An Au­di­ence With...

From BASIC to VR: Epic’s founder talks us through a life ded­i­cated to videogame tech­nol­ogy


Epic Games’ Tim Sweeney tells us about his begin­nings in games – and his con­cerns about the fu­ture

Epic Games founder Tim Sweeney has the look of the pro­to­typ­i­cal nerd engi­neer, with thick­lensed specs, a ner­vously twitch­ing nose, and a wardrobe full of black hood­ies. Where Sweeney breaks from cliché, how­ever, is in his arch out­spo­ken­ness. “[Mi­crosoft is] sub­vert­ing the rights of devel­op­ers and pub­lish­ers to main­tain a di­rect re­la­tion­ship with their cus­tomers,” he wrote in The Guardian ear­lier this year of what he sees as the com­pany’s anti-con­sumer poli­cies. It’s fight­ing talk from a man who has made mil­lions from Gears Of War, a se­ries pub­lished by Mi­crosoft. Sweeney, how­ever, can af­ford to be bol­shy. His com­pany, which he founded in 1991 at the age of 21, has pro­duced not only some of the best-sell­ing block­busters of the past 20 years, but also Un­real En­gine, the mid­dle­ware that’s been li­censed by other game devel­op­ers to power hun­dreds more.

Where did you grow up?

In Mary­land, on the out­skirts of Po­tomac. I got a computer, an Ap­ple II, when I was 11 years old. I com­pletely fell in love with it. The sheer power of that thing… To see what other peo­ple were build­ing with it and to think that I might be able to do it my­self if I could just learn the com­mands – that was hugely em­pow­er­ing. For the next ten years I spent more time at my computer than I spent at school. I’d say it took about a decade be­fore I felt like I was ac­tu­ally get­ting good at it.

What con­vinced your par­ents to buy you one?

It was around the time of the computer revo­lu­tion and my brother, who is 15 years older than me, was liv­ing in San Diego, where he was work­ing for hi-tech com­pa­nies. I vis­ited him and he showed me his garage, which was full of com­put­ers. He taught me how to pro­gram BASIC while I was stay­ing, over the course of two days. He worked at IBM at the time but, even so, he bought me an Ap­ple II that Christ­mas. Soon enough I found that BASIC was too slow, so I taught my­self ma­chine code and Assem­bler in­stead. I pro­grammed that way for 14 years. I made some pretty ma­jor pro­grams that way – some games and also a basic bul­letin board. I was su­per into bul­letin boards at the time.

What sort of games were you cre­at­ing in those early days?

All kinds of things – prob­a­bly about 50 games in to­tal by the time I switched to a PC. The very first game I made was a Pong clone. I didn’t know how to cre­ate graph­ics so I just used hex sym­bols to form the bats and an as­ter­isk sym­bol for the ball. I started writ­ing an­other game and it was go­ing well, so I in­vited all of the neigh­bour­hood kids over to play it and see what they thought.

How did it work?

It was an ac­tion RPG that used text in­stead of graph­ics. A lit­tle like NetHack, you could say. Any­way, I had all of these kids com­ing over – some were younger than me and some were older than me; they ranged from age five to 25. I watched them play­ing and tweaked things and, over the course of a few weeks, I be­gan to feel like ev­ery­one who played the game could fig­ure out what they were sup­posed to be do­ing and have fun. So I put it out as share­ware. I copied the model that 3D Realms and Apogee used at the time, whereby you give away one episode of the game for free and then peo­ple can buy new episodes via mail or­der.

The model has come full cir­cle nowa­days with on­line episodic games.

Ex­actly. I made about $300 that way and thought to my­self: ‘Oh my god, I’m rich.’ Soon af­ter that I wrote ZZT.

How did you teach your­self graph­ics pro­gram­ming?

I read the doc­u­men­ta­tion and taught my­self that way.

What were you study­ing at col­lege at the time?

Me­chan­i­cal engi­neer­ing at the Uni­ver­sity Of Mary­land. I’d been pro­gram­ming com­put­ers for half my life by that point and I felt like computer sci­ence would have just been a waste of my time. Me­chan­i­cal engi­neer­ing turned out to be a good ma­jor be­cause I got to learn all kinds of math­e­mat­ics that’s al­most im­pos­si­ble to learn on your own. By the time it came to writ­ing 3D for Un­real, it was crit­i­cal. There was like a Karate Kid mo­ment, a day that I re­alised: ‘Oh, wow! I need to rep­re­sent a point in 3D space as a vec­tor. And for a vec­tor to work I need a ma­trix’. I learned all of this seem­ingly use­less math and then it turned out to ac­tu­ally work.

So you grad­u­ated with a me­chan­i­cal engi­neer­ing de­gree – how did you move from that po­si­tion to be­com­ing a full-time game-maker?

I ac­tu­ally didn’t grad­u­ate. Epic was grow­ing so fast, and I was one credit short af­ter four years. I didn’t fol­low through and get the de­gree. I’d put out two games by

that point. But I’d also re­cruited some other peo­ple via bul­letin boards and other early com­mu­ni­ca­tion tools of the In­ter­net like Usenet.

How were you get­ting on­line in those days?

Mainly via the uni­ver­sity net­works. I re­cruited about five other devel­op­ers from around the world in this way, and was help­ing them with their projects as a kind of pro­ducer. I’d pro­vide ad­vice with game bal­ance and so on.

Were you al­ways plan­ning to pub­lish their games?

Yes, I was kind of set­ting up a com­pany, although noth­ing that for­mal, re­ally. I was still work­ing out of my par­ents’ garage. By the time I grad­u­ated col­lege I was spend­ing my time in that garage copy­ing disks and mail­ing them out to peo­ple, while work­ing with these other devel­op­ers vir­tu­ally around the world. It was a very rudi­men­tary pub­lisher setup, you could say. That’s how Cliff Bleszin­ski got in­volved – I met him over the In­ter­net when he was work­ing on an ad­ven­ture game called Dare

To Dream. It wasn’t a very big hit but, af­ter that game, I put Cliff to­gether with a bril­liant pro­gram­mer I knew, Ar­jan Brussee, and they built Jazz Jackrab­bit to­gether. It’s funny: af­ter years of sep­a­ra­tion, they’ve re­cently got back to­gether again.

You’d started out as the some­what dry-sound­ing Po­tomac Computer Sys­tems.

Yes, I did. At the time I had thought I’d be­come a computer con­sul­tant.

Were you called Epic by the point at which you had mul­ti­ple share­ware games com­ing out?

No, in the early days the op­er­a­tion was called Epic Me­gaGames. I thought that made me sound like a se­ri­ous, ma­jor com­pany – some kind of big op­er­a­tion. Once we ac­tu­ally be­came suc­cess­ful, I de­cided to drop the ‘Me­gaGames’ from the com­pany name. It sounded a bit much.

How suc­cess­ful were these games that you were putting out?

At their peak, some of these games were mak­ing sev­er­al­hun­dred dol­lars a day. They more than paid ev­ery­one’s mea­gre sus­te­nance at the time. I was just su­per-happy to be in a po­si­tion where I was able to make a basic liv­ing mak­ing games. We had no cen­tral of­fice or any­thing like that. Ev­ery­one was work­ing from their homes, garages and bed­rooms. It worked well. Then the in­dus­try started chang­ing re­ally quickly. Dur­ing this time, Wolfen­stein 3D came out. Videogames started to grow larger and larger. At some point it be­came clear that I wouldn’t be able to sup­port teams of the nec­es­sary size on a share­ware model. I struck a pub­lish­ing deal with GT In­ter­ac­tive, and we had to re­think the way that we did busi­ness en­tirely. We took all of our best devel­op­ers who were work­ing on smaller projects and put them to­gether to work on a much more am­bi­tious game. None of us had done any­thing in 3D be­fore. I had to learn 3D pro­gram­ming and Cliff had to learn 3D level de­sign.

How many of you were there on staff at the time?

There were ac­tu­ally about six of us, which quickly grew to 25. It was a pe­riod of ma­jor growth. We started with­out know­ing what we were do­ing. About a year into pro­duc­tion we had a playable demo and we started show­ing it around as we needed fund­ing – we needed a pub­lish­ing deal and pub­lic­ity for it. Soon, devel­op­ers started con­tact­ing us say­ing they were in­ter­ested in us­ing our en­gine. I was like: “What’s an en­gine?” That’s when we started to li­cense the use of our tools. It pro­vided us with some in­come that in turn helped to fund the de­vel­op­ment of Un­real. We started in 1995 and shipped in 1998. Three and a half years’ de­vel­op­ment for a game that we never thought would take longer than six months to make.

Did these other devel­op­ers get in touch be­cause they’d seen your en­gine in mag­a­zines?

Yeah, we had good press con­tacts from our share­ware days, so we used those to get pub­lic­ity. We had a cover on PC Gamer, which was re­ally the first le­git­i­mate 3D screen­shot to be used full-page on the cover of a mag­a­zine, I believe.

Were other com­pa­nies try­ing to do a sim­i­lar thing with the busi­ness model at that time?

Id Soft­ware had li­censed the Wolfen­stein en­gine, but en­gines didn’t re­ally ex­ist as a con­cept un­til 3D came along. 2D is so sim­ple that you don’t re­ally need an en­gine to han­dle it. Every­body wrote their own stuff up un­til that point. And then it be­came hy­per-spe­cialised. For about 20 years now, that has been our busi­ness model at Epic. What do you think was the rea­son for Un­real’s suc­cess? Pre­sum­ably the tim­ing helped, in terms of the tech­nol­ogy you’d de­vel­oped. It was about nine months be­fore Half-Life. That was im­por­tant tim­ing. Half-Life com­pletely changed ex­pec­ta­tions of first­per­son shoot­ers – it of­fered a story that you’re a part of. They broke down bar­ri­ers that nei­ther Id nor we had ever re­ally at­tempted be­fore. Un­til that point, the story was ba­si­cally: you’re some badass with a gun shoot­ing at stuff. If we’d have come out af­ter

Half-Life I think peo­ple would’ve said that our game looked pretty but had no sub­stance to it. So in that re­gard we were for­tu­nate.


The neg­a­tive as­pect, how­ever, was that we shipped with com­pletely bro­ken net­work code. We’d only tested the game on our lo­cal net­work. This forced us to work on that as­pect of the game for months af­ter re­lease. We cre­ated an ad­don to fix its prob­lems, with a bunch of new mul­ti­player maps. That grew and grew un­til it be­came a stand­alone prod­uct, Un­real Tour­na­ment, which came out in 1999. That turned out to be a lot big­ger than the orig­i­nal game.

Was that due to the gap that Id Soft­ware left when it went on to try other things? An op­por­tu­nity ap­peared in the mul­ti­player mar­ket?

Yeah. It’s funny. Every com­pany has been founded on core prin­ci­ples. Every step of the in­dus­try’s evo­lu­tion has com­pletely chal­lenged those, and com­pa­nies have to change. The medium went through so many rapid changes while peo­ple were fig­ur­ing out the rules and the bound­aries. The move from 2D to re­al­time 3D graph­ics was seis­mic. Then story’s in­flu­ence. Then on­line mul­ti­player. These ma­jor shifts, con­stantly clos­ing and open­ing op­por­tu­ni­ties for game com­pa­nies.

How did that grow into Gears Of War?

We had a real prob­lem. The In­ter­net be­came vast. There was no Steam at the time, so it be­came re­ally, re­ally con­ve­nient to pi­rate games [on PC]. With­out a ma­jor chan­nel through which to sell games on­line, we were find­ing that, for every one copy of a game we sold, ten more were pi­rated. That ra­tio made it im­pos­si­ble to com­pete with the con­soles, where piracy rates were so much lower. This led us to sign a deal with Mi­crosoft to make Gears Of War. It took about three years. Then, dur­ing that pe­riod, Steam came along. Piracy be­came a lot less con­ve­nient than buy­ing games le­git­i­mately. Mul­ti­player be­came much more im­por­tant. And so PC gaming be­came vi­able once again.

Just prior to Steam’s launch there were end­less ar­ti­cles pro­claim­ing PC gaming dead or dy­ing. It was a very quick turn­around to the land­scape we have to­day. Was Un­real En­gine li­cens­ing the core of your busi­ness by that point in terms of rev­enue?

It bounced back and forth. In years when we haven’t had a ma­jor game on the mar­ket, the en­gine has been the num­ber one rev­enue earner. It’s sta­bilised us and has en­abled us to fund the com­pany when, with­out the en­gine, we might other­wise have gone un­der.

At what point did you de­cide to change the busi­ness model for your en­gine li­cence to al­low com­pa­nies to use the tech for free? Was that mainly in re­sponse to Unity, where you hadn’t been chal­lenged in the mar­ket­place be­fore?

It just shows the value of com­pe­ti­tion. We’d got­ten fat on the li­cens­ing model with block­buster devel­op­ers. The indie revo­lu­tion re­ally started with mo­bile app stores. Unity hap­pened to be at the cen­tre of that. Sud­denly every­body was us­ing its en­gine for these in­de­pen­dent mo­bile projects. That re­ally democra­tised ac­cess to game devel­op­ers. We’d al­ways nur­tured a com­mu­nity around

Un­real Tour­na­ment mod­ding, but it wasn’t re­ally a com­mer­cial thing; it was more un­der the um­brella of the game. Whereas Unity had the same con­cept: you could use this en­gine for any­thing, in­clud­ing mod­ding other prod­ucts. That forced us to open up our en­gine to every­body. Our deal is straight­for­ward now: for three years af­ter the game’s re­lease, we take a five per cent roy­alty on rev­enues. That way we’re not a tax on peo­ple’s in­come be­fore they make a profit – we only make a profit if the game’s suc­cess­ful. It’s also been cru­cial to open up the en­gine to cus­tomi­sa­tion by mak­ing it open source. In block­buster game de­vel­op­ment, the first 80 per cent of tech­ni­cal de­vel­op­ment is straight­for­ward; later, you re­ally need to have the op­por­tu­nity to be able to go in and mod­ify parts of the en­gine.

You’ve re­cently made highly crit­i­cal state­ments about Univer­sal Win­dows Plat­form. It’s un­usual for a high­pro­file fig­ure to at­tack a high-pro­file com­pany in such a pub­lic way. What ag­i­tated you to that ex­tent?

I could see that some­thing very bad was hap­pen­ing

Sweeney’s MS-DOS ti­tle ZZT was re­leased in 1991, just prior to his game com­pany’s trans­for­ma­tion from Po­tomac Computer Sys­tems to the less mod­est-sound­ing Epic Me­gaGames

Like his one-time com­peti­tor John Car­mack, Sweeney is com­mit­ted to VR: Epic’s Bul­let Train demo was con­ceived as a show­case for Un­real En­gine 4 in the con­text of an Ocu­lus Rift shooter

In April 1998, Un­real saw Sweeney’s coding skills un­leashed in the 3D ac­tion game arena for the first time. It was, the pro­gram­mer says, good tim­ing

Newspapers in English

Newspapers from Australia

© PressReader. All rights reserved.