Re: [h5md-user] Metadata on physical quantities and their units

[Pierre de Buyl]

Peter Colberg <address@hidden> a écrit :

> On Thu, Aug 01, 2013 at 10:52:19PM -0400, Peter Colberg wrote:
$>> I have prepared a proposal that adds an H5MD metadata group for
> > physical quantities and their units. Please read the proposal
> > carefully, to avoid any inconsistencies to slip through.
>
> Our current discussion of units is quite abstract, so I thought it
> would be nice to explore a few concrete examples. I revised the above
> proposal to illustrate a real-world scenario, a system of SI units as
> available in the LAMMPS package [1]. The attachment contains a Python
> script that produces such a “quantities” group.
>
> [1] http://lammps.sandia.gov/doc/units.html

Hi Peter,

I'm sorry I didn't reply to the original message of this thread.

One possibility, with H5MD, is the ability to have subgroups. Imagine
combining two molecular structures that come from separate sources.
With a global unit settings for units, as opposed to a named datatype,
this would be error-prone.

Given the lack of actual use, as of now and in the context of H5MD,
could we not wait?

Pierre