The Making Of: Dust 2
Despite the success and overwhelming popularity of Dust, the thought of making a sequel took a long time to cross my mind. Although considered by some to be it's spiritual successor, Cobble failed to find the same audience, and there was clear demand for a 'real' sequel. Dust 2 didn't arrive until March 2001 - nearly 2 years after the original.
It was never going to be easy creating a successor to the most-played FPS map in the world. Even with the benefit of the same texture set, an established theme and innumerable number of salivating CS players, it was an incredibly daunting task. I didn't want to call it "Dust 2" for this very reason. Instead, on the basis that the third instalment of any movie trilogy is typically never as good as the first one, I decided to call it "Dust 3", and hoped no-one would notice. My assumption was that everyone would just carry on playing the original.
So, I opened up Hammer, and started laying out "Dust 3".
Sticking to a Theme
First things first, I had to ensure that this new map had everything in common with Dust, without actually being Dust. I had to identify the elements that defined a Dust map.
Perhaps the most iconic aspect of Dust are the arches that separate main gameplay areas. Aesthetically they neatly compartmentalise the map, framing entrances and exits between areas and helping players create a mental picture of the layout. From a gameplay perspective, they produce tight choke-points between objectives, and block out unwanted lines-of-sight. Consider how Dust would have looked without the arches:
It's worth reading the original Making of Dust if you haven't already, as you'll see the arches (amongst many other elements) were lifted directly from Team Fortress 2.
The stone roads through Dust were like arteries, giving players clear paths to follow towards the main objective areas. The sequel would do the same, connecting spawn points to the primary objectives.
As an example of the importance of these paths, here's what Dust would have looked like without them:
The much-abused 'trim' along all the walls and ceilings of Dust was perhaps the single element that gave the world a hint of relief and purpose, and helped separate distinct geometric elements. It was incredibly important in a world that was almost otherwise entirely made of the same colour stone.
I therefore tried to use the trim very carefully, only exactly where needed, and not just as filler. I didn't want to repeat the mistake of many Dust-alikes where it had been peppered everywhere it fitted, flooding the visual cortex of players with the excessive complexity. Neither did I want to under-employ it, and end up with a world looking flat and barren.
I tried to formalise the rules I'd employed. The trim would never appear on a floor or ceiling, or anywhere a player could stand on it. It would never be striped or tiled vertically across a surface. Finally, for any given flat wall, it would never appear more than twice (e.g. at the top, and at the bottom, but never in the middle as well.)
This was perhaps the element of Dust that underwent the most tweaking - every BETA release of Counter-Strike saw it moved ever so slightly, lengthening the shadows each time. It was key to making Dust distinct amongst the other maps in rotation, and keeping it accessible and pleasant to play in. I considered giving the sequel a night setting, but in doing so would have removed a vital element of the Dust theme. The best I could do was make it exactly the same.
This new Dust had to bear resemblance to the original beyond the texture set and sunlight. It had to copy important design elements that players had become accustomed to: the 'Dust doors', simple structures, ramps, crates and arbitrarily raised concrete areas. In fact, I'd have to copy some areas nearly verbatim to ensure the lineage was present.
At the same time, while trying to differentiate the sequel from the original, I'd have to be conservative about any new elements I was introducing. These came in the form of the half-spiral staircase, and the rock face adjoining the two bomb spot areas.
There were also two distinct boxes that Dust 3 needed to tick - a place for close-quarters battles, and a place for long, drawn-out AWP fights. Dust had the central corridor and the underpass to fulfil these requirements, and I'd have to create something similar in the sequel to cater for both types of player.
It was a lot to juggle.
Keeping It Simple
Of course, the most important element that would make or break the sequel was the overall map layout. It had to be Dust-ish, without being Dust, different enough to give players a fresh challenge, but maintaining the balance and pace of the original.
I had already spent a long time trying to work out what vital layout ingredients had Dust tick, and reached the conclusion that the simplicity and concise connectivity were key. In it's most basic form, Dust was little more than a figure-of-eight that had grown a pair of arms and legs, centralising the battles but providing tactical wiggle room.
To maximise the sequel's chances, to make it feel like Dust and play as well as Dust, I'd have to adopt this same structure.
Unusually, I opted to draw out some doodles, rather than just diving in as I typically would. This sequel was important - it deserved careful, considered, detailed and thorough planning, a complicated process that would eventually span literally minutes.
One of the most peculiar decisions I took at this stage was to add rock to the Dust theme. The textures had always been there, but I had never used them. It was a (miniscule) risk to adopt them now, knowing the challenges of creating believable rockery with the Half-Life engine, but I took it anyway. I needn't have worried.
The Early Dust 3
It took a couple of days to convert the paper design into a playable map, although thankfully was a relatively straightforward process. Through habit, I strayed a bit from the paper designs, but maintained the overall layout that I intended, correcting the scale as I went along.
The biggest problem was dealing the relative proportions of gameplay areas. I had inadvertently built the map and filled in the details in such a way that I was now hitting the very edge of the permitted space allowed by Worldcraft. But, rather than just moving the entire map a few thousand units in the opposite direction, I stuck with it, and that's how the Terrorist spawn area ended up so long in the final revision.
The alpha version of Dust 3 was pretty close to the paper design, but lacked any character. While it was clearly a Dust map, and had both old and new Dust elements, it was far from being finished.
You can see in the shots below that the area I was most unhappy about was the second bomb spot. It only had a couple of entrances and a few crates for cover (just like the original), but clearly needed something more. The road leading from it was also a straightforward junction. All these elements were clearly derived from Dust, but detrimentally so.
Around this time I invited Gearbox's Brian Martel to offer some feedback and advice. He subsequently suggested adding arrows pointing toward bomb locations to support the existing floor markings, features that would radically improve the accessibility of the map (and were later introduced into the original Dust and indeed all other official maps.)
Heading to beta, I massaged the bomb site with Dust's bomb site B in mind, adding a raised platform and shuffling more crates in to break the space up and offer more gameplay opportunities.
In the process, I managed to find a way to open it up by creating a hole in the wall connecting it to the junction to bomb site A. I considered this risky - Dust had no such elements, and I wasn't sure quite how it would look or indeed play, but it 'felt' right, and the objective became a far more comfortable area to navigate. I pictured how much fun I'd have armed with a Steyr Scout perched as lookout for enemies trying to breach the gates.
I also took some liberties with the crates, which were sitting nervously at awkward angles, as if the laws of physics had tumbled them around a little. Again, I was wary that this was a break from the rigid, grid-like nature of Dust's crate arrangements, such was my concern that such minor elements had been what made Dust a success.
The staircase was also unnerving. Dust never had stairs of more than a couple of steps, and this map was set to have a half-circular staircase, and in a tight area too. However, I needed some passage between those two areas, and the staircase seemed to fit and provide some interesting conflict.
Despite these changes, this map had some familiar elements too. In fact, some elements were ripped nearly verbatim from the original.
The ramp pictured above always seemed particularly 'Dusty' to me, perhaps because it harked back to the TF2 screenshots that inspired Dust.
Again, notice the similarity - two ramps which both head down into darkness. In Dust 3, the T's would have started at the higher end of the ramp. This later changed to the CT's in the underpass in Dust 2.
This had all been happening in relative secrecy. I had been intentionally keeping it quiet, not wanting to raise expectations, assuming that once it launched on a new website called GameHelper that would be the end of it. I was convinced that compared to Dust, "Dust 3" would be a failure, despite my efforts.
Alas, it was not to be. Jess insisted it became part of the official Counter-Strike rotation, except under the far more sensible name of "Dust 2". As with Dust, Jess was vital to getting the final balance of the map just right, suggesting new spawn and bomb locations.
Dust 2 was released in March 2001, as part of Counter-Strike 1.1. Unlike Dust, it received no further layout tweaks after that.
Just like Dust, Dust 2 excused itself from my pessimism and became incredibly popular, not just on public servers, but in clan matches too, which Dust had become ill-suited for. Dust 2 later also stole the "most-played map" crown, sharing it only occasionally with Aztec as Dust fell down the charts to newer and better maps. I had never expected Dust 2 to compete with Dust for longer than a few weeks, but like Dust, "Dust 2 24/7" servers sprang up everywhere and it's popularity meant it earned places in later versions and ports of Counter-Strike.
Counter-Strike: Condition Zero
Like Dust 1, this version of Dust 2 shares much in common with the original, and is largely based on the original brushwork. The map is notably more vivid and strong in its appearance, with an overall more colourful appearance helped by the additional detail. A few changes have been sneaked in, such as crate placement around bomb sites, but otherwise the map serves as a nicer updated version of the original.
This version was primarily worked on by Ritual, although some finishing touches were added by Valve just before release of the game.
This Dust featured a whole host of improvements, from the improved skybox (it really does feel like the middle of a desert now), improved building structures, additional detail, village clutter, as well as slight changes in the layout of the bomb sites to improve the game.
From the overview above it's clear that there were a number of small changes to map proportions versus the one that shipped in CS 1.1. Various passageways have been widened and made more accessible, while others have been lengthened or squared off to improve gameplay and rendering performance. The additional detail also cuts of some lines-of-sight and adds cover where there was none before, but otherwise it's reasonably faithful to the original.
This renovation of Dust 2 was performed by Valve following the renovation of the original Dust.
Counter-Strike: Global Offensive
Far and away the most beautiful version of the map ever created courtesy of Valve's crack team of artists and designers.
Dust 2 has also made appearances in other games, courtesy of the modding community. One of my favourites is the Dust 2 conversion for Far Cry 3, showing the mapping process from start to finish:
There's also the obligatory Minecraft version.
As with Dust, none of this would have been possible without the help of Joe Markert of GameHelper, Jess Cliffe, Minh Le, Chris Ashton, Brian Martel, Richard Gray, Kristen Perry, Ido Magal, and all sorts of other people.
Also, of course, ultimately I have to thank Valve Software for the idea I stole in the first place. Twice.
This work is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.