seen from that the fact that he has for many years taken care of his mother in Marl, until to her death 2017 (I know from his letters that he took care already for his mother since 2000 and for example drove her to and from hospitals or rehabilitation clinics). I was told that he lived alone and isolated after his mother's passing. |

- Frank Josellis: Genealogy project
- Global periodic orbits for Hamiltonian systems on T
^{n}x R^{n}Dissertation 1991 [PDF] - Lyusternik-Schnirelman Theory for Flows and periodic orbits for Hamiltonian systems on T
^{n}x R^{n},Proceedings London Math. Soc, 1994[PDF] - Morse Theory for forced Oscillations of Hamiltonian systems on x R
^{n}, J. Diff Equations 1994, [PDF] - With Wilhelm Klingenberg: Structural aspects of Closed Characteristic [PDF] Preprint
- With Oliver Knilll: The Lusternik-Schnirelmann theorem for graphs [PDF] Preprint
- ipfw2shield Software
- Kal calendar Software
- Moon phase Software