Re: [h5md-user] fields of observable group

[Felix Höfling]

Am 30.08.2011, 19:09 Uhr, schrieb Pierre de Buyl
<address@hidden>:

> > > I have one objection: it might be of advantage to have some
> > > parameters standardised, e.g., the box size. Otherwise, an H5MD
> > > reader could not retrieve the box size from an H5MD file without the
> > > trajectory group. (And the box volume is certainly an interesting
> > > quantity independent of the trajectory.)
> >
> > That is not an objection, but an endorsement of my proposal ;-).
> >
> > We would definitely not want to mix non-specified (i.e. program
> > dependent) and specified parameters in the same group. This is
> > why it is a good idea to give program-dependent parameters
> > a program-dependent group name from the beginning.
> Ok, then the idea would be to have no parameters group but the  
> specification would suggest that programs creating H5MD files create a  
> "program_name" group for that (or other) purposes.
> An observable with the box size as a function of time for variable size  
> simulations would fit this need, also.
Treating the box size as well as the particle numbers as observables is a
very good idea. This supports both, fixed values as well as values that
change over the course of the simulation. And (for the latter in
particular) it makes semantically sense to consider them as "observables".
But different from the usual observables, these should be mandatory for a
valid H5MD file.

Let me repeat my suggestions from 17 June:

box
   \-- offset [var][d] e.g., (-L_x/2, -L_y/2, -L_z/2)
   \-- edges [var][d][d] e.g., ((L_x, 0, 0), (0, L_y, 0), (0, 0, L_z))
   \-- time [var] e.g., (0)
   \-- step [var] e.g., (0)

d refers to the dimension of space. The examplatory values are for a
cuboid box of fixed size with the origin in its centre.

Felix