What I'm doing right now is taking my big-picture concept, and breaking it down into a core engine, and a number of smaller, manageable sub-projects that eventually can all fit together as a cohesive whole.

My favorite example is Rock Band. It's basically just FreQuency/Amplitude, Karaoke Revolution, and Guitar Hero all put together, all played at once (give or take a drum controller).

Looks pretty sweet. You have a downloadable, by chance?

Loving the delightful minimalist style of it.

There's plenty of game engines with user-generated content and games with level editors out there; you might want to begin with one of those.

You should look up Zelda Classic, Spelunky, VVVVVV, MUGEN, and StepMania for examples of these.

Another solid alternative is Terraria(2D)/Minecraft(3D); which in themselves are fully-editable level/world generators with inherent gear and items/equipment that you can modify pretty freely.

Usually, there's 4 key puzzles per area to figure out.

1. TO GET "THE KEY."

The key could be a persistent token item like Zelda's Big/Boss Key, or it could be some methodical change in the environment to change the terrain and how you can navigate it.

2. TO GET "THE ITEM."

This usually goes into Metroidvania/Zeldalike progression, where a new ability is introduced to the player, and a new gameplay mechanic for which it's used.

3. TO GET "TO THE NEXT PART."

One token new gameplay mechanic puzzle is often a gateway to the later part of the level, where the gameplay mechanic is worked into the level design.

If you really want to be interesting, you can switch these two around, using the environmental change "key," and have that open up a kind of backtracking that includes some freshness to it.

You can also have extra goodies/loot scattered about behind these gameplay mechanics as well. But in those cases, the areas should be clearly optional, and the mechanics not required to navigate the area.

4. TO BEAT "THE BOSS."

A good boss encounter itself is a different kind of puzzle to solve, one more spacial and tactical in nature; and it may or may not utilize the gameplay mechanics introduced previously.

It doesn't mean you have to use the "3 phases" rule akin to more casual games (although that approach can be used to make for some interesting AI). You can combine raw risk/reward damage tension to it - making a kind of damage sponge that isn't 10,000 hits to beat, but more like 20 to 30. And then, it's still usually a good idea to have a "trump card" so that tactical or experimental players can still find a way to "outsmart" the boss by applying some kind of super-effective weakness.

I prefer some influence in air control, but not outright direct control. This is why for the MegaMan/Sonic/Metroid hybrid subproject I'm working on, I'm aiming for more Sonic-like physics.

Don't get me wrong, MegaMan/X/Zero could not be the level of coolness they are without precise air control; because those games are built to expect that kind of precision out of the player, especially in post-8-bosses fortress levels.

That said, having techniques that can influence jumping can make air game still really fun to play. Double jumps, wall jumps, midair dashes... there's still a lot of player influence you can put into a jumping system that doesn't have immediate directional response.

An elevator object, and a switch you can hit with your sword to take it up and down, maybe?

Nifty level designs so far. J-Snake's gonna love you, it's a lot like TrapThem.

And I grew up on old-brick GameBoy, too... remember how linear those physics were?

So long as it's mostly around other positive developers, you shouldn't have any more leakage than you'd need to create buzz anyhow. And as long as your pricing is reasonable, it shouldn't be too much of a concern. Spelunky and La Mulana were complete freeware for 3 years before their commercial releases, and they're both doing okay.

Having other developers contribute levels is a great idea, it was a very nifty bonus with VVVVVV, for instance.

I just downloaded Oryx's tileart Brogue 1.7.2, the TrapThem demo, and Bunny Land.

Gonna play those over the weekend, during dev breaks.

I thought about associating symbols with words or numbers (as hints around the area?), then using either math/lingual puzzles, or sequence/logic puzzles.

Really, it's usually about completing a successive sequence of events most of the time, although Zelda pushblock puzzles are a classic as well.

Now that's a good one I'd like to check out. Especially if there's any La Mulana in there.

Well, the negative grinding aspect comes down to JRPG's measuring their quality in playtime hours. A 40-hour game totally sucks if 35 of them is necessitative grind.

But compare that with grinding 1ups or powerups in SNES-era platformers (or SMB3). Super Mario World had plenty of "grind" by having multiple exits per level - you're not going to get both exits on one run. You can have some fun grind in JRPGs too, like Chrono Trigger or Cthulu Saves the World. Heck, Chrono Trigger even creates 2 or 3 locations *specifically* for grinding/training, where losing a fight doesn't even Game Over you. (Gato and Spekkio for sure. I thought there was a third with Masa/Mune/Masamune somewhere, but I'm not totally certain on that.)

The active battle system makes SMRPG grinding so much fun, you could do it forever.

Speaking of Yoshi's Island, I'm surprised there hasn't been something that's just a collection of minigames like that. Those things are always fun! XD

I guess it's not totally complete without the whole inventory/reward cycle that plays into the core game, though...

How does it feel? To take that to-do list down to double-digits, after so long?

Timing puzzle. They twist at 90-degree increments with rhythm; you can stand on them when they're sideways and they damage/drop you when they're vertical.

Looking sweet. I can tell from the influences in that map design, you've got the stuff. SO looking forward to this.

Derek (still). Red. Dom2D. PixelProspector. FinalSin, if you're paying attention to what he's up to. John Sandoval and Blademaster Bobo's 3-ring circus. JWK5 has been a lot too.

Northernlion. (You did not say developers specifically!!)

There's a good reason why. Most grind out there - especially in MMOs and MOBAs is complete garbage.

One example of GOOD grind is the covenant system of Dark Souls - it encourages you to explore the different dynamics of the game, and rewards you for success with them. Also, the souls in that game can be put to very versatile use.

[that]

Yeah. Trig is fun stuff. Writing formulae for Trig? Not so nicely. So here's a thread to rant, rave, poke fun at, or help with calculating weird crap in your games. I'm just kicking it off with a platformer-related post, is all.


SO! What am I doing here now?

Trying to create some trig-magic formulae: using a "speed" variable and a curve-radius constant to calculate the appropriate change in "direction."



The convex curves like this one have a radius of 48 px.



The concaves have a radius of 96 px.


So, clearly the first step is to calculate the difference between "speed" and how far to continue regular contact first. Then... Grr. It gets nasty and confounding.


I am this far (in theory/practice, but not in code yet!):



Question is, what order am I calculating these all in?

I have
> yellow hypotenuse = curve radius
and
> magenta hypotenuse = speed radius (blue)

I need to calculate

> magenta angle

...if the angle of the yellow hypotenuse can be calculated, how does it relate, or does it at all? Is it intrinsically perpendicular to it?

Using that hypotenuse/opposite combo to then calculate the relative angle change in "direction" is what I'm going for; because that is what has to be applied to the character object.

 


I know it's not just me here. I've seen enough Sonic fangame vaporware to know what's going on.
____________________

  I AM SUCH A DUM!!

Okay, so both yellow radii are equal, so there's no way that's a right triangle, it's an isoceles one!! And since those are the constants, so are their squares...

SO!

(2304 if 48 px; 9216 if 96 px) / (speed / 2)² = the perpendicular difference²

Then those can be used to triangulate the central angle of the split isoceles, and the adjustment angle IS perpindicular to THAT!!

  Now... how do I put that in GML code? ...

This. This kind of stuff right here. It's what I'm procrastinating on, because I've never had an actual Trig class yet...