next »

[SolarLune]

@Geti - Pretty good! I like how you can change how fast the sand falls. I'm guessing you're going to optimize the sand to only fall when on-screen to help make things run faster right? I'm not sure, but could you split the sand blocks up into individual squares and move those, rather than large planks of them? Anyway, nice sand physics!

[Sakar]

Made this

Thinking of making it some sort of weird exploration type game thing.

[namragog]

Made this

Thinking of making it some sort of weird exploration type game thing.

You NEED to make this. I was so excited when I fund out the floaty things were solid!!!

[peanutbuttershoes]

I fell off a floaty thing. I don't know if I stopped falling and hit an invisible floor or if I just kept falling. Twas hard to tell.

Cool stuff.

[herror]

Made this

Thinking of making it some sort of weird exploration type game thing.

It was so weird and so nice. I enjoyed it a lot.

[SolarLune]

Definitely weird, but it definitely also was interesting to look at and seemed worth investing time into. You should make something with that...

[Sigz]

Today I implemented mouse support for the main menu, full screen toggle, and rough controls for the paranormal mystery/adventure game I recently started working on. Not much, but I'm glad I got something done despite my house being flooded with visiting relatives.

Continuing on this (and now that the family is gone), I implemented the animation system, camera system (so the view follows the player around on levels larger than the level), and worked up a new concept for the character design:

Original mockup:





New mockup:





Still tons of work to be done on every aspect, but it feels good to be making little bits of progress each day :D

[Geti]

@Geti - Pretty good! I like how you can change how fast the sand falls. I'm guessing you're going to optimize the sand to only fall when on-screen to help make things run faster right? I'm not sure, but could you split the sand blocks up into individual squares and move those, rather than large planks of them? Anyway, nice sand physics!

The reason it's in "planks" is that it's kind of faster to do things like that with tiles than with discrete pieces unless you've got some sort of parallel processing, which I don't in flash. In KAG, it'll be optimised (algorithm-side and code-side) and implemented in C++, so it should be considerably faster regardless. There's no auto-tiling at the moment because I'm lazy, but it should really have angled smoothing tiles in as well to look nicer.

The thing with only running when on-screen is that in KAG there's many a player in-game. Perhaps we can distribute the CPU-load there, somehow; it'd be more costly net-wise but much faster than having one machine (or worse - each machine) slogging out the sim. I guess the server could run the fluid sims itself if its got enough grunt, it'd have to send a buttload of tile updates though...

Either way, glad you like it.

EDIT: derp, took some debug logging statements out and now it runs ~30 times as fast. goddamn string concatenation, eh?

[SolarLune]

30 times as fast, huh? If you wanted to implement 4x4 blocks instead of 16x4 'planks':

16x4 planks = baseline (30x faster at this point)
4x4 blocks = 4 blocks in each plank (30x / 4 = 7.5x faster)

If you implemented 4x4 blocks instead of 16x4 planks, it seems like it would run 7.5x faster than it ran before you took the debug logs out. I may be wrong, though. That may or may not be in your budget, but it's worth considering.

[Geti]

30 times as fast, huh? If you wanted to implement 4x4 blocks instead of 16x4 'planks':

16x4 planks = baseline (30x faster at this point)
4x4 blocks = 4 blocks in each plank (30x / 4 = 7.5x faster)

If you implemented 4x4 blocks instead of 16x4 planks, it seems like it would run 7.5x faster than it ran before you took the debug logs out. I may be wrong, though. That may or may not be in your budget, but it's worth considering.

Considering it's still likely to be synced over network to ~30 players in real time anything we can do to optimise is a good idea.
Algorithms like this (essentially a large stack of conditional checks on an array full of the sand tiles) perform in linear time, however the operations per unit N is quite high. A high number of elements will still be crippling. We're looking into other ways of optimising it - that was just proof of concept in flash - and we'll keep "proper" fluid dynamics in mind, but we'll likely stick with something tile-based with some effects thrown around to hide the falling plank effect (ie don't draw them as planks, use pretty water flashy instead)

[SolarLune]

Oh, of course. If you can make it look good (which I'm sure you can with some particles and such), it would look awesome. And you're right - with an online game, especially that features building and destroying, every millisecond counts.

[Brayzen]

Today I created...

... some more code for my first proper game. All I have is a devlog and print screen:

http://gnrgamedev.tumblr.com/post/8737706367/dualris

I've only spent around an hour or two on it today, but I touched up the logo, implemented some basic collision detection (and set up a framework for future expansion), implemented keyboard controls (and debugged) and correctly positioned the stack elements. Few more hours on it and it should be ready to test and release 

I'm learning a lot the intricacies of Java libraries and some special techniques, especially in graphics, but that's for the "Today I learned...." thread.

[leonelc29]

today i created...

one more special effect. the game look better now

[davidp]

I created a webpage for my 'shmup A week' project:
http://shmups.itsmelol.com/

I'm planning to release a new flash shmup type of game every week. Let's see how this turns out.

[Psy]

I added a level selection menu which slides between levels in a fancy-pants manner.