Haven't read the whole thread, but will just chime in with some factual references relating to how things are in game (cuz a lot of people are maing references to things that are not factual wrt how our code and the BF2 engine works).
As a point of general context, y'all have to keep in mind that we're trying to inject realism into the BF2 engine which is inherently more of an arcade and fun oriented game than a simulation engine. So this has some obvious challenges and as we inject realism into one area, typically we discover problems that creates in another area.
Again at a more sort of "directional" level is our intent to use all of the assets of the BF franchise. By this I mean tanks, APCs, helicopters, jets, etc. We've spent a lot of time trying to make these assets and their weapon systems perform similarly to their real world counterparts. This means that they are incredibly lethal, but still very vulnerable because .. well .. military budgets have taken care of a lot of aspects of game balancing for us.
But in order to utilize these assets effectively at the level of fidelity we're trying to inject, we need larger maps and more room for engagements. Kashan Desert is the first all terrain combined arms map of it's size ever released in a BF product or mod. We didn't even know if it was going to work from a map size perspective, never mind taking into account that we're trying to model 1km view distances.
As we figure out more things we can do there and more ways to optomise performance and as computers get upgraded in horsepower, we'll hopefully be able to add in more terrain and static objects features into these large maps than what we have in Kashan today.
I am quoting this part because it's a deviation from the thread but explains a lot of what turns my crank about PR:
The original "Falcon" game was released by Spectrum Holobyte in 1987 for microcomputers before many of you were born. This was a game that.. heh.. in some ways changed my life. It made me discover the concept of networked computers (at the time via a null modem cable).
In 1989 Microprose released a game called M1 Tank Platoon which, in it's day, was a stellar Tank sim. Eventually Microprose and Spectrum Holobyte merged together and geeks like me thought that "any day now" we'd have a combined M1 Tank Platoon and Falcon game world.
As Falcon matured into Falcon 3.0, Spectrum Holobyte promised an A-10 Warthog add-on and began talking more and more about their concept of an Electronic Battlefield. The vision for this was high fidelity simulations all connecting to each other to form a virtual world of electronic combat where the only thing that gets hurt is your *** and your marriage.
Alas none of this materialised. They even ran advertisements for the A10 product that never made it to market. The closest thing to an electronic battlefield has been WWIIOL which .. um .. well.. if this was 1999 might look OK .. but it ain't 1999.
I've always wanted to see a high fidelity combined arms virtual warfare simulator come to fruition.
Hence my interest in BF2 / PR and the hopes that somehow we can create some sort of concept along the lines of an Electronic Battlefield.
wrt the injection of realism .. case in point is the accuracy and lethality of the small arms. We've spent a lot of time optomising these for long range engagements through v0.6 development. We used 1km engagement ranges as a benchmark of the upper end of the engagement range scale. That more than covers the effective range of the assault rifles modelled in game.
wrt iron sight zoom. there is a SLIGHT amount of zoom when you raise the weapon to iron sights view. We're talking 1.1x zoom, effectively immaterial, just a minor "focusing" of your FoV to reflect staring down the iron sights.
wrt the zoom levels of the Rifleman weapons, we've tried to match that to the realistic zoom levels found on those weapons in real life. Most rifleman sights are 4x zoom.
wrt accuracy .. well... that's a little more complex. the BF2 engine has a number of parameters we can work with to model accuracy, but that model is not one based on real world physics, rather it's one based on being "good enough" to sell millions of copies of a game.
For example the "bullet drop" modelled in BF2 is linear. In real life it's exponential. ASCII art:
BF2 bullet drop:
-
.-
..-
...-
Real Life bullet drop:
---
....-
wrt the incongruence this creates and what we need to look at doing about it... I don't believe the right approach is to "nerf" the realism out of one area. In contrast I believe the right approach is to raise the levels of realism in the areas that are deficient.
The main issues in my view are:
- view distances
... if you can watch troop movement from 800m, you can manoeuvre to engage / avoid / ambush. This has a lot more survivability than a guy popping out of the fog at 250m.
- cover and concealment
... if I can see grass in front of me, the guy 800m away should also see grass in front of me. This is a major issue we need to resolve.
... real life terrain has a lot more places for hiding the human figure. We need to resolve this.
... camouflage. irl you are generally looking for signs of movement to locate an enemy. in BF2 / PR you are looking for the tell tale colour swatches of infantry uniforms and player models. Most modern camoflauge appears very well blended into the surrounding terrain at anything beyond a couple hundred meters.
... lighting. similar to above... real life has a lot of subtlety to the lighting that makes hiding a lot easier than it does in the game engine. Generally I think soldier bodies are too reflective of light and many maps have areas that are too brightly lit (for example forests are quite dark even in broad daylight).
I am sure there is other stuff.
So erm.. long babble short... accuracy may be an issue we need to evaluate, but I think there are other issues to evaluate that can raise the level of realism and achieve the same end goals around game play dynamics.