In Memoriam France Allard
With very great sadness I have learned about the passing of my long time
friend and collaborator France Allard.
I have known France since about 1986 when we were both students at the
ITA in Heidelberg. She had just moved from Montreal to start as a
graduate student and we ended up with the same advisor for our
dissertations (Rainer Wehrse). With her thesis on atmosphere models of M
dwarfs she was pushing the state of the art a decade or two ahead. At
the time we collaborated loosely, around 1993 our collaboration
intensified with her suggestion to merge our, at the time, separate
codes to expand modeling capabilities. Here is the relevant comment in
the PHOENIX changelog:
* 4.0.0: exchanged the old EOS with the molecular EOS of
* France Allard. Added Brent-style solver for pure ion
* eos, added molecules consistently. (phh, 21/jul/93)
From there on, we (together with Eddie Baron) collaborated closely and,
I dare say, very successfully on many many PHOENIX versions and
generated a vast number of models. We spent countless hours and days
discussing new ideas (hers were often brilliant and almost always good),
code problems and possibilities, fixing innumerable bugs etc etc. I
remember intense discussion sessions at the CRAL in the late 90's and
early 2000's when I (and later also Travis Barman) was visiting her over
the summer. Around 2005 our scientific interests drifted apart a bit and
France continued to work on her fork of PHOENIX, constantly expanding
A few closing random thoughts:
- Before starting her work in Heidelberg, France took a six week crash
course in German at the Goethe Institute in Freiburg (if memory serves
me correctly), with no prior knowledge of the language. I could not
believe how well she spoke and understood German when she came
- We had a number of rather epic scientific fights. She was always right.
Her Fortran code was horrid, though.
The PHOENIX comments in French are all from France. I have no idea what
they say (this was before Google Translate). I suspect that was in part
intentional, whereas France of course could read all my comments, in
English or German.
She cared very much about observations and trying to reproduce them
with models and synthetic spectra. She was very annoyed when I did not
seem to worry about observations as much as she did.
28.2% of 'my' refereed papers are co-authored by France.
France was very competitive and intense, but never unfriendly to the
When following a new idea, France was immensely focussed, sometimes totally forgetting
everything around her. Including colloquium talks scheduled long in advance, buying food
for the weekend and other such trivialities.
My visits to France at the CRAL often went like this: In the first days
France explained a new idea/concept to me (getting upset when I was too
slow to get it), which obviously needed to be added to PHOENIX like two days
ago. After some initial work, it turns out that a simple implementation
was either horribly complex and/or incredibly slow (e.g., requiring a
few micro-Hubble-times in core-hours to compute, nowadays that is not that big
of a deal, but in the late 90's it was bad news). This meant a major
redesign, which sometimes was mostly finished at CRAL, sometimes
requiring a dedicated project (quite a few dissertations started like this).
I thought that France's talks were too much data and too little theory,
I am pretty sure that she thought that my talks were too much theory and
not enough data (if any). Guess who got (rightfully!) invited to give
presentations at conferences. Which was also a good thing because:
France loved going to conferences, give presentations and travel all
over the planet (I am pretty sure she would have travelled off-planet
for a conference on the ISS or the Moon). She often seemed to have
complicated travel arrangements that she changed/fixed on the fly.
From time to time France sent me emails asking about the reason for a
crash or error message from a PHOENIX run. More often than not she did
not include enough information to guess an answer (I suspect that was
intentional). In recent years that was her way to touch bases, I think.
did this to great effect to get us to think about extending PHOENIX
capabilities, i.e., by running it outside the intended parameter ranges
and watching it crash or produce nonsense (a few times it even worked on
the first attempt). That's one reason why there are no safety checks in
PHOENIX, that would stifle creative new uses (but makes it close to
impossible for non-experts to use it correctly).
Astronomy lost a great researcher and role model, far too early.
I will always remember.
This page was last modified by PHH on