Suggestion: New type of AA gun/cannon

[BulletMagnet]

Per's right; both are plenty fast... unless you're doing many thousands of these every cycle.

[Zarel]

Per's right; both are plenty fast... unless you're doing many thousands of these every cycle.

It's O(n^2), so assuming a max of 25 FPS, 8 players, and an average of 100 units+structures (not unthinkable, for midgame in an 8P game), the visibility code does the double multiplication about 16 million times per second. The targeting code uses another 16 million cycles.

So yes, way more than "many thousands".

...hey, isn't the visibility code poorly optimized enough that no one wants the Z sensor view to be auto-on? Maybe my optimizations could be useful there.

[TVR]

... the visibility code does the double multiplication about 16 million times per second ...

It's a good thing we have TeraFLOPS processors these days.

... isn't the visibility code poorly optimized enough that no one wants the Z sensor view to be auto-on?

Range display is never the default view in any RTS, as it becomes exceptionally confusing when circumferences overlap.

[Zarel]

It's a good thing we have TeraFLOPS processors these days.

I'm pretty sure there are a few Warzone players who don't have supercomputers.

Range display is never the default view in any RTS, as it becomes exceptionally confusing when circumferences overlap.

...erm, it's default in nearly every RTS. Look at StarCraft, for instance. I'm talking about the Z-view (visibility view), not the range view.

[Olrox]

[TVR]

... I'm pretty sure there are a few Warzone players who don't have supercomputers. ...

Even the Pentium II could perform ~100 MFLOPS double-precision, well above combined 32 MFLOPS load.

Fortunately technology progresses with time, thus consumer X86 processors can now perform 10+ GFLOPS per core.

... erm, it's default in nearly every RTS. Look at StarCraft, for instance. I'm talking about the Z-view (visibility view), not the range view. ...

Fog of War? That as the default would be computationally free to implement - visibility is already pre-calculated.

[Zarel]

Even the Pentium II could perform ~100 MFLOPS double-precision, well above combined 32 MFLOPS load.

Fortunately technology progresses with time, thus consumer X86 processors can now perform 10+ GFLOPS per core.

Regardless, change the O(n^2) to an O(n^3), which I think parts of the visibility code does, will mean that it might be felt even on 10+ GFLOPS.

... erm, it's default in nearly every RTS. Look at StarCraft, for instance. I'm talking about the Z-view (visibility view), not the range view. ...

Fog of War? That as the default would be computationally free to implement - visibility is already pre-calculated.[/quote]
...go play a game of StarCraft. Or just watch one. Like here, at 8:06: http://www.youtube.com/watch?v=B-4DN8RKNDc#t=8m06s

There's black (fog-of-war), there's darkened (no-visibility), and then there's regular (visible).

Warzone does the same thing, but only while holding down Z, because it's kind of slow.

[TVR]

... change the O(n^2) to an O(n^3), which I think parts of the visibility code does ...

IIRC, Warzone 2100 map format is BSP, therefore visibility calculations are O(log n).

Either way, the time spent on accurate simulation is well spent, as demonstrated by Warzone's niche of armchair generals populating this forum.

... There's black (fog-of-war), there's darkened (no-visibility), and then there's regular (visible)

Unexplored terrain is shown as unrevealed on the map, whereas Fog of War is explored terrain that is not currently covered by sensors.

Visibility display should be as simple as darkening the unseen sections, as all the visibility sets are already calculated.

[Per]

Warzone being a true 3D game with true raycasting visibility simply does not compare to most other games that use far simpler visibility systems. We need to improve Warzone's visibility system, not just to get the 'z' button behaviour permanently but also to implement player-specific map knowledge, so that you do not get information leaks about what is happening in areas of the map you can no longer see. For example, you can see buildings vanish when they are destroyed if you have only seen them once, and see oil resources vanish when being built on now that we have full map knowledge of oil resources.

So there is a lot to do, but the difficulty is finding good algorithms to cache information to avoid having to recalculate visibility all the time.

[praefectus]

LOL some one hijacked my topic ... xD

ok then, if you can provide me with a complete description of the mechanism you have in mind I can make a model next week, so we can see if someone gets interested on testing it.

Thanx for the offer Olrox, but I don't think I'll be able to.
All I can contribute with is the idea, which I can't describe any better than I've already have.

I still do think WZ need an effective weapon against massive VTOL formations....

[Olrox]

LOL some one hijacked my topic ... xD

Those are closely related, to say the least.

I'll think about your turret and will post it as soon as I can make something.

[winsrp]

hmm.... new type of AA.... what I would do, is combine some type of plasmite shell with flak cannonn, and make some sort of air napalm mini shell. It can be easily done with current game mechanics, just need a new model and I will set up the weapon stats.

[Olrox]

hmm.... new type of AA.... what I would do, is combine some type of plasmite shell with flak cannonn, and make some sort of air napalm mini shell. It can be easily done with current game mechanics, just need a new model and I will set up the weapon stats.

Would be an Anti-Aircraft flame weapon of sorts, I get it right? I was thinking about that, because flamers are my large group fighting weapon of choice, they are devastating when used properly. Considering, of course, that I'm not fighting many dozens of Twin assault cannon Mantis Tracks - that would require much more firepower.

Perhaps we should go this way. The flak gun would fire incendiary shells, with delayed activation proximity fuze (just an explanation for it to blow at the right time. If anyone dislikes that, there are trained mice inside each shell that activate the fuse at the correct time: they are good)

I'm searching for flak cannons for reference, I'll begin my artistic works again after day 17.

[winsrp]

Would be an Anti-Aircraft flame weapon of sorts

Yeah this is exactly what I was talking about but, I just didn't want to put it that way... jejeje.

The plasmite on the other hand has an extender fire radius on it, that would make it for an upgrade from a simple flame Flak AA... so we could have

Fire Flak AA (in lack of a better name): when shell fires it does fire damage on target (or mice take out the grenade lock, you name it.)
Plamite MIssile AA, since missiles can carry bigger loads they can have a more range explosive fire damage type on them, but I wonder how fire burn radius would look if shells explodes in the middle of the air... would have to try that out.

[praefectus]

...there are trained mice inside each shell that activate the fuse at the correct time

KAMIKAZE MICE

I LIKE IT!


OT : Seems like you two caught the idea spot on... and improved it. Exciting!