Castle Paradox

Bagne · Posted: Sun Dec 06, 2009 6:11 am Post subject:

Sure, if you like.
_________________
Working on rain and cloud formation

Bagne · Posted: Sun Dec 06, 2009 10:27 am Post subject:

Is the high speed caterpillar stretching desirable?
I'm guessing not - I also can't think of any uses for it.
I don't know of any games that exploit it.

I have some thoughts on caterpillars:

If a new kind of caterpillar (based on the one you just explained) was done in the following way, it would prevent stretching/squishing, while also paving the way for non tile-based movement.

I am assuming that the magnitude of the minimum movement speed is 1 pixel per tick.

1) We have a history of leader positions (size 40 ... er ... 41).
2) The leader's position is stored at slot 0.
3) We have a pre-history for each tailing hero (size ... 20 I think).
4) The current position of a tailing hero is found in the last entry of their pre-history
5) Every time movement occurs, the leader's history entries are shifted down by n slots, where "n" is the leader's movement speed in [pixels/tick]. This will create empty entries between slot 0 and slot n - so they will have to be interpolated.
6) Positions stored at indices 0, 20 and 40 will feed into each tailing hero's pre-history (at a rate of n entries/ tick).
7) Whenever "resume caterpillar" is used the tailing hero pre-histories are re-interpolated (between current positions, and positions stored in slots 0, 20, and 40)

Note that:

This lets us use any integer speed, so long as it isn't smaller than 1.

Point 4) will guarantee that tailing heros are always 20 pixels away. If you're using a speed that isn't divisible into 20, rounding errors in the interpolation may misplace the tailing heros by 1 pixel, but I suspect that if you walk continuously in the same direction, the offset will be consistent, so the heros won't jiggle.
For tile-based movement, we're restricted to speeds of 1,2,4,5,10,20,40 etc anyway, so rounding error isn't a concern (we don't need to worry about misaligning heros to the tiles).

Also, "resume caterpillar" interpolation would have to be slightly different for tile based and non-tile based movement - the difference being that in tile based movement, we don't ever want to risk misalign a hero with a tile, and so they need to be moved to their proper place in the time of one step.

With non-tile based movement, we're free to interpolate over however many ticks we please 4 or 5 is good.

If we want to be over-the-top extravagant and make the caterpillar length customizeable, really large histories (and pre-histories) could be kept, but the feed-in indicies and hero positions indicies would change.
For example, to make the caterpillar twice as long, the history feeds into the pre-histories at indicies 0,40,and 80, and heros are positioned at entry 40 of the pre-history.

Yeah, that's it.

Okay,
In the meantime, I've thought of a use of the stretch-contract effect.
The team is walking really fast, when the leader slows abrubtly, and the queue contracts. Everyone has bumped into one another a-la-Jungle-Book-Elephants. "Hey! Watch it!", barks the party leader.
_________________
Working on rain and cloud formation