next »

[io3 creations]

And of course, how can you argue with:

"Unity is the ultimate game development platform. Use Unity to build high-quality 3D and 2D games, deploy them across mobile, desktop, VR/AR, consoles or the Web, and connect with loyal and enthusiastic players and customers."

Well, they still haven't fixed the most basic of VSync bugs, leading to maximising CPU usage on all Linux machines even when a game idles.

Yeah, I know.  I've come across a few that have been reported years ago ... yet still no progress.  One of those is having shadows showing up on objects on layers that should be ignored.  That is affecting my "portal" views.  I wondered if perhaps it was something in my code, but looking at a few-years-old portal demo, the shadows show up there as well.
As with any organization, they probably have their "priorities".

[JobLeonard]

At first, based on people's and cars' movements and pixellation, I thought that it was animated.  But the the longer I watched, I wondered if there was a "simpler" way.

Is there a diagram or some kind of visual that illustrates the process?

So I somehow read over this so I forgot to answer this, but here is an early upload:

https://imgur.com/a/xqfif

I'll write the explanatory text when I get back from work

[io3 creations]

Sometimes I find that when I click the *NEW* button, either an earlier or later post is displayed instead of the actual post.  Maybe something similar happened?


Oh, I see!  I was thinking that the T,Y plane actually meant slices parallel to the T,Y plane.  It makes sense now.

The  lava-lamp-people are so mesmerizing to watch.

[diegzumillo]

Made a cool gif:



There's really no progress to show here, despite looking like it's now "on curved space". I mean, I guess it technically is but all I did from then to now is write a function on curved space that draws lines using its own metric. Just trust me on this, I made no progress. But it is a cool gif!

The current challenge is getting this physics simulation to incorporate my curved space system. And the challenge in that is the transition. Because local position will be something between 0 and 1 (for x and y) so it the current patch of space is like a second coordinate system on top of it all. Anyway, I'm just venting really. Hopefully that makes a bit of sense. Very hard to find time to work on this lately.

[JobLeonard]

So you're saying it's just projecting a regular grid, with regular gravity, to curved space?

That's still good enough for fooling players though, no?

[diegzumillo]

I wouldn't say it's fooling the player, that's pretty much what it is. At the end of the day all the matters is the topology of the level; how everything is connected. And scale. Of course, I have plans of making the local geometry have a role here and there but it would not be a natural consequence of it, if that makes sense.

An example is my old implementation of curved space with those curves that simply rotate the player, where momentum was neglecting the curve completely. I took advantage of that on a couple of levels of the demo; you jump on a mushroom spring, move through a curved region and hop over some spikes. But you can only go high enough if you climb through the 'compressed' side of the curve because it's 'shorter'. In actual curved space your momentum would also be shorter though, so that wouldn't work now by default.

Any thoughts on this? Was that momentum violation a fun thing? watching people play that demo and get stuck on that part or beat it by trial and error and have no idea how they got it makes me think otherwise. Then again the mechanic could be there, just used with caution.

I think topology and scale are good enough but it could be fun if local geometry played more than a cosmetic role. Another way of doing that besides the momentum thing would be having special objects that ignore the curve in some way.

[io3 creations]

Hmm... not sure how I missed the demo link in the first post.  Maybe, I was just skimming ...

Unfortunately, I didn't get very far.  After the first checkpoint tree, I kept dying at the spikes (right before the first wormhole).  I wondered if I missed something but based on your video at 0:28



it seems like that the character always jumps too high and hits the spikes.  I was using the keyboard.  I did notice that if I press the spacebar longer then the character jumps a bit higher but couldn't find a way to jump low enough.


As a first intuitive impression, I agree that using topology and scale only is enough.  But the other aspects could also be interesting if introduced well and result in fun gameplay.

[diegzumillo]

Ah yes, that old outdated demo. Full of bugs and idiotic decisions. That part on the first level is super dumb. I'm pretty sure there's no bug, you just have to tap the jump ever so slightly. And that's really dumb because the demo never uses that again, it's teaching you how to do something you won't need again. My main incentive in getting a new demo out is so I don't have to show that to no one else ever again  But right now that's better than nothing.

[io3 creations]

Ah yes, that old outdated demo. Full of bugs and idiotic decisions. That part on the first level is super dumb. I'm pretty sure there's no bug, you just have to tap the jump ever so slightly. And that's really dumb because the demo never uses that again, it's teaching you how to do something you won't need again.

Yeah, that's exactly what I expected since I noticed that there was a difference when I held the spacebar longer.  I did restart the demo and that didn't help either.  However, when I tried the demo after reading your post, it worked fine.   

My main incentive in getting a new demo out is so I don't have to show that to no one else ever again  But right now that's better than nothing.

That's a good motivation!   Looking forward to it!

[io3 creations]

Looks like the Antigrid devlog is heading in a similar direction.

[danovo]

Of course, my very precious idea not only already existed, but it's more polished, better coded and has an actual direction.
Really superb work you're doing here. I just started with the idea and reading your devlog it seems I'm kind of following your path (right now my maps are a list of quads which store pointers to other quads).
What a bittersweet sensation I'm left with after knowing of this project  If you ever want to share ideas about curved space, drop me a line

[diegzumillo]

Of course, my very precious idea not only already existed, but it's more polished, better coded and has an actual direction.
Really superb work you're doing here. I just started with the idea and reading your devlog it seems I'm kind of following your path (right now my maps are a list of quads which store pointers to other quads).
What a bittersweet sensation I'm left with after knowing of this project   If you ever want to share ideas about curved space, drop me a line

I know that feeling too well. It's very easy to get attached and possessive about your ideas. Everyone who followed this devlog will remember a few episodes where I thought my idea wasn't original anymore and it made me irrationally demotivated. Deep down I know it's not what matters but emotions are what they are.

But hey, curved space is a general enough idea that it can lead to drastically different games. You could have a contra-like game or a mario-like game in curved spaces, and we know how different these two would be.

I hope you can learn from my mistakes and avoid wasting as much time as I did haha. It seems like you're doing what I call the second iteration here, with individual quads having a linear relation to a square. It was working but now I settled on large bezier surface quads. And you could try an idea that I never tried implementing of having the game happen in separate square regions with dedicated cameras, then get these cameras to textures on a curved mesh. The connection between region is the tricky part.

Good luck on your project!

[io3 creations]

Of course, my very precious idea not only already existed, but it's more polished, better coded and has an actual direction.
Really superb work you're doing here. I just started with the idea and reading your devlog it seems I'm kind of following your path (right now my maps are a list of quads which store pointers to other quads).
What a bittersweet sensation I'm left with after knowing of this project   If you ever want to share ideas about curved space, drop me a line

I know that feeling too well. It's very easy to get attached and possessive about your ideas. Everyone who followed this devlog will remember a few episodes where I thought my idea wasn't original anymore and it made me irrationally demotivated. Deep down I know it's not what matters but emotions are what they are.

Well, there's always the "First to finish" i.e. release.   

[diegzumillo]

Haha That's true. But I can't compete against anyone with a little more time on their hands.

The good news is I'm done with this grad program this year! Not sure if summer or later. That means I'll have a lot more time to work on this. That also means that this year will also be the least productive of all for this project. I have to dedicate myself completely to the program. And I don't have anything yet. No results, no papers, no dissertation. Somehow all of that has to happen now. If this miracle happens I shall answer only by DOCTOR DIEGZUMILLO MCQUANLOOPSIE III.

[io3 creations]

The good news is I'm done with this grad program this year! Not sure if summer or later. That means I'll have a lot more time to work on this. That also means that this year will also be the least productive of all for this project. I have to dedicate myself completely to the program. And I don't have anything yet. No results, no papers, no dissertation. Somehow all of that has to happen now. If this miracle happens I shall answer only by DOCTOR DIEGZUMILLO MCQUANLOOPSIE III.

Best of luck!

[diegzumillo]

Did some work on the physics implementation. It's messy but seems to be coming along. Intersections are quite tricky! Normal intersections work nicely but if there's a change in scale there's no simple way to achieve a smooth transition. I suspect the best I can get is to make it less jumpy and weird. Maybe I should consider splitting the physics body into two, one for each region and blend them as seamlessly as possible. I'll try that once I mess with this transition scheme for a little longer.

Here's the current state of transition of rigidbodies on normal intersection and scale varying intersections:

[diegzumillo]

OK, this is glitchy as hell but it's all good progress on the physics engine. Most of the weird lines all over the place are just my lazy drawline code that gets a little lost when the object is transitioning between two regions and isn't worth the trouble of fixing, seeing how it's only for debugging.

But there are some pretty ugly collisions there that I need to figure out, specially at corners of walls. I don't know yet if it's caused by the physics code or the region transition code, maybe there's some mistake in my function that calculates distances incorrectly at the transition and it's only noticeable during collisions.

This verlet algorithm is a little limited. Collisions don't cause torque, so sometimes you see a triangle on top of another without falling over. I don't like that very much. I'll see if I can live with that or try to implement torque by hand or switch to a different algorithm altogether.

[diegzumillo]

Turns out my sunday wasn't over yet! Seems like I squashed the main bugs. Apart from a few blocks here and there freaking out and leaking out of the world and some collisions that should have happened, it seems to be working pretty well now. I even fixed the debug line rendering.

This gif uses the same configuration as the previous post but now it's easier to tell what the hell is going on. They all start falling in the same direction but they fall through the left and right holes and meet in the middle going opposite ways.



Performance is absolutely abysmal at the moment but that was expected. While trying to get it to work I intentionally did some extremely dumb and slow coding. Now that the logic is in place I can clean it up and rewrite to make it a little better. But it's not so bad. That's a job for next weekend diegzumillo though... Actually, I might have to do a seminar the following monday so next weekend is probably not gonna work. Anyway, I might procrastinate midweek working on this.

[oahda]

i like the inky, scribbly notebook look of these debug lines

[diegzumillo]

i like the inky, scribbly notebook look of these debug lines

Me too! It was completely by accident, I was just trying to make it easier to see and suddenly I had notebook lines and blue pen drawings