[kde-edu]: [Step] Suggestions

[Vladimir]

On Friday 22 February 2008 22:17:14 RalfGesellensetter wrote:
> Am Donnerstag 21 Februar 2008 schrieben Sie:
> > The idea about t=0 as "reset time" is good by we need to somehow
> > educate the users about it. For the near future I'll probably
> > implement something like Simulate/Pause/Stop buttons where Pause will
> > do what Stop now does and Stop will do what Stop+Undo now does. I'll
> > think more about resetting to zero and probably will implement it
> > later.
>
> Hi there, how about an EDIT mode that has to be entered before the user
> can change the settings? If the simulation is paused, only dependent
> objects (like particles) can be moved or added - which is reset hitting
> the reset button. However, if the setting of the esperiment is to be
> changed, the simulation needs to be stopped, only then the EDIT mode
> can be entered.
>
> Could this work out as solution?
Probably, but EDIT mode should be visually distinguishable from normal mode, 
in the other case it can easily lead to accidental mistakes. And how to 
integrate this with undo system ?

Currently I'm thinking about the following solution:
- there are three buttons: Simulate/Pause, Stop and Reset.
- all actions done when simulation is active or paused are not
  recorded to the undo history
- when the user clicks on Stop the whole simulation is
  recorded to the undo history
- when the user clicks on Reset the whole simulation is
  also recorded but is immediately undone (it means that
  Reset equals Stop + Undo)

This way its obvious for the user what Reset does. Moreover, this solution is 
consistent with the fact that currently the user is able to alter experiment 
(move/change/create/delete objects) even when simulation is active, but that 
actions are not undoable: paused mode will be very natural, it will behave 
similarly as altering without pausing. However I'm not entirely happy about 
this solutions because it can lead to complains like "Oh, I've forget to 
press Stop and now I can't undo my changes".

I can also implement undo for actions during simulation (which is non-trivial 
but doable), but this can lead to further confusions: what if the user 
reaches the time before simulation was started by pressing undo ?

>
> about t=0: In some cases (only deterministic models I think) it might be
> interesting to invert time axis and rewind time to t<0.
Yes, simulation back in time is on my TODO list. Actually there is a little 
chance that everything will just work if I set negative time delta for the 
simulator, but I've never tried it.

>
> Regards & thanks for your dedication!
Thanks for you feedback !


-- 
      Best Regards,
        Vladimir