Ashes of the Singularity

They're going to support multi-GPU through DX12. The entire point is that makes SLI and Crossfire redundant, not supported.

GRRRRAHHHHHHHHHHHH !


http://forums.ashesofthesingularity.com/473306/page/1/#3602571

Dang. How do you mess up your pathfinder so badly that it cannot deal with detached sections of pathable terrain?

IFKR???????!!

Well, that is easy.
You don't even need know how path finding works in detail, common sense is all you need for this.

Just think about what path finding in a more "formal" way means:
Path finding is the search for a valid path from a start point to a goal point within a given space. For an RTS you usually have a map, the start point is where the unit currently is and the goal, well that's where the player clicked on the map. Pretty easy so far.

But how to find a valid path? Well, just search all possible candidates of paths to check if one is valid. If a valid path is found, path finding is basically over.
The tricky part in path finding is how to order those candidate paths in a way to find a valid path really (I mean really!) fast. This is almost all there is to do in path planning (some examples in ascending order of ordering those paths are: depth first search, breadth first search, Dijkstra, and A-star).

And how to know there is no valid path at all? Well, that is a really hard problem. To know that there is no path, you need to look at all possible candidates of paths and check if they are not valid. And usually there are a lots of them. Path finding is only fast because we order them really intelligently to find a valid path fast, but that doesn't help us if we need to know if there is no valid path at all.

This is the reason why the worst case of path finding is the case where no valid path exists. In this case you need to go through all paths there may be.
In the case of an RTS map, which is a discrete planning problem, the number of paths is finite so you will get to an negative result eventually. But if you have planning problem in an continuous state space, there are usually an uncountably infinite number of paths, in such cases the problem becomes semi-decidable, which means you planner will run literally ( <- and this is the correct usage of literally) forever if there is no valid path.

(There are some cases which a trivially decidable, for example when the goal is in un-pathable terrain. Additionally in an RTS the map is usually known beforehand, this would allow to mark connected "islands" of pathable terrain with an unique marker. If you try to plan a path between an start point and an goal point with different markers, it is obvious that no path exists. But this gets a bit more trickier when you have different unit types with different path constraints (e.g. a path for a small unit may exist, but a larger one won't fit through the small gap on the path) or you use a real fog-of-war which affects path planning to (which would mean real SLAM)

Nonetheless, as a closing word for the brave people who made it through this wall of text: The idea of this post was to point out the fact that because finding an positive answer to a problem is easy, doesn't mean that finding a negative answer is also easy. In fact it is sometimes impossible in a finite amount of time.)

and my teachers say wikipedia is unreliable.

I'm just going to tell them that Crizmess uses wikipedia, so it must be reliable.

@crizmess
Dude I don't get why three simple lines of code wouldn't be enough to solve this "really hard problem". all you have to do is draw contours at every land border and make them meet ends. when they do you know that the area inside is pathable.

then repertoire all these areas and refuse pathing queries emanating from one area and attempting to go to anothert.

Voila. you've kept your extremely fast pathing code and allowed for islands to exist and function correctly.

why are you explaining this pseudo-complexity/ near impossibility of pathing two unpathable points when I can cite dozens of RTSes (including PA! ohoy! wake up!) that feature islands and path correctly on despite them. (I'm quite certain that when you try to move a tank that's on an island to another island in FA, it uses the technique I described above but also the game simply draws a line between your cursor and the tank and find where that line intersects with the island's border and simply places the move order there instead and then the pathing starts).

Why has this suddenly become so impossible?

I'm loosing my reason here. Is the world spinning backwards? are we travelling backwards in time?

A weird kind of past where games traded out any sort of fun for astounding graphics. Pah!

Doing that stuff on static maps that have islands certainly is not impossible, but if Oxide decided not to do island maps until later than it makes sense for them not to spent time on optimizations they do not need for what they want to do.

no. from what I understand they're absolutely befuddled by us wanting islands at all.

They have no naval units per se.

That said, I get it to an extent. These things are hover units. They can go over water, so why not?

And would you PLEASE cut these guys a break? They're deep in EA and are doing a decent job at responding to feedback. I mean, I wrote one paragraph on their Setons repro and a week later they implemented most of my suggestions. Might not have been related to me, but it sure looked good for them haha.

thinking I give these guys a hard time is missinterpreting it.

I give the game a hard time.

standards are high and it's not my fault.

I was born into this world without a choice on capitalism basically with the gun pointed to my head. I've earned the right to remind the world that participates in these doctrines of it's own doctrines : placing yourself in a situation of concurrence drives the demand for more up.

besides how am I to take this demand seriously when

you yourself fall in these guy's pitfalls.

we can't help ourselves but to take things for granted and then get slapped in the face with a fish because shiit be asss-backwards in this joint yo!

no they can't.

there was a big discussions about that where we failed to have any sway: http://forums.ashesofthesingularity.com/462351/page/2

Tatsu, now you're just playing devil's advocate for the hell of it. That's sickening.

Being forgiving of game devs decisions because you don't understand or know every facet will lead to more satisfaction in the long run

I'm not. are you implying that I should have no right to such backlash?

I'm implying that it wears thin when you do it all the time

regardless Ashes is going to fail D:

too many poor design choices .
It starts with their huge turtlephobia and ends with sticking to logistics.
They have the completely cut out baseline they must stick to with no room for creativity.
I tell ya whats gonna happen: Ashes - or more their balancing team - is going to tell the players how they are ment to play their game instead of players deciding that for themselves. Instead of employing any possible strategy the have already destroyed and rooted out plenty of them. It's absolutely ridiculous to the point where I'm just so disapointed that I threw my money at them- not to say outraged.

Yeah I don't know how to feel about the game at this point either. The first version will be content minimal and it appears that they are going the route of the civilization franchise where you need a minimum of two expansions before the game becomes decent.
I really really want this game to be good, especially since there aren't many interesting RTS games out there now days. (For example grey goo and act of aggression not being that good). But as time passes this is looking less and less likely to be the case. I'm still going to follow the game to see where it goes but at this point in time im not feeling that it's really going to go far. The minimal basebuilding is something I'm not a massive fan of either. Ah well all I can do is hope for the best now.

Not satisfied with the recently released rts games? Why not giving Total Annihilation Zero a chance? It actually is superior in gameplay to all the recently released rts games. The only downside is the outdated graphic.

No, tashadan, it actually isn't. TA lacks as simple a thing as an infinite build queue, and there's no way to get around that. It lacks area commands, line move, priority building, pushing units to the front of the build queue. If the graphics were the only thing outdated about TA, then I would be playing it, but the fact of the matter is that they aren't, and unless your mod includes some modifications I haven't seen before,TA Zero doesn't either.

Too bad Uber didn't win with their bid for the TA ip. John Mavor said he would have released the source code to the community - that would have been awesome, although it must be hard dipping into ancient code.

are you fukkin kidding me !!!

why even post that !!! what a tease ! I would have been happier not knowing ! WAAAAAAAAAAAA!!! uber shoulda won!!

TT_TT