Re: [h5md-user] proposal 100 - update

[Pierre de Buyl]

Hi,

On Thu, May 28, 2015 at 06:19:51PM -0400, Peter Colberg wrote:
> On Wed, May 27, 2015 at 11:17:43AM +0200, Pierre de Buyl wrote:
> > 2. I propose to extend it by allowing
> > 
> > /step                    Dataset {4/Inf}
> > /time                    Dataset {SCALAR}
> > /value                   Dataset {4/Inf, 32, 3}
> 
> What does 4/ mean?
> 
> Both step and time should be scalar datasets, no?

Ok, I should have been explicit. This was the output of h5ls. The idea is that
you check "step" first. If step is scalar, then the storage is linear with
step[i] = i*step

and if there is a scalar "time", time[i] = i*time. Both come with optional
"offset" attributes.

But, in another likely usage, imagine that step is not linear and that
time[i]=step[i]*DT for all values of i.

My "new" extension is that if you have a non-scalar dataset for "step" and a
scalar dataset for "time", you can compute the value in time as
time[i]=step[i]*time

> > Then, time[i] = step[i]*time (I hope this is self-explaining, if not let me
> > know).
> 
> I suggest keeping step and time independent:
> 
> step[i] = step_offset + i*step_interval
> time[i] = time_offset + i*time_interval

Hopefully, my scheme above avoids ambiguities.

Formatedly,

Pierre