garykl / human-noise

0 stars 0 forks source link

saving data format #12

Open garykl opened 9 years ago

garykl commented 9 years ago

I would handle the data saving part in the following way. A reasonable time interval is chosen: this will be about the time, an agent needs to go half through its environment of perception. (cutoff radius). For each such time step, a new file is created, that contains N lines, where N is the number of agents, and four columns. The first column is the id that is saved by the server, column two and three are the position in cartesian coordinates, and the fourth column contains the orientation angle. Alternativly, the coordinates of the velocity may be saved, if we want to have different velocities, later. Actually, I'd like to have different velocities. There may be some predators for example.

protection2501 commented 9 years ago

sounds good in principle, just one thing. do you think creating so many files is the best way to handle the data? is that better than a few big files? specially since opening and closing fiels takes some lag-time... I honestly don't know, just asking...

ManfredLevi commented 9 years ago

Where should we save the data? I think it may not be such a good idea to save all the data on an open access forum?

garykl commented 9 years ago

@protection2501: You may be right. The reason for this is, that the number of players may change during the game/simulation. One could also have one file per player, but then, we would also need to save the time step. @ManfredLevi: yes, we shouldn't save it on github. But I don't really know yet how to handle that gracefully.

Manfred commented 9 years ago

@garykl Wrong Manfred