commandline arguments in mcstas

[Thomas C Hansen]

Dear Kristian,

In a future version of McStas, could you imagine some additional funcitonality
in the commandline arguments of mcstas simulations? E.g. for a huge number of
parameters, in the actual version the commandline becomes too long, so the only
way is to give all parameters manually, or by a commandline like
'./mysimulation.out -n 1e6 < input.arguments'. That workaround is fine, as long
as I do not want to use gscan which won't work by no means, if there are too
many parameters. I could imagine a commandline like './mysimulation.out -n 1e6
-p input.arguments Lmin=1.2 Lmax=1.4', which means, all input parameters are
taken from a file input.arguments, except the parameters Lmin and Lmax, which
are overwritten by the values given in the commandline. This procedure would
permit the use of a slightly modified version of gscan (taking account of the
option '-p' or '--param' of the mcstas generated simulation). The file
input.arguments might contain as many 'param=value', separated by linefeed,
blank or comma in any order. If any 'param=value' is missing in the file
input.arguments, as well as in the commandline itself, the up to now used
convention may be used, which is asking interactively the user for some input.
(That's in fact another, new, suggestion, as in the actual version the
simulation stops if some parameters are given, others not. Why not asking
interactively these missing parameters instead?)

These are just suggestions, in order to ripe some new ideas, don't take it as a
strong demand right now!

Best Regards,

Thomas


--
Dr. Thomas C HANSEN, Instrument Scientist D20, Diffraction Group
Institut Max von Laue-Paul Langevin, BP 156, 38042 Grenoble Cédex 9, France
ILL4-1st floor room 100, Phone +33(0)476207044, Fax +33(0)476207648
mailto:hansen at ill.fr, http://www.ill.fr/YellowBook/D20
Private: 13, allée des étangs, 38240 Meylan, France, +33(0)476180553