Re: [h5md-user] Monotonicity of the time step

[Pierre de Buyl]

On Mon, Sep 30, 2013 at 11:25:42AM +0200, Felix Höfling wrote:
> Currently, we have the constraint that step and time have to
> increase monotonically. If somebody is interested in simulations
> with reversed time, this can not be satisfied.

A-ouch.

I suggest that those people store their time as increasing and use

actual_time = - time_found_in_file

> I would like to relax the constraint from the time and to require
> only that the "step" increases monotonically. (Which would allow the
> physical time to jump wildly, it could be made monotonic as well,
> but either increasing or decreasing.)

unless there is a strong incentive that I do not see clearly now, I am not very
enthusiastic at this idea.

P