Few LAN Users

[Joemadden1989]

Hi,

A few of us have started playing 0AD in work (As we are only allowed to play open source games this was ideal!), durning our lunch our. The most we've had on at once was 5 people all with there own army in a 5 way fight.

The game ran pritty good, the biggest problem we found was that there was a noticable lag delay in telling units where to go but no one crashed, had any problems etc etc.

Keep up to good work, and if there is anything you want us to try out specific for you and give you feedback please let me know. Unforuntatley I am not a coader so will not be able to help in that sence.

Thanks

Joe.

[sanderd17]

A bit of delay is normal. In a multiplayer game, every turn lasts 0.5s. During that time, players can send commands to the host. Then at the start of a new turn, the host distributes the commands to all players. So taking around 0.5s to see a reaction is normal.

If you have more delay, it's either because the network is slow (which is not that likely on a LAN), or (more likely) that the calculation of the simulation state takes longer than the give 0.5s.

I assume you want to play for fun, so you want to stick with a stable version. But if you guys would be willing to cope with a few bugs (sometimes minor bugs, sometimes major crashes), you could also test the development version (http://trac.wildfiregames.com/wiki/BuildInstructions). But as that requires you to install it on every computer, and update it for every game, and usually makes games less fun due to bugs, I would understand if you don't want to do this. But it would be a great help if you could test it (certainly after we committed the new pathfinder in the next weeks).

[Lion.Kanzen]

...(certainly after we committed the new pathfinder in the next weeks).

That is too faster, I figured this can take you guys whole year

[WellDoneToast]

That is too faster, I figured this can take you guys whole year

Is a new Path finder happening soon? I'd love to see a reduction in lag.

Edited June 3, 2015 by WellDoneToast

[sanderd17]

Yes, we will include the new pathfinder soon, but the new pathfinder will have more bugs and hiccups than the old one.

Like a few days ago, we saw that games with the AI had lag spikes of turns that took around 300ms (while singleplayer turns should be completed in 200ms of calculation) every time a building was constructed. The structure of the new pathfinder is better, so it can be used to allow faster code in other places. But the new pathfinder itself isn't optimised or bug-free yet.

[av93]

Yes, we will include the new pathfinder soon, but the new pathfinder will have more bugs and hiccups than the old one.

Like a few days ago, we saw that games with the AI had lag spikes of turns that took around 300ms (while singleplayer turns should be completed in 200ms of calculation) every time a building was constructed. The structure of the new pathfinder is better, so it can be used to allow faster code in other places. But the new pathfinder itself isn't optimised or bug-free yet.

But overall, the new pathfinder will be faster in the first release?

[sanderd17]

It will be theoretically faster (the JPS algorithm is a lot faster than plain A* on grids). But it will lack a lot of optimisations, which might bring the practical speed about equal to the speed of the old pathfinder. However, the optimisations needed will probably be simple things like avoiding memory copies, or recalculating stuff a bit more clever. The old pathfinder couldn't be optimised any further.

Moreover, it should be consistent now. With the old pathfinder, it was possible that one part of the pathfinder believed that a unit could fit through a hole, while another part knew it didn't fit. Which caused problems like units getting stuck in obvious situations. This was especially noticeable in forests, where there are many small clearances.

[Prodigal Son]

One major related thing to consider is the handling of forests. I see two viable options (there might be more):

• Like in most RTS games, forests being impassable. This needs to come with a way to place trees in random maps and the scenario editor with no random spaces between them, aligned to each other, maybe with slightly increased obstruction size and distance, so that not the entirety of the upper parts go into each other. A forest paint tool in the editor would come handy as well.
• Passable forests, coupled with some ambush mechanic for some civs and units. Either with obvious gaps between trees or no obstruction for them.

A better looking (not perfectly round) version of single large forest entities like in Wow's Delenda Est might also be worth considering if it proves too hard to totally fix the pathfinder.

[Lion.Kanzen]

Like in most RTS games, forests being impassable. This needs to come with a way to place trees in random maps and the scenario editor with no random spaces between them, aligned to each other, maybe with slightly increased obstruction size and distance, so that not the entirety of the upper parts go into each other. A forest paint tool in the editor would come handy as well.

I suggested that but the answer was " it's not realistic"

[Prodigal Son]

I'm mostly for it as well, even though I wouldn't mind the other option that much.

It can be called realistic: You can pass through a dense forest, but generally armies would avoid them for being a pain to cross (losing direction, carrying baggage etc) and to avoid being ambushed, as most armies were better fit to fight in open or relatively open ground. Units represent larger numbers in such games so for me it's not a bad compromise in realism. Gameplay-wise it's also better on most occasions since vision/control of your units in forests isn't the best even if using silhouettes. Their function as natural borders is interesting as well, having player bases open on all sides on most maps isn't that great for an RTS game. It also fits better in a macro oriented game (it's getting old I know:p).

That said, representing some units' and civs' strengths in forests (or uneven ground) would also be interesting, like ambushing careless armies passing next to/in forests, or crossing them to launch an unexpected raid.This made me wonder if a higher ground bonus could be easily implemented for melee units as well (fighting up/downhill).

Like on most fields of game design there's pros and cons on each decision, both on realism and functionality. Whatever the choice, the terrible issues with woodcutters/military units being struck between trees and the uncertainty on if a forest is passable or not need to go at some point, with no distance between trees or fully passable forests.

Edited June 7, 2015 by Prodigal Son

[sanderd17]

The thing is, the pathfinder is completely fixed now (apart from some bugs that we need to find, and some optimizations). So I don't know why we would suddenly have to change forests as well.

[FeXoR]

I agree with sanderd17

As forests are handled now is the way actual forests would influence units:

- Dense forests are unpassable

- Light forests can be passed but in general units can't walk in a straight line

Formations might have trouble to find their way through but we should test that with the new path finder first to think about a fix (that might actually be not needed in the first place).

In my opinion IF there still are problems with formations in light woods it's a problem of formations, not the way trees are handled. And I don't think it's good to change something else to make the "broken" part work better while staying "broken".

[auron2401]

Or we could just entirely make forrests impassable and give that buff to them iberians we've always been talking about.
Would make for fun gameplay.

probably easy to achieve, too.
their occlusion (is that the word?) could "glue" itself to close by trees, so you don't have to make them all bigger, thus making individual trees... er.. bigger than they look.


I didn't say it was easy to code.

Edited June 7, 2015 by auron2401

[FeXoR]

Or we could just entirely make forrests impassable and give that buff to them iberians we've always been talking about.

Would make for fun gameplay.

probably easy to achieve, too.

their occlusion (is that the word?) could "glue" itself to close by trees, so you don't have to make them all bigger, thus making individual trees... er.. bigger than they look.

I didn't say it was easy to code.

There is no special concept "forest" in the game, there are only trees. In real life as in the game a "forest" is a collection of trees (that might be so tightly packed you actually don't want to go through or be easily passed by one person but still avoided by cars/siege engines). So the density of trees make the difference if a forest can be passed or not.

What's the point in replacing a realistic and working (besides eventually formations that are still not satisfactory as is anyways IMO) approach with a less realistic concept?

[Prodigal Son]

Glad to hear the pathfinder is getting better and better lately:)

My point is that there needs to be a clear forest attribute for gameplay reasons. Creating "light forests" with individual trees or an extra option would still be possible, but dense forests need to be by default impassable (or passable), and easy to place without random, hard to spot, exploitable and buggy gaps.

Edited June 7, 2015 by Prodigal Son

[FeXoR]

Which gameplay reasons?

"Exploitable" in which way?

"Gaps" are the absence of, well, entities mainly. How can gaps be buggy? Maybe the pathfinder can be buggy or formation movement.

I'm sorry if sound like a broken record, forgive me ^^.

[Prodigal Son]

Could also be I'm not explaining clear enough even after all those posts, lacking the technical knowledge on some fields as well.

Forests currently can either be impassable or randomly passable through small, hard to spot gaps. Besides the pathfinder issues with units being stuck or slowing down, this has other effects as well. It's hard to be sure on what's the passability of each forest mass (without using the developer overlay). That's bad for gameplay. You need to be able to easily spot all pathways, not having to manually path-scout each tiny bit between trees to know if you need troops or walls to guard a place. It's kinda mean when the AI or a lucky/more observant player pass through those gaps and the opportunities for this are currently many and hard to monitor in most cases. Also gaps that allow some units through but not all of them aren't that great for gameplay either and very small gaps in general are bad for they can cause even small numbers of units to get stuck. It's not by accident or foolishness that games like the AOE and Warcraft series have no gaps in their similar forest masses.

Ofc the "real forests are random" argument can be thrown in, but it doesn't help gameplay at all. Another rather major thing to consider is trees often being almost on top of each other, with both aesthetic and gameplay issues (more workers try to gather there than possible, and this might have to do with the worker limit per tree and not the pathfinder, I'm not sure about it though - happens with other resources as well in random maps).

Edited June 7, 2015 by Prodigal Son

[av93]

So the new pathfinder will be on the next alpha? Maybe making announcements on social media (fb, twitter, moddb) with screenshots and videos saying that you need testers would make come more players.

[sanderd17]

That is the plan, we're currently making an announcement draft that should explain the changes so regular players can understand it.

[WellDoneToast]

That is the plan, we're currently making an announcement draft that should explain the changes so regular players can understand it.

Fantastic news . Looking forward to seeing the new path finder.

[FeXoR]

Prodigal Son: I Think what you describe is what I would describe as a "positive game-play feature evolving from sane and clean implementation".

I am failing to see the drawback, sorry.

For overlapping trees in random maps: Some maps use non-fitting random offset and average distance parameters. Another case is the unit group placer where this can happen. This should be avoided but it's not a big deal either IMO.

The map where you will likely find the must unsatisfactory tree placement (from your point of view) is Deep Forest. There, however, you won't find overlapping trees.

While writing this I found me realizing the ability to place "forests" in many different ways (by being made of separate trees) shows the strength of this approach by granting diversity IMO.

[niektb]

I think Schwarzwald is Prodigal Son's worst nightmare concerning this forest issue