Yeah, that's what I mean. You were right, but I was just expanding on why it works that way.

So do unhittable MOs contribute to the MOID limit then?

No. To be counted for the MO limit, an MO has to have an ID. To have an ID, an MO has to be able to be hit.

That could prove extremely useful in cutting down on MOID limit depletion on things which don't have to be hit or should only be able to be hit after other stuff has been destroyed already.

It's also incredibly annoying that there's a limit at all.

I think there's a limit on the number of MOIDs because of the way the engine works. My theory is that Data actually uses the colors in the palette as the ID of hittable objects. Since there are only 255 colors (besides the null color) there can only be 255 MOIDs. He could probably link IDs to a number in a list, instead of colors in a palette, to have more (or infinite) MOIDs available.

I think that data limited it to 255 MOs because his collision detection probably checks that list of 255 to determine which one to affect.

It's limited to 255 to keep the time it takes to check the list down.

That's a pretty inefficient system, but, hey, think of what we have to do with Lua to check simple things like proximity. The fact that there's no dedicated engine functions to check whether MOs are near each other kinda leads me to believe that Data's engine coding isn't the best.

Quadtrees in the future perhaps? I've been thinking the same thing recently, as that's what most engines these days seem to get out of the way as quickly as they can, but hey, it's Dan's game..

I have a feeling the collision system works off of these colors we see in the MOID view, by checking if colors hit or intersect other colors. Which is why you can only have 255 hittable objects at a time.

I agree with Grif as I have noticed several times playing Cortex Command that if there is more than 255 MOIDs, actors that normally should not have a MOID (ie they are MOID 302 or something like that) other actors can walk right through them. It is as if they lose the GetsHitByMO's and HitsMO's from them to get rid of their ID. Doesn't always work, however. Also, I may be wrong about this and Data may very well have linked MOIDs to colors, but I believe that using colors for the MOIDs would be impractical and unnecessary.

So what you are saying is that the collision system is limited to an 8bit palette? \:|
If so then all it would take to get past this issue is switching to 16bit (For the MOIDs only).

Or making the collision system work off of numbers/arrays.

Like, instead of an image, you have a list with 3-numbered entries. 2 numbers are for "pixel" position, and the third is for ID.

(1,1,39),(1,2,39),(1,3,39),(2,1,39),(2,2,39),(2,3,39),(3,1,39),(3,2,39),(3,3,39)

^That would be a 3 by 3 pixel sprite with the MOID 39.

Take a look at DeepGroups and AtomGroups. As far as I can tell, they work the same way you're describing. When you set AutoGenerate to 1, it's generating that array for you. The Resolution is related to the number of pixels around the contour it skips. I'm sure the are reasons higher up than color for the 255-MOID limit, but it may be one of them.