To me it seems I should reserve tags to be used only for processing purposes. I mean, tags are not to be set once and for all. If i use obstacle detection now i can use tags in the process. Later I will tag the same object in different way in different method. So it is not like I am gonna tag this object as player and that is the end. No, I see a tag as a dynamic thing that you can use as needed, for different purposes.
Yeah but the problem is once you tag an object as one thing you can’t tag it else. So you basically have just one tag per object, which is not so useful.
I don’t like using tags; i prefer detect the different kind of objects by the class attached to.
However, as illiterate and Afisicos said, tags “must” be use as “group label”.
Mind my noobieness but use them for defining things like enemies or player or death cubes! When things collide I check if it was an enemy and then BOOM do things.