[sdpd] Structure prediction

[Armel Le Bail <armel... @fluo.univ-lemans.fr>]

Structure prediction is a very attactive subject to me.
Indeed, effective structure prediction would suppress
the need of so many things, including SDPD.

Unfortunately, structure prediction progress quite
slowly, in a manner similar to weather prediction,
maybe. The true structure predictor should only use the
periodic table of elements, like a weather predictor
in a TV buiding, consulting weather reports and trying
some extrapolation.

Locally, it is just fine to look at the window and see
clouds coming for predicting some rain possibility.
The more the predictor is using chemical composition, 
cell parameters, molecules, diffraction data, and the
more he looks at the window, becoming a simple
structure solver, like us, and the less it is easy to
publish in Nature or Science ;-). In spite of a random
starting configuration used in the ESPOIR software,
this is no more structure prediction than applying the
direct methods black box (though ESPOIR is a 
white box, due to its extreme simplicity or rusticity).

I would recommend the reading of "Are Crystal
Structures Predictables ?" by A. Gavezotti, Acc. Chem.
Res. 27 (1994) 309-314.

If someone has references for more recent review papers,
please let me know. That structure prediction topic 
may be the cause of our firing, folks !

Best,

Armel Le Bail - Universite du Maine, Laboratoire des Fluorures,
CNRS ESA 6010, Av. O. Messiaen, 72085 Le Mans Cedex 9, France
http://www.cristal.org/