The Making Of: Dust

Introduction

Dust is the work for which I am best known. I wish it was due to my brilliance, intelligence, level design knack and astonishingly good looks, but actually, the main element was luck. It was all the right bits in the right place at the right time. The result far exceeded anything I could have imagined or planned to make - Dust just 'happened'.

Some say the reason for Dust's success was its simplicity. Others say it's due to how balanced it is. In reality, the two are closely linked - the simpler a map is, the easier it is to balance. In essence, it's hard to make a simple map play badly.

Or perhaps Dust is popular for another reason entirely - because, compared to the other maps which shipped with Counter-Strike at the time, Dust was very different. It was bright. It was open. It felt comfortable and appeasing to play in. New players found it welcoming.

This article covers where Dust came from - my inspiration, my various decisions, contributions from other people, and how different it could have been. Much is from memory and so many of the images are reconstructions, and some parts of the process have been left out. For this reason I won't cover the whole process, but all the most memorable and important moments are.

The Origins

First things first: I stole Dust from Valve. Anyone who had seen the Team Fortress 2 (TF2) screen shots and videos from around 1999/2000 may have noticed some similarities with Dust. For example, consider the following screen shots:

Early Team Fortress 2 screenshots were the inspiration for Dust.

It was those screen shots which established what I wanted to create. I wanted to play TF2 so desperately, and knew it wouldn't be out for a long time, so it felt right to use my hobby as a method to get just a little bit closer. I knew I could recreate what I had seen in the screenshots - it wasn't anything particularly complicated. My aim was simple - to recreate this TF2 map without it looking too obvious.

The first thing I required were textures. I had already decided this was going to be a CS hostage rescue map, and so cliffe (Jess Cliffe) put me in contact with MacMan (Chris Ashton) who had inadvertently made the textures for my previous CS map. My request was simple - create textures like those I had outlined in this shot:

Highlighting needed textures

Chris did a fantastic job and returned textures far exceeding anything I had expected. This was the first batch of custom textures I had ever had custom made for me, so it was an extremely exciting time and when I remember being absolutely amazed when I saw them.

I had been hoping for replicas of the textures I had outlined in the shot, but as it turned out, I much preferred these ones. I saved them in a texture file called 'cs_dest' - at the time, I had called the map I was going to build 'Destiny'. It would have been 'desert' but that had been done already...

The Beginning (counter-terrorist start)

I had a clear idea in my head describing to me how one area of Destiny would look - it was the exact same area as in the TF2 shots. There were three main features in the TF2 screen shots (repeated below) that I'd have to copy:

• The stone path 'road'. I determined at this time that the road would need to travel all around the map - from one team spawn to the other.
• The stone arches - again, this would need to feature predominantly.
• The stone trims.

Combined, it was the above three features that would define the overall look and style of the map.

Early Team Fortress 2 screenshots (again)

The map started out quite simple - it was essentially a straight piece of road with a couple of buildings and walls, plus the arch. It wasn't quite the same as the TF2 images, but it was a start.

Early Dust screenshots based on TF2

The House

See that building on the left? It would have made a lovely little 'house' from which the area could have been defended. In the worst possible way to build a map, I had no idea what this area was going to be for - whether it would be central, or team-oriented. The building just called for windows. They could be sniped from, and the entrance would be from behind - from that hall I made. It would be ideal.

Some initial ideas including the 'hut' and a dubious light-fitting.

Ultimately however, they didn't provide the view of the entrance that I wanted, and the windows just didn't seem to fit in with the architecture at all. However, one by-product of the room was lighting - I wasn't able to fake light flooding into the room realistically, so I let reality take a back seat and put an electrical light in. It lighted the room well, although it wasn't ideal, and I was going to remove the room anyway - so I removed it. Little did I realise that it would return to light the darker parts of Dust.

How It Compares

Side-by-side, the amount that I stole from TF2 is really rather obvious:

Note the archway in the centre and the buildings on the left.

The same buildings are on the right.

What surpriss me now is how ugly Dust is, even with the advantage of being able to see the TF2 screen shots and improve upon them, Dust still looks empty and bland. If I were to remake Dust today using the TF2 influence, it would certainly bear more resemblance to the original and look less 'basic'.

The Arch

The arched doorways of Dust have become a hallmark in the Dust theme - no Dust map can possibly not have at least a few of those arches between areas. They differed slightly from those I saw in the TF2 screen shots by having a rounded top and vertical sides, as opposed to the 'curved half diamond' shape that TF2 exhibited. However, you'll notice the arches are always slightly extruded from the wall to which they're attached - something I took directly from the TF2 arches. I remember at the time trying to work out how the brushwork would be done, and considering it something of a challenge to recreate.

The Middle

After the success of the first area, the middle area had to follow. Unfortunately, I was inexperienced and found this the hardest part of all. It took many iterations before it reached the 'underpass and hallway' formation that it has today - many of which were never fully formed.

However there was one design which has stayed with me:

Dust's original underground section

This room was at the end of the road from the underpass. It was a departure from the style that Destiny was taking - for a start, it was indoors, for the first time a real room. Fortunately, it had no real purpose and I removed it. Had it stayed, I doubt it would have been accepted into the official line-up, and it certainly wouldn't have played right. I believe it would have been less of a departure from the other maps as Dust turned out to be. The only similar outdoor map at the time was Desert, and that was set at night so it had its dark areas.

I believe the main reason I removed and tried again was because it just provided no opportunities for expansion. At the time I designed levels without extensive planning - I took the map where I felt like taking it. If I built a bit which I couldn't work out how to extend, I'd scrap it and try again. That's what happened with the room - and it was just as well.

Timing

Part of building the centre of Destiny involved me running from each side of the map into the centre and seeing how long it took. The idea was to ensure each team would meet in the middle at approximately the same time, to focus the battle there. I like to think this was quite an obvious thing to do, but I'm quite proud of it. I've done it ever since on my CS maps to get some idea of where the conflict will be focussed.

The End

The rest of the map was built just like that described above - bit by bit, as my imagination took me. Of course, I had a slight idea of what I needed - like the 'ring' at the terrorist start. This had to lead into the underpass and into the hallway - that was certain. I also had another route leading off which provided nothing, and so in the final version of Destiny, blocked it off with crates, and still it remains in Dust.

Bomb site Placement

It was only once I finished the layout of Destiny that I thought about the bomb locations. Originally, there was one by the hallway (bomb site A as it is now, but closer), and another one underneath the underpass. Cliffe suggested (not sure if it was before testing or after - before I believe) that the latter bomb site should be moved to the CT spawn. That I did, and there it remained.

Naming

It was only during the testing phase that Dust was named. Destiny had always been a terrible name, but I had not been able to think of an alternative. Given the filename was 'de_dest', I think it provided some inspiration to rename it to 'de_dust'. 'Dust' obviously fitted the map far better.

The BETA Years

While CS was still in its beta stages, Dust had a remarkable ability to change with almost every new release. Often these changes were somewhat experimental, more often they were aesthetic changes, and sometimes they changed the game play entirely beyond what was expected.

BETA 4

Dust first appeared in BETA 4, as one of the maps to debut the new 'defusion' game-type that had been implemented. Comparing the BETA 4 version of Dust to that of CS 1.5's shows some considerable differences - it certainly feels different when I load it up. In CS 1.5 for example, Dust has many more crates than it did originally - many of which were added in the BETA stages to provide either team with more cover and defensive/offensive strategies. Aesthetically, CS 1.5 is far 'cleaner' and the sunlight feels nicer - it's a bit dimmer and at a different angle.

Comparison between BETA 4 and CS 1.5. Note the crates have moved to provide cover.

Furhter comparison between BETA 4 and CS 1.5 showing the change in sunlight and sky.

BETA 5

The BETA 5 version of Dust wasn't that much different to the original. Mostly it consisted of aesthetic changes. One of the most fondly remembered by various players, including myself, was the 'break' in the wall of the CT's sniper nest. I took part of the wall out with the intention of giving the Terrorists some increased chance of hitting a sniping CT. However, the increase view area gave the CTs more of an advantage than I anticipated. The wall was later fixed again.

Features of BETA 5 - the sniper nest cutout and the sparse underpass.

The underpass is also worth examining - in the early stages, there was a significant lack of cover for any advancing players, which did mean the underpass was treated by many as minefield. Crates were later added to add plenty more cover.

BETA 6.1

Perhaps the change I made to Dust which upset/overjoyed the largest number of players was in BETA 6.1. I had been playing with the timing - when each of the teams would meet in the middle, and who would reach the main bomb site first. At that time I felt that the CTs were perhaps a little overpowered by the Terrorists, who could easily rush the primary bomb site, plant the bomb, and win. To counter this, I moved the CT spawn positions closer to the centre of the map - a distance which, when running, took about 2 seconds to cover - so the CTs would arrive at the hallway 2 seconds earlier than before.

Once 6.1 was released, the effects of the change became quite apparent. It was now easier for the CTs to hold down the hallway, and harder for the Terrorists to rush the bomb site, as intended. However, the balance was now too far in the CTs favour. I recieved many complaints about this at the time, as well as one or two compliments. Ultimately, it was obvious that the spawn positions from the previous Dust were fairer, and so in 6.5 the CTs lost their 2 second head-start. The spawn positions have remained the same ever since.

Rejected Ideas

Dust has remained the same ever since CS 1.0 - which is both a reflection on how popular it is, and how little I want to change it (since there's never a guarantee it will improve). I also don't like changing it because I consider it unfair to the new maps - which is admittedly slightly childish, but that's the way I like it.

However, there was one major change that I almost made at CS 1.1 time which would have changed the dynamics of the entire map and would have destroyed many proven strategies. A new route between the two central hubs - the hallway, and the underpass - a staircase connecting them.

Further ideas, including stairs between the two levels.

The consequences of this change would have been quite remarkable in terms of game play. The change would have given Dust a little more variety perhaps, but ultimately - due to the age and acceptance of the basic layout - it would require more balancing work than perhaps one would imagine. It would certainly present new challenges to the player.

I daren't think how Dust might have played with the new route. Looking at it, it would give the CT's an advantage if the T's didn't defend the underpass well enough, but traditionally the hallway needs the most defence, seeing how many players liken the underpass to a minefield. The dynamics of the map would have changed too much, and so I didn't include it. Had I tried, I'd would probably have met plenty of resistance.

Conclusion

Today Dust is - according to the statistics - one of the most played maps in the world, and certainly racks up a remarkable amount of multi player time. It seems remarkable considering how random the development of it was, and how little planning I put into it - which is both a comforting thought and a slightly concerning one.

There is no indication of how well the map would have been taken if some of the changes mentioned in this article remained - for example, the bomb site in the underpass, the sniper house, or the stairs between the underpass and the hallway. If I were to add them today, I would probably receive a lot of criticism, but probably also some appreciation from people who feel Dust is old and tired. I have no intention to make any changes to Dust in the future except those which are required by technology changes and/or bugs in the map.

If I were to make Dust for a next-generation engine, I'm not sure what I'd change. Certainly, the layout would need to stay exactly as it is now - change that and it is no longer Dust. The only real changes would be aesthetic - the various windows could do with some indentation - maybe some of them could be open, with the wooden shutters just swaying gently in the wind. The sand would be more uneven. Small things.

On a large scale, some more passive detail in the background distance might be good - some additional buildings far behind the perimeter walls to give some impression of vastness, and to reiterate that this is just part of a small town. Perhaps some mountains, as currently taken care of by the environment map. Gusts of wind whipping up sand would be nice but perhaps too distracting. More 'village clutter' scattered around the place would help secure the idea that people live in this village - but it would have to be out of the way and not affect the game play. The basic architecture would need some improving - and that would be the most difficult of all, given the basic layout which I'd need to adhere to.

Ultimately, Dust is a peculiar map - it's popular, despite a terrible development method. I attribute it's success to luck more than skill - which, being inexperienced, I lacked. Perhaps it was its perceived originality. Whatever it was, it worked.

Addendum: Modern Dust

It is perhaps slightly worrying that if Dust were a child, it would be of an age able to string complete sentences together, and even read. Since I wrote this piece originally, Dust has moved on from Counter-Strike and was graphically renovated in both Counter-Strike: Condition Zero and Counter-Strike: Source.

Counter-Strike: Condition Zero

This version of Dust 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.

Dust as it appears in Counter-Strike: Condition Zero.

This version was primarily worked on by Ritual, although some finishing touches were added by Valve just before release of the game. The Levelord was chiefly responsible for updating the map and sneaking in all the phallic references.

Counter-Strike: Source

This represents how Dust looks today, and is easily the most realistic Dust ever, representing the theme and ambience of the location far better than any earlier version. Spookily, it more fairly accurately reflects the conclusion I reached earlier:

If I were to make Dust for a next-generation engine, I'm not sure what I'd change. Certainly, the layout would need to stay exactly as it is now - change that and it is no longer Dust. The only real changes would be aesthetic - the various windows could do with some indentation - maybe some of them could be open, with the wooden shutters just swaying gently in the wind. The sand would be more uneven. Small things.

On a large scale, some more passive detail in the background distance might be good - some additional buildings far behind the perimeter walls to give some impression of vastness, and to reiterate that this is just part of a small town. Perhaps some mountains, as currently taken care of by the environment map. Gusts of wind whipping up sand would be nice but perhaps too distracting. More 'village clutter' scattered around the place would help secure the idea that people live in this village - but it would have to be out of the way and not affect the game play. The basic architecture would need some improving - and that would be the most difficult of all, given the basic layout which I'd need to adhere to.

This Dust features 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 bombsites to improve the game.

Dust as it appears in Counter-Strike: Source.

This renovation of Dust was done at Valve, primarily by Kristen Perry and Ido Magal who were given the unenviable job of determining appropriate architectural references for Dust, working on top of the Condition Zero version to give the map some sorely-needed life. My response to seeing the work they had done (given that I was unaware of it) was along the lines of a severely elongated "Wow".

Conclusion 2

To this day I am still amazed that such a simple map could go as far as it has, but a lot of people have helped along the way from Jess Cliffe and Chris Ashton at the very beginning, through to Brian Martel, Richard Gray, Kristen Perry, Ido Magal, and a whole plethora of other behind-the-scenes workers. I haven't touched the map for several years now, and it keeps marching on.

At any point in time, Dust (1) is host to around 10,000 players, around 5% of all online gamers (source). In its lifetime it has been host to millions of matches, responsible for billions of kills, and just won't stop.

Addendum: Dust Nutters

Some stuff just defies explanation.

• de_dust sandcastle (of unknown origin)
• Dust crates in real-life (YouTube)