The level design is great! The mechanics were introduced properly. And the harder levels are hard enough that you have to think and not just fiddle around. Solving the levels felt very awarding.
Other notes:
It would be nice if pitch move worked in space.
Sometimes I felt like I was teleported around a few nodes. Haven't found out if this a subgame or engine bug, and that's feaking me out.
Some of the effects, like the fire pushing you away, the feedback you get when you're in the way of a klotz, the bouncy walls, or the portal are very satisfying!
(If low latency is important for a better experience, you could also recommend decreasing dedicated_server_step to something like 0.016. This influences how frequent the server sends packets (also in singleplayer).)
I experimented with 3 different forms of movement control (minetest-style pitch-move, not-minetest-style pitch-move, and non-pitch-move) and settled on the one I used because (1) it seems like players found all 3 about equally unintuitive, and (2) MT doesn't (at least at the time) provide a way to find out what the state of the client-side toggle is. If I could detect it, then I could support the toggle automatically and transparently.
Sometimes I felt like I was teleported around a few nodes
I've seen some movement jank when a lag spike hits. I thought it could be related to the Klots movement code, but I've seen some of the same jank when using standard MT movement, and even in other games. I think mostly it's just easier to observe in Klots becauase the movement is novel so you're noticing it more.
dedicated_server_step
Unfortunately, this ALSO affects the animation timing for object:move_to(pos), which affects how smoothly Klots appear to move when you push them; shortening it tends to shorten the animation to the point where you lose the sense of motion. Would be nice to have these things decoupled so we could get the best of all worlds.
I want more levels
Me too! I've got a few ideas for level designs, but if you have some yourself, there's actually a fairly robust level editor built into the game, and a level editing guide that explains how to set it up, use it, and ways to submit levels if you want them included upstream.
If I could detect it, then I could support the toggle automatically and transparently.
Actually, in singleplayer you can just read the pitch_move setting. But I admit this is ugly.
movement jank
I wasn't entirely sure if you're attached to an object. But if you are, such jank is at least not new. x)
ALSO affects the animation timing for object:move_to(pos)
Ah I didn't know that I think. It looks like the server could send a separate update interval for each move_to actually. I've tried using a larger update interval than send inerval, and it works without major issues, so decoupling might work. Added a bullet to my large personal TODO list, maybe I'll make a PR or issue in a distance future if nobody else does.
Finally played it.
The level design is great! The mechanics were introduced properly. And the harder levels are hard enough that you have to think and not just fiddle around. Solving the levels felt very awarding.
Other notes:
dedicated_server_stepto something like0.016. This influences how frequent the server sends packets (also in singleplayer).)I experimented with 3 different forms of movement control (minetest-style pitch-move, not-minetest-style pitch-move, and non-pitch-move) and settled on the one I used because (1) it seems like players found all 3 about equally unintuitive, and (2) MT doesn't (at least at the time) provide a way to find out what the state of the client-side toggle is. If I could detect it, then I could support the toggle automatically and transparently.
I've seen some movement jank when a lag spike hits. I thought it could be related to the Klots movement code, but I've seen some of the same jank when using standard MT movement, and even in other games. I think mostly it's just easier to observe in Klots becauase the movement is novel so you're noticing it more.
Unfortunately, this ALSO affects the animation timing for object:move_to(pos), which affects how smoothly Klots appear to move when you push them; shortening it tends to shorten the animation to the point where you lose the sense of motion. Would be nice to have these things decoupled so we could get the best of all worlds.
Me too! I've got a few ideas for level designs, but if you have some yourself, there's actually a fairly robust level editor built into the game, and a level editing guide that explains how to set it up, use it, and ways to submit levels if you want them included upstream.
Actually, in singleplayer you can just read the
pitch_movesetting. But I admit this is ugly.I wasn't entirely sure if you're attached to an object. But if you are, such jank is at least not new. x)
Ah I didn't know that I think. It looks like the server could send a separate update interval for each
move_toactually. I've tried using a larger update interval than send inerval, and it works without major issues, so decoupling might work. Added a bullet to my large personal TODO list, maybe I'll make a PR or issue in a distance future if nobody else does.