The purpose in copyrighting a piece of software is to control its copying and/or commercial use. Richard Stallman has extensively questioned the desirability of this purpose. I'm going to do something simpler: Question its feasibility. Let's consider three cases.
Laws that cannot be enforced are vacuous. Laws whose violations cannot be detected cannot be enforced. Violations occurring in an office cannot be detected unless an employee blows the whistle or the violator is careless. Whistle-blowing is not a common phenomenon. Consider conscientious office worker Clara Tillinghast. Her company pollutes the environment, at least one male executive she knows of makes improper advances to female employees, her boss regularly lies to customers, and the person in the next office has an unauthorized copy of Lotus 1-2-3. What will Clara do?
Even if Clara doesn't get on the phone to Lotus, the person in the next office might get careless, leaving a hand-labeled diskette lying about when Jim Manzi is visiting the office, say. But with the prevalence of hard disks, there need be no illegal diskette to observe, and the illegal copy can be erased in seconds. No, the sad fact is that user carelessness is not something software vendors can count on.
One strategy is to issue site license contracts that are cheap enough to make even the small risk of detection unacceptable. But this leaves pricing out of vendor control. Since there is no clear formula for determining the ratio of site licensing price to single copy price, tight competition in any product category could conceivably drive this ratio arbitrarily close to one. That is, the vendor will effectively sell one copy per office.
But it has been mooted that what really sells multiple copies of a program into an office is the need for manuals. If so, the trend toward online documentation could be disastrous. And the trend is compelling: The user really ought to be able to get the answer at the point and in the medium where the question arises; software can be produced more economically without hardcopy manuals; and it can be produced to tighter deadlines, hence more competitively, if two distinct production lines don't have to finish simultaneously. If hardcopy manuals are all that makes multiple sales to an office possible, the elimination of hardcopy manuals will also lead to one copy per office.
Here the risk of detection due to carelessness and the reward for whistle blowing both drop to zero, which is why copy protection hasn't gone away in game software. But games are generally cheap and ephemeral, and copy protection has been so discredited in the broad business software market that it is unlikely to come back. What happens if the trend toward working at home continues?
Consider the pressures. The company supplies ambitious home-office worker Bob Slocum with mediocre spreadsheet program X. A barrage of ads tells Bob that spreadsheet program W will give him the edge over his rival for promotion, Jack Green. Bob can't afford to buy W himself and the company won't spring for it, but he's sure that Jack is just the sort of guy to take unfair advantage by pirating W. As a home-office worker, Bob naturally belongs to a user group, where he knows he can pick up W for nothing. What will Bob do?
It's easy to imagine a vicious cycle developing, with home-office workers copying software because they can't afford to buy it and vendors raising prices because of the diminished sales. Pretty soon, they're selling one copy per user group.
Let's postulate what many would call the worst scenario, and what I have been painting as the most plausible scenario. No office buys more than one copy of a piece of software, and all home users get their software from friends in user groups. This still lets the software vendor sell a copy to each office and each user group, right?
It's actually possible to imagine an industry configured this way, although pricing for software would have to be high because every purchase would be a group purchase. In this scenario, the direct customers would be defined by access considerations. Since copyright violation would still be illegal, sending unauthorized copies through the mail might be seen as too risky, and sending them by phone would surely be too expensive. Copying would all be local: User groups would be sneakernet-bound and offices would be site-bound. User groups would have to find a way to get money to pay for the software they'd buy, but selling support suggests itself as a possibility. It might work.
Maybe. We'll have to see what the bit-per-second pricing for phone service is when wide-bandwidth phone service goes national in 1992.
Case I: Copying at the Office
Case II: Copying at Home
Case III: Copying in Cyberspace