Build grid suggestions/discussion

-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
<<Don't read the next section if you are only interested in the "build grid discussion".>>
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------

Hello,

I recently bought early access to PA on steam.

I am a mod maker for both SC2 and MineCraft as well as working on the Vega Trek project and I was a major Diplo mod maker for the Blizzard scene (now dead as I'm sure you know) and I see huge potential here for this game. I am also a University CS student with an interest in path-finding engines and AI (I've created one working pre-mapped road path finding engine to date using ArcMap data for a class project in Java, I am moving to a video game variant that I am also using to learn C++. I am playing around with various "path-finding memory" systems to reduce path-finding loads by allowing large groups of units to share path-finding calculations as well as other ideas).

I am really interested in this idea of spherical path-finding. It will be interesting to see how you are doing it because so far it doesn't appear to be to bad. I haven't actually loaded it up yet though.

Before I begin on the build grid I would like to inform the web developers about an issue between this forum and IE 11 + Win7 x64. I put in a ticket but then appear to have solved the issue.

In order to be able to put text into this box with IE 11 and Win 7 x64 you need to add uberent.com to the IE 11 compatibility list otherwise the text box that you use to enter this text will not show. So that's a bug you may want to put a heads up on.

-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
<<Ok you can start reading.>>
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------

Since you are using "spherical" maps I going to have to suggest looking into build grids made up of anything from pentagonal to octagonal blocks, much like the Civilization series.

Blizzard with SC2 and previous RTS games kind of made those build grids and the terrain generator one and the same. I haven't looked to closely yet but I'm hoping you used a 5-8 sided polygon for your terrain instead of a 3. Now I don't mean that you actually make them only 5-8 sided polygons I mean you have an assortment of triangles that easily makes up a clean 5-8 sided polygon. A hexagon can be tessellated in numerous different ways to obtain different shapes. That could also simplify shading by having the light shade system only need to calculate for the hexagon "cell" (as Blizzard calls them) instead of each individual triangle.

But this since, due to latitude and longitude lines making a square shape different sizes depending on where you are on the sphere, anything with 5 sides and up is the only real option, not to mention being something new, it is also going to affect art and design.

Some structures, like the power plant, will be ok with a hexagonal design, but others like some of the factories may have issues. In essence more circular designs for a lot of these structure models may be appropriate.

I think this build grid should be a very high priority though because what I saw happen on my very first game is a unit get stuck between 2 structures and a build grid allows you to force some "tolerance" like Supreme Commander has between structures as well as being easy on the path-finding engine since it appears that structures are currently be treated as units.

What do you guys think? Can the build grid be done like this?

No matter what shape grid you have, it's not going to work.

Planet sizes vary, so the grid size would have to vary, or the grid would only work at a very very specific size.

[EDIT] https://forums.uberent.com/search/?type=post The community has discuss grids a dozen times before.

I am willing to bet my life that I could go to a college math professor and get an equation that would tell me exactly how many sets and what the sets are of hexagons/pentagons/octagons etc would be able to perfectly fit together to make a sphere like object. With a Hexagon I ought to be able to make multi-size spheres easily enough.

The fact is that without a build grid this game will probably fail. Its lacking basic RTS aspects right now. RTS is an extension of board-game-style turn-based strategy and the more elements of its heritage it retains the better its going to be. Otherwise you have a tug-of-war or MOBA style game.

Please do show me that equation.

Google "Spherical_Geometry".... its a wiki link. Apparently "new users" are not allowed to post any links even though I did technically pay for access to the game...

Didn't even have to go to a professor to prove it could be done. Its done in regular CG.

Now I'm not gonna pay the professor the money just to prove this. Their programmers ought to know it.

Their own sphere is probably just make up of triangles, you can divide other geometry up into triangles for rendering.

I assume you mean this link on Spherical Geometry.

Not seeing anything in there to support your claims.

Well after doing a little homework it turns out I was wrong about there being a solution using a single shape of geometry. It may be possible but not with anything the developers would be willing to take that many pain killers working with.

But I do not believe I have come up empty handed.

It occurred to me that with latitude and longitudinal the problem is that they are drawn off a specific and locked axis. The "square" at the equator is going to be stretched from the one near the pole.

However the scaling around the sphere is going to be the same no matter the orientation from the poles. So instead of locking the build grid to the poles, instead unlock it, in increments (from the poles), and allow the "equator" to always be underneath the placement model.

Now this is going to cause overlap with any nearby structures but the great part is the overlap is small and consistent. So you can build a tolerance into the structures from the edges when you make the model and the path-finding.

In addition that same tolerance you build in (that is in Supreme Commander too) will allow units to prevent themselves from getting stuck.

This won't override any current path finding mechanics, and its just one extra mesh to render (and you probably don't need to see the whole thing). This won't make pre-mapped path finding any easier of course but it would be a cost-efficient way to work around the current problem.

So basically, after declaring a certain "grid" unit size, use the equation s = r * theta (theta being the angle in radians and s being the length of the build unit and r being the radius of the circle/sphere) to determine how much to "tilt" the build grid by (in the nearest north/south direction) to adjust it.

You could also render it on the map itself and just use a rotation offset to match the placement models current target while its "hovering" (for lack of a better word).

Ok I believe I am out of the deep stuff for now lol.

You are exaggerating the problem to the point of silliness. CnC:Generals had no grid and it didn't make it any less of an RTS. And with ongoing improvements in pathfinding and building placement UI, lack of grid is not a big problem at all.

RTS is an extension that broke the rules. Going from turn based to real time is even larger step than going from grid to free placement, I don't see why modern games should necessary hold to this mechanic.
And MOBA is a completely different genre, defined by separate set of tropes.

Through, I would like buildings to be (unless purposefully rotated) oriented parallel to nearby buildings if there are any, and oriented north if there aren't.

Download this, script up what you propose and show us that it works.

You need to free your mind from squares and right angles for this game. There are no corners to be backed up against, nor edges of the map.

Line and area build commands solve any problems that may have existed previously; more importantly, however, it helps to remember that this is a game of escalation and expansion. If you can't fit more buildings in your base, don't bemoan the lack of a tidy square grid. Just go conquer some more real estate to slap them down on.

Pretty much this. A grid would be a completely pointless waste of the developer's time. Here are some frequent reasons people like grids:

+Keeps things nice and tidy. Are you really worried about tidiness in the midst of utter annihilation?
+Allows units to travel through the base freely. Just space your structures when you're building them, it actually makes for a better flow than a grid system.
+In order to achieve greatest packing efficiency for power farms. Just wat? Packing efficiency is bad for structures that are vulnerable to AoE attacks.

None of these reasons have any basis in reason. If someone can give me a legitimate reason backed up by fact and importance as to why a grid system is necessary, please state it.

there is a problem.
you're dealing here with a shape that already has a ton of approximations around the edges. what edges you ask? well the edges of the cube obviously!

these planets are cut from cubes. Each face is bent concave until you get a sphere. and the borders are ironed out.

as a result. it'll be pretty complex for any machine to start griding in the center of one face and finish the grid without any overlapping stretching and tilting of the grid shapes on the other side.

Ok now I did not use your linked software to make the demo. There is no scripting in it and I have better modeling software on hand to do this with (Rhino 4).

I did however make a demo. The idea is clean and works. And in fact is rather easy to implement and involves no extra taxing of computer hardware or the graphics engine whatsoever from the rendering standpoint. As far as path-finding goes, you are just checking the same collision the game already does. So this is a rather small and easy fix all told.

I do however have an issue showing this around. The big issue of course is that I cannot link the stream so I will have to make a youtube video. I understand that I cannot use any in-game content and I also understand that I won't need to in any case given the audience. However a youtube video is several hours base-line of compression etc. and if I am to counter many of the fan-boy -
-------------------------------
(I'm sorry, but some of this is so blind it cannot be anything but. Watching TotalBiscuit's channel I can see that I am not alone in my opinion of just how damaging fan-boys can be to a video game if they are allowed to run riot. I am not mindlessly whining here, I am actually spending many hours thinking these suggestions through (minus several parts of the first post) and doing home work and putting forth actual solutions. If community management does not move the fan-boys aside soon then I can already see the kind of future this game is going to have).
-------------------------------
"explanations" and other posts then the video can be up to several hours long. If no responsible developer is actually going to be seeing the video then there is no point for me to make one. I have already gone to quite a bit of effort to suggest practical solutions to the problem as it is. Here is a screen of the demo in question: http://imgur<dot>com/6Ocf2mE. Yes it is a 3d model but a you can see from the top edges there the actual over lap for a 4x3 build grid ("huge" by SC2 standards) isn't bad at all using the method I've been talking about.

This does not change how the path finding engine currently operates, its just a very resource-inexpensive build grid overlay. It can be projected onto uneven spheres if need be with acceptable distortion.

Basically what you are seeing is actual meshes, not simple projections, being used with edge spacing. The idea is to do a temporary rotation of the main coordinate grid so the build "mesh" always centers at the "local" equator even though that is rotationally offset from the "model" or model equator. If there is a shift in mouse position to any side you just recalculate from the "global" or "model" equator and what you are seeing is the overlap for "huge" (4x3) structures. Put a model size and path finding collision offset just like Supreme Commander so units don't get stuck and this should not become an issue. There is no need to render any additional geometry, just a texture projection. You can do such projections pretty cheaply in a JavaScript web page so I assume a full-blown purpose built 3d graphics engine should not have a problem with it.

Now why is a build grid needed? Well as a mod maker I know something the developers probably have not had the opportunity to actually test. You see I can try out new GUI ideas much quicker and cheaply (pretty much free) on the worst consumer base possible (that would be the owners and active players of StarCraft 2) to see if they bite. And I (and others) have tried this very placement system on the much more simple and flat terrain of StarCraft 2.

NOTE: The following is NOT a promotion, I'm just explaining how to exactly replicate my point.
First to replicate this yourself, just grab the StarCraft 2 Galaxy Editor, Go to the data editor and through some complex and annoying process (potentially) remove certain (or all, I cannot remember off the top of my head) placement "foot prints". This gives you almost exactly the same placement and path-finding behavior that this game has for small areas (like a "base" for instance).

Trying this on a set of StarCraft 2 owners and players (the current "mainstream RTS") and I found that not one person found themselves capable of giving it a positive review. As I recall the most common word I remember being used to describe it was "annoying". So I have proven via a mod that a placement system almost exactly like this has already been proven to fail. And fail hard enough to kill a mod. People are in general more forgiving of things that they do not pay for. They did not pay extra for my mod. They do pay quite a bit more, however, for this game. So the average consumer (as represented by StarCraft 2's extremely 4-letter word player-base) is less likely to be forgiving of something they "perceive" to pay for (key concept: perception being more important in marketing than reality) than something they "perceive" to get for free. So if a nearly identical building placement system has been proven to kill a "free" mod, then is most likely going to kill a "not free" game.

This is not the only evil I have found the developers have committed in the design here. And it may not even be the largest. But the core issue comes down to this:

How big of a "step" can you take in "innovation" and still have it sell to the consumer?
So far I am seeing the consumer being asked to swallow several large pills:
1. Spherical maps.
2. Managing several maps at once ("planets").
3. Almost entirely different GUI.
4. No build grid of any kind.
5. <And probably more I am missing here, I am giving feedback based on incremental exposure>.

And that's just the major aspects I can name from a mere 2 games. Now me who has tried to take the lessons James Burke in his Connections series tries to point out-
-----------------------------
(Bill Gates is quoted as same James Burke is his favorite author, and before you judge him, please watch at least a few shows series 3 (series 1 and 2 being more "hard core"). James Burke put almost ever major work he produced up on youtube so you can watch all of the Connections episodes for free).
-----------------------------
-the "new thing always starts out looking like the old thing it replaces". There is a very long list of serious changes here and I think its too much for a the average consume to swallow here. And of these, parts 3 and 4 are unnecessary and, as I'm arguing and going to argue, harmful to the game. There has evolved a Standard RTS GUI that Blizzard has used for well over 15 years (I first saw it back in WarCraft 2) and has been used by Supreme Commander and in series very similar to this in game play mechanics. The same issue is with the build grid. It doesn't actually have to be that way behind the scenes (or "mechanically") but the user should feel the familiar aspect. And I hope I've just demonstrated and it can been done and experienced with relatively little hassle.

You have 4 huge "innovations" or in some cases just "changes" and I think that's too much for the average user to swallow at once. I have personally tested parts 3 and 4 in a modding environment for large-scale strategy and I have alarm bells going off in my head here. Both can cause the game to fail when it's luck is being pushed a lot already in parts 1 and 2.

As for RTS being descended from board games and how "all" of the rules were broken. I suggest you go back to the early generation RTS games, like StarCraft 1, Dune, very early CnC etc. There are far more board game aspects, including path-finding, then you seem to think. It eventually evolved from that point as hardware and programming methods became better but it was a gradual thing over 10-20 years. The incremental change is the key aspect. I am liking many of the changes and "innovations" but I am also seeing the boat being rocked too hard and certainly, as I've begun to point out, harder than it needs to be. The more familiar the game is "perceived" to be by the average consumer the better the chances of a second one being made and the better the chances of this "evolution" succeeding.

So I am going to be back to the main point of this post. If I make a full video to demo the concept as well as possibly counter some of the other arguments made here, will a responsible person actually be looking at it?

I'll look at it.

I originally wanted a grid. Then I realised that it's impossible without hand-waving in fudge-factors.

Don't shy away from mathematical proofs either. This forum is filled with people with university degrees.

You had proven that grid is required in Starcraft 2, not that it is required everywhere all the time. Starcraft has very specific map design that was built from ground up with grid in mind. In PA I can see places where grid would be downright harmful, unless you tie it into the core of planet generation engine, which is not an easiest thing to do. You are basically asking for grid for the sake of it being there because "it has to be that way, otherwise everything will fail".

PA is a daring game, it takes risks in a lot of ways. Is taking risks a bad thing to do? You may say "yes", but at the end of the day best games are made when developers are trying to make a game they wanted to make, not when they try to appeal to a wider audience.

And, again, lack of grid did not kill the Generals.

Also, maybe it's because I'm not a mathematician, but your picture makes absolutely no sense to me.

Are you actually a "responsible" developer? I don't see a tag on your account any where.

Teod, you are talking like I can actually force them to make a decision that I want them to make. If I could really do that I wouldn't feel compelled to write such long posts where I do my best to prove my idea with other reasoning.

My original argument still stands. From an artistic aspect, this is great. Innovative, new, etc. But from a business aspect (where they put dinner on the table every night and be able to afford to make another game) its an extra and as I see it, unnecessary risk. Spherical maps can be hard enough for the "average consumer" to adapt to. They are already rocking the boat very hard. But they are also designing and experience. Its 3d vs 2d chess all over again.

Remember the whole fiasco with Windows 8? What was the number one issue people had with it? That's why Windows 8.1 exists. It rocks the boat pretty hard in many ways but the rock that threw everyone off wasn't an internal mechanics issue. It was not having the familiar user interface.

In fact Windows 8 provides a perfect example on an even larger scale of exactly why this different command interface (right instead of bottom right) is setting off so many alarm bells in my head. Simply removing the start button made Windows 8 sales tank and so far they have not recovered despite a new re-release and they probably never will. You cannot battle human nature. Humans are a lazy creature and they rarely ever want to actually learn something new.

And what difference does the command card location make to the over-all gameplay? Absolutely none what so ever. It will not change how the units interact with each other in the slightest. Hotkeys are going to be used by the majority of the consistent players in any case. What will that button frame's location do to sales when you factor in human nature? Make it or break it? It sure did with Windows 8. In fact it wasn't even a frame. It was the lack of a single button that made Windows 8 fail more than any other reason (there were certainly other reasons, but the biggest gripe was removing that single start button).

You are, again, basing your argument on the assumption that grid-less placement is unintuitive and unfamiliar to the players. It is not. For the majority of players it feels completely natural in here, like it did in Generals. All the way through the alpha and half of beta there wasn't a single complaint about it until yours. Yes, some people were asking for grid, but they were basing their argument on aesthetics (where it is a minor improvement) and pathfinding (which will be significantly improved and will stop being a problem).

@Triplitz88 that's like the worst possible example. It's 100% a matter of opinion.

Its crazy you metion this right now as right now i am copying the last of the files i need from my C drive.

I am about to format my drive clean of 8.1 and install 8.

Lost of things went wrong with 8.1 that had no need to be and I never had any problems with 8's interface since i feel it is largely adressed to the linux public.

First off widows 8.1 added the windows button back into the taskbar. and worst of all, left no option to remove it again, so i have to have an executable to remove it.

I liked that they let you boot to the desktop but my workaround to get to the desktop automatically worked faster than the one to remove the windows button. most times i never even saw metro on win 8 now I see the start button every startup.

then there's the e-mail thing. It was a big stress during install but I phoned a friend who had it and he told me i could just make a junk one then de-associate it from the O.S. a bit of a pain in the arse but it does turn the O.S. back into a normal account type system.

Then there's the libraries. Now that's a definitive fuc-k-up right there. Libraries were like the best thing that ever happened to windows and i'd still like to have an equivalent under linux.

But they decided Fu ck libraries! why ? no idea. Anyways through an extremely roundabout way you can bring them back halfway but only under favorites. not as their own section so you can't browse sub-directories in the left-hand side bar like you used to. plus there's the added folders in the my computer section that you keep cliking on by accident because you're so used to the old browsing from libraries but those are only on the C drive so they're more than utterly useless.

also the new IE is crap and nothing works on it (I know nothing new, right?) but i'm saying this compared to windows 8's IE.

So they broke it for no reason. Embedded youtube vids don't play, you can't post on this forum, ect.

then there's sleep which no longer frekin works. everytime you tell the computer to sleep it just wakes up after a minute!

it's the most irritating thing!

what else?

I forget but basically. my advice to people: don't upgrade to windows 8.1.

Teod ok but I just got this thing because steam had it on sale for what looked like the final sales price and given that I don't have any other valid RTS to play (Save the Star Trek Armada 3 mod which I don't want to over do) I decided to finally give this a try. I have been watching for a while.

Consider who is currently testing the game. These are not mainstream market players. This is for the most part hard core fans. There have been complaints running rampant in several spots about various aspects of this game.

I mean I had a bug where an I unit got stuck between structures. Build grid will not fix that issue for a game like this, only tolerance in pathing around the parameter of the structure will solve that just like in SupCom. The issue with this game just like in SupCom is that it overloads too quick. This is not unique. It when you see units start to "teleport" its the path-finding engine being over loaded.

Now as I am a CS student who is focusing in path-finding and AI I think I have a solution for the problem. I've already done a prototype with a pre-mapped road system in Java but now I'm busy learning C++ to take a wack at a dedicated video game system. The core concept of the solution is very simple but there are dozens of different ways to implement it, some of which may be better than others.

But back to GUI. First of all to the other poster your experience appears to be roughly unique. And the re-release of Windows to 8.1 is kind of the proof of my point. It cost Microsoft a lot of money to do that re-release and they wouldn't do it without a very compelling reason.

As far as the build grid goes. As I said I have found that people think its annoying. But it may be easier and cheaper to build in sooner rather than later. Gameplay wise it allows for much quicker structure placement for things like path-finding blocking ("walling-off").

A "Geodesic grid" is what you're looking for, I think.