You have to

a) pick the two colors to match the sky textures correctly

b) use the skybox fog setting

c) put a fog_controller in the map itself

d) make sure they match EXACTLY or it's really obvious.

e) use VERY BIG AND LOOSE func_smokevolume's up against the sky

I originally had smokevol's working as mist in all the valley's and it looked really cool. So the effect was even better. This light mist just floated above the ground. But there is a limit on the number of smokevol sprites you can use, and it' seems to be gated in the engine, so I couldn't make it all work without exceeding that limit.

While there is a performance penalty using them, it's actually pretty lightweight.

-LN

This is a problem with source inthe current sdk. See "cubmaps bug" in VERC. THere is a workaround but it seems not to be perfect. I'll use it in the next build.

-LN

I don't understand. Power of 4 would generate MORE planes.

Not long at all. the map is constructed as a grid of 1024 unit cubes over which the displacements form a loose skin. THen the entire map is pretty much made of func-detail and models. So a glview of the map is a very tidy very clean set of vis areas. Almost everything is func_lod that can be, and all the models have an LOD set. So, rendering is controlled by lod not necessarily vis.

This is the technique used in the hl2 coast_12 map. It is also how many other engines handle the problem of exterior space.

-LN

First of all, I really like alot of things about this map - the attention to detail and obvious thought/care that has gone into it. Now, allow me to go off on a tangent. I don't see you changing this map much, and I understand that, but I'll lay all this down anyway - food for thought for future projects, maybe.

One of the benefits of higher detail maps, larger worlds, and more freedom to build is the ability to closer mimic realism. This is also a very double edge sword. Last generations games, in their more simplistic form, forced a sort of 'suspension of disbelief' and made it ok for developers to define the gameworld in a very rigid fashion. Basically, it was very obvious that you were playing in a confined box - everyone accepted it, no one really cared. Now that this constraint is somewhat lifted, we can create the illusion of LARGE environments.... the only problem is.... you are still playing in a confined box. Only now, you don't realize it until you 'smack' into an invisible wall that defines the gameplay space.

This creates a conundrum... we can make the player 'think' they have much more space, but we actually can't give them that much more space.

Another problem with our newfound next-gen capabilities, is the opportunity to create much more confusing gamespaces. The simple forms of previous gen games made the developer and player focus solely on 'playable space'. This meant that the game path was clearly defined and usually easy to understand. You could typically drop into a map, and know right off the bat where to go. Now, however, we have to build very defined game paths that fit cohesively into a more realistic environment, but still read well, are well lit, easy to navigate...etc. This takes more skill and forethought.

There within, I think, lies the major flaw of this map. The illusion of space and touches of half-realized realism, are detremental to the gameplay and overall quality of the map. The game path is confusing and unclear in lots of places. There are far to many invisible walls, and the sheep feel like statues.

If you've never heard of the theory of the 'uncanny valley', you should look it up, because this is a good example of that principle in action. To sum it up, the closer we come to 'realism' the more jarring or unsettling the subtle flaws in that realism become. The uncanny valley deals with simulated humans, but I'm extending this to realism in general, because I'm just that kind of pioneer =D

On entering this map, I immidiately see this wide expansive view, which it looks like I have access to. What do I do? Of course, I run towards it - only to hit an invisible wall. realism broken. Then, I see these cute little sheep, that look very good at a distance - I run up to one and study it. No movement. 'realism broken' Now, I'm trying to figure out where I can go, and where I can't - if there were people in the game, I would be dead because at this point I'm somewhat confused about where I can actually go.

So... the question you're probably asking is "what would you do, mr. critic... with all your critical critiques'. Well....I would probably do just what you have done, in trying to accomplish all of the great ideas you had when dreaming about this map. Then... one day when I had the time to go back and do it again, this is what I would do.

A. Define a gameplay path that is very clear. building a distinct ground path/game path, with cohesive, yet unique textures, to signify exactly where you can go. Using very defined, but asthetically pleasing boundaries to signify where you can't go. Rock walls that are seperate meshes, almost verticle, obviously too steep to climb. Broken brick walls, that you might find on the exterior of a building obviously too tall to climb, but with nice chunks missing - allowing you to peek through and see the expansive fields beyond them. Rivers on the boundary, obviously too deep to cross, but still allow a nice view. Wooden fences that feel a little more restrictive, but natural. Basically, every place you are restricted from must have a subtle or obvious hit that it is off limits. The major gamepath must somehow stand out and draw the player into it, without feeling akward or too defined.

B. Develop a series of rock walls, that are large and work as a more detailed middle-ground between you and the large mountain. These would intersect with your buildings, and be the boundary geometry for portions of the map. Then, I would push the mountain back more, to better develop a sense of scale. Also, with it further back, you can cheat the texture by scaling them up alot, making it feel less repetitive.

C. Change the waterfall from a steep angled waterfall, to a full free-fall verticle drop waterfall. I would also push it back more. You can still have the cool bridge going over the creek that has flowed from this waterfall. I just think that the waterfall you have now feels like HL1 sitting in a HL2 world. With a freefall drop, you wont have the akward transition from rock to water that you have now. the waterfall would basically pour from the mountain top, or a hole in the mountain, straight down into a pool. With some good steam effects and more work on the scrolling alpha texture (so it feels more scattered), I think you could get this looking good.

D. Hide the sheep. Sorry, although I really love the sheep model, the fact that they are not animated actually detracts from the map more than the bonus of having them. If nothing else, put them further out, beyond the gamespace, in a pin that is dimly lit. I think you can still get the same atmosphere with the hay, some animal pins, and the sounds of sheep coming from different parts of the map, without actually showing the sheep. It's like putting civilians in a map, but not having them move - it just seems weird, and the downside outweighs plus.

WHew..... ok, now that I've spewed all of that your way, let me just say that I see much potential in this map and the theme. I think the execution may need to be rethought for it to really hit the highest quality mark possible, but I 'get it' and like it. I wish you much luck with this and future maps.