Re: [h5md-user] Variable-size particle groups

[Peter Colberg]

On Tue, May 29, 2012 at 06:48:49AM -0400, Peter Colberg wrote:
> On Tue, May 29, 2012 at 11:38:22AM +0200, Olaf Lenz wrote:
> > I just remembered another issue that comes up with varying particle
> > numbers in a simulation: the particle identity. When a particle
> > disappears from the simulation, are all other particle data afterwards
> > shifted one down? In particular for visualization, this information is
> > crucial.
> > 
> > Therefore, in a simulation with varying particle number, it is probably
> > necessary to add some way to identify a particle in the array, e.g. by
> > adding another subgroup "particle_id[variable x N]" to the trajectory
> > group.
> 
> Yes, then we should also specify a “particle_id” subgroup in H5MD
> along with the optional “range” dataset for the subgroups.

Maybe “id” is a better name, shorter and more general?

Or my inital proposal of “tag”.

Trajectory groups are not necessarily restricted to particles, a
trajectory group could describe arbitrary objects with positions,
velocities, …

Peter