In the fall of 1991, a heated debate erupted on the technical group's electronic mail system about how best to protect Release 12 from piracy, it having been determined that patching one bit in Release 11 enabled it to run without the hardware lock. As the debate spiraled upward into ``if we do this and they do that and then we do this and they respond in this manner, well, then we'll do the other thing to keep them from sneaking around it by...,'' I thought it was high time to step back and consider just why people pay for software at all rather than stealing it, especially since many of the most highly profitable contemporary software packages contained no protection whatsoever against unauthorised use. My hidden agenda was to get people thinking about what benefits Autodesk could provide our legitimate customers which could be easily denied to pirates and which might, in the mind of a moral mugwump, tilt the balance toward buying a legal copy.
In this letter I rolled out the idea of an ``Autodesk Global Village'' network of bulletin board systems to a broad audience for the first time--I'd been discussing it various people in private E-mail since for a month or so.
by John Walker -- October 14, 1991
Let me start by establishing some terminology for discussing the problem and potential solutions. When faced with the possibility of something nasty being done to you by somebody else, there are basically three approaches you can take to avert the unpleasantness.
Defence consists of active measures intended to make the event you're worried about impossible or at least unlikely to succeed. ``If you try to hold up my store, you'll have to contend with the .45 Auto I keep in the cash drawer''. Or, ``Go ahead, invade; we have three times the troops and twice the tanks, and we'd like the opportunity to even the score with you.''
Deterrence is a strategy based on persuading the opponent that doing whatever you're trying to prevent is a really dumb idea. Defence and deterrence are usually interrelated since a convincing defence is an excellent deterrent, but sometimes deterrence can be achieved without mounting a full defence. The ``massive retaliation'' strategy of NATO in the 1950's is a classic example: ``Cross the border and we'll bomb you back to the Stone Age.'' Bombers and missiles may seem expensive, but they're dirt cheap compared to the multi-million-man army it would take to actually defeat an invasion. Closer to home, consider radar traps on the highways. If you never know when you may be nailed for speeding, you're less likely to drive fast. At least that's the theory.
Incentive strategies attempt to tilt the balance your potential adversary uses to weigh alternatives. Certainly knowledge of an effective defence or belief in a credible deterrent creates incentives to think good thoughts and do good deeds, but incentives needn't be the products of negative, defensive measures. A country enmeshed in a web of trading relationships with other countries, commerce on which its own prosperity as well as that of its trading partners is based, is far less likely to ponder aggression against them since doing so would impoverish itself.
The three strategies are ``nested'' in the sense that defence always creates a deterrent and deterrence creates incentives, but it doesn't go the other way. Incentives needn't involve deterrence or defence at all, nor need a deterrent necessarily be based on a defence capable of defeating the adversary. In terms of cost, defence is the most expensive, deterrence less costly, and pure incentives often have a negative cost--if you create incentives by cooperation and mutual benefit, you generate additional wealth for both parties.
I believe there's a tendency to focus on defence and deterrence and neglect both the value and the efficacy of incentives in inducing the desired behaviour. Think about it. Fundamentally, why doesn't everybody go around holding up gas stations? Is it because all the guys at the pumps are armed? No. In practice, very few are. Is it because the cops might be lurking behind the Pepsi machine, or you're afraid the tireless minions of the law will track you down and drag you to justice? Perhaps, but not very likely. Basically, I think the reason you and I don't hold up gas stations, or at least don't do it any more, is because we'd rather live in a world where we can go to the gas station and fill up the tank in peace rather than live in something resembling an armed camp. Consider red lights. Do you really stop at red lights because you're afraid a flic is hiding behind a tree, or because you'd rather live in a world where you can drive through a green light without slowing to a creep and looking left and right for fear somebody who didn't stop at the red?
What about the guys who do hold up gas stations, roar through red lights, and, for that matter, invade neighbouring countries? Have they convinced themselves that the defence can't stop them? Almost always. Is it because they aren't deterred by what might happen if they try? Sure. But is that why they do it? I don't think so. They do it because they believe that it's worth trying; that what they stand to gain outweighs what they risk losing, all things considered. In such a situation, it's often wiser to try to work on the incentives rather than getting stuck on defence and deterrence. Consider the following alternatives a convenience store chain might adopt when faced with a problem of frequent hold-ups:
I would argue that item (3), once it had seeped through the thick skulls of the Beagle Boys, would do more to reduce the incidence of stickups than any of the other means.
What does all this mean for Autodesk and our problem of software piracy?
First, let's look at the different strategies we have adopted and can adopt to the end of preventing the theft of our products.
The hardware lock is our ultimate line of defence. We deploy it on the front lines of piracy and its mission is as simple and clear as that of a tank--to prevent any and all potential malefactors' running extra copies of AutoCAD by main force--by making it impossible. Like most kinds of defence, the hardware lock is expensive, virtually doubling the cost of goods for every locked copy we sell and burdening us with additional development, QA, manufacturing, and product support costs which are difficult to calculate but certainly run into the millions of dollars a year. Like the cost of maintaining an army, we justify these expenses by arguing that, however high, they're still less than the cost of risking the alternative--we'd lose even more in sales if we abolished the lock. Other copy protection schemes also constitute defence. And like defence in the modern world, there are never-ending promises of ``cheap, guaranteed effective'' technological fixes which never seem to pan out in practice or only trigger a costly game of technological leapfrog between the developers of defences and those attempting to circumvent them.
We exercise deterrence by such measures as serialisation, personalisation (in Release 11 and afterward), and by the efforts of the AutoCops and their brethren and sistren abroad in tracking down and bringing to justice those who use our product without paying for it. This is deterrence in the purest form; nothing prevents you installing a US-domestic version on 10 machines, or purchasing a copy of DONKEY to patch the AutoCAD executable to allow it to run without a hardware lock. But there's that lingering worry.... Might there be a little piece of code in there that, one bright Monday morning, will wipe everything on my network? Will that guy who quit and went to work for the AutoCAD dealer turn us in? What if that fella I lent my discs to lets them get away and suddenly there's half a million copies running around with my name and serial number on them? Do I really want the president of my university to have to issue a public apology after settling a lawsuit because I used 20 bootleg copies in my lab class?
I'm not saying that any of these means are ineffective, nor that there aren't ways in which we could reinforce our defence and increase the credibility of our deterrent. But keep in mind that anything we do in those veins is essentially defensive in nature and negative in effect. When I say negative, I don't mean that Autodesk and its dealers don't benefit from the reduction in piracy engendered thereby, but that no defensive or deterrent measure benefits the user in any way except in the most tenuous and indirect way conceivable--by improving the profitability of the vendor, thereby funding updates and upgrades. That's pretty abstract though, especially when you're faced with the alternative of buying a lock buster or forking over DM12,300 for a legal copy of AutoCAD R11+AME.
What are the incentives to own a legitimate, fully-paid copy of AutoCAD? Well, you get a nice, hardbound manual--generally of much higher quality than what you get with a bootleg copy, and certainly more complete and useful than the various ``Buccaneer Books'' which masquerade as ``simplified user guides'' but are primarily bought by people who knock off the software. You have a dealer you can go to who may be able to help you with various problems...but then you could just as well go the local user group or ask your brother-in-law. You have the right to buy updates without finding a bootleg copy of the update. And then...and then?
What could we do, if we applied the same creativity we've used in squeezing a mainframe CAD system into 640K or building a global sales and support organisation by developing our dealer channel, to increase the incentives for a user to own a legitimate copy of AutoCAD rather than running a pirated copy? What additional value could we provide to legitimate users which is denied, in principle, to those who make illegal copies?
I'm asking you, not trying to sell a list of my own.
Here are some examples of incentives I've stumbled across in the week or so I've been turning this issue over in my mind. These are intended to stimulate your creativity, not constrain the alternatives.
Editor: John Walker