Kratos: [Output] output_frequency

[GuillermoCasas]

The time that passes between successive outputs is held in the json variable 'output_frequency'. I find the name very unfortunate, since it clearly is not used as a frequency (units 1/time) but as a period (units of time). I suspect the difficulty of finding a suitable term guided this final choice. But I still think we could do better.
I do think it is much more intuitive to fix a time interval than a frequency (which would require calculations in most cases) or a number of time steps (one could fix the number of time steps in between prints).
So I propose to discuss possible alternatives (@KratosMultiphysics/implementation-committee ). Here are a few:

• output_period
• output_process_period
• output_time_interval
• time_between_outputs
• time_between_consecutive_outputs
• time_interval_between_consecutive_outputs
• time_interval_between_outputs
  Or, my favourite so far:
• output_time_period

[philbucher]

related to #3766

[josep-m-carbonell]

If the time discretization is given by the variable "time_step", maybe a good name for printing the output is "output_time_step"

[AlejandroCornejo]

If the time discretization is given by the variable "time_step", maybe a good name for printing the output is "output_time_step"

This name I think that it is confusing if your "output_control_type" is "step" for example... Maybe output_period? IDK

[josep-m-carbonell]

As confusing as "time_step" is. At least is consistent.

[GuillermoCasas]

@josep-m-carbonell Yes, I do not know why it did not come to mind. output_time_step sounds very good to me and it is consistent in the way you pointed out.

[GuillermoCasas]

The @KratosMultiphysics/implementation-committee has come to the following decision concerning this issue:
The name output_frequency is definitely not appropriate and it should be changed. There are two possibilities and we could not reach a definitive agreement on which one to follow.

1. Change the name to output_interval so that it is neutral to either integers (number of time steps) or doubles (time interval).
2. Create two different variables, so only one of them is used depending on the control parameter (step or time). Their names would be output_time_interval and output_every_n_steps. This has the advantage of making it clear what the variable means from its name, although it would require a lot more re-coding.

We pass the issue on to the @KratosMultiphysics/technical-committee to decide.

[philbucher]

I think I would go for the first option

[pooyan-dadvand]

Thanks @KratosMultiphysics/implementation-committee for their clarification. Any opinion from @KratosMultiphysics/interface-committee?

[dbaumgaertner]

Speaking for @KratosMultiphysics/interface-committee, I would also vote for the first option above. The reasons is its increased flexibility/usability while reducing overhead implementations through the distinction of two variables which area technically standing for the same thing.

[pooyan-dadvand]

@KratosMultiphysics/technical-committee agrees with the decision of the @KratosMultiphysics/implementation-committee and @KratosMultiphysics/interface-committee

Thanks all!

[RiccardoRossi]

this is done, at least concerning the @KratosMultiphysics/technical-committee

now someone needs to step up to implement it

[roigcarlo]

Ping @KratosMultiphysics/implementation-committee Is there anyone doing this?

[AlejandroCornejo]

It is assigned to @GuillermoCasas but he has to find time to implement it I think :)