VMS: The antithesis of UNIX?

9.3 VMS: The antithesis of UNIX?

If MS-DOS was the king of the PC operating systems before Windows came along, Digital Equipment Corporation's VMS^[3] was the king of minicomputer operating systems. Before Unix had caught the world's attention, no other single operating system in the minicomputer space had earned such a wide, loyal following.

VMS owed much of its success to Digital's VAX line of computers, where the same software could run unmodified on a range of systems from the desktop machine to the large, room-filling lab system. This compatibility at the binary level made it relatively easy to construct huge, fully integrated environments with a uniform style of interaction throughout.

Digital's approach to building systems was at one time the envy of the industry. What Digital had done with VMS and the VAX line had no equal in the minicomputer world. Other companies, such as IBM and Hewlett Packard, had tried to unify their entire product lines around a single architecture, but had failed. The success of Intel's x86 architecture running Windows is an affirmation of Digital's "one system one architecture" concept as well.

VMS was a closed-source, proprietary operating system, meaning a single company developed the source code, maintained a strong grip on it, and derived all profits from it. Although written by Digital's VMS Engineering Group in Nashua, New Hampshire, it really represented the aggregate thoughts on software that Digital's engineers had carried in their minds for years. The similarities to Digital's other operating systems were many. For instance, users of RSX-11, an earlier Digital operating system, felt very much at home with the VMS command interpreter. Like earlier Digital operating systems, VMS relied on compatibility across an entire hardware product line to strengthen and extend its user base.

The VMS developers had strong ideas about what constituted the VMS philosophy. Like the developers of the Atari Home Computer's operating system, they assumed not only that the user should be amply shielded from any vagaries lurking within the system, but that the user should be well informed of their existence. If the Atari approach was to build a tunnel through solid rock, then the VMS approach was to put lights in the tunnel.

Second, in the VMS environment, if big was good, then bigger was better. This led to the development of large, complex environments, such as Digital's All-In-1 office automation product, which integrated many closely related functions under a single user interface. VMS had been quite successful in this area and had one of the strongest office suites available before the PC-based ones came along.

A third element of the VMS philosophy was an artifact of the effect that the VMS mind set had on Digital's customers. Many people who purchased VMS did so because they had a defined task to accomplish and were unaware or unsure of the kind of information technology required to provide a solution. While the typical Linux user today says "I need this box" or "Give me that application," the typical VMS customer would ask, "How can I do this with your system?" The difference may appear subtle, but it profoundly affected how VMS engineers wrote software.

Applications developed by VMS engineers tended to be feature rich. But they were often rich under the covers, where the user couldn't see all of the wondrous complexity. The choice of features shown to the user at any one time was not very large, despite the fact that a large number of interesting options did exist. Unix and Linux applications designers usually make no such presumptions about what is interesting. The choices on a typical Unix system often include everything: the good, the bad, and that which may be safely ignored.

To a certain extent, an underlying belief of the VMS philosophy was that users are afraid of computers. Given the ubiquity of modern PC's, this might sound a little strange. The fact is, many people still feel threatened by computers. If they could, they would prefer to ignore the benefits of this exciting technology. Many wish that computers would simply go away.

Jon "maddog" Hall, a longtime proponent of Unix systems and more recently of Linux systems , has always asked a tough question: What about Mom and Pop? Mom and Pop, he would say, have a microwave oven, but use only 59 percent of its capabilities. They have a CD player, but don't know what quad oversampling is. They own a VCR and have yet to program it. They don't know what all the buttons on the remote do. In short, they refuse to learn about any technology that is not easy to use.

Mom and Pop-most Unix advocates hate to admit this and yet readily do so-would have a terrible time with the "have it your way" approach of Unix and Linux. Proponents of the VMS philosophy dealt with this notion a long time ago. The VMS developers were in the business of making the world's computers more accessible to those people who needed bounded solutions, not do-it-yourself toolkits.

Besides the obvious user interface differences, there were other ways in which VMS differed from Unix. Because they are at opposite ends of the spectrum in so many respects, one could claim that VMS was really the antithesis of Unix. For example, VMS usually provided a single path to a solution; Unix often provides a dozen ways or more. VMS tended toward large, monolithic programs with lots of options to accommodate the vast general user population; Unix leans toward small programs, each of which performs a single function with a limited set of options. VMS was originally written in assembly language and BLISS-32, languages highly tuned for the underlying hardware architecture; Unix is written in the higher-level C language for porting to many CPU architectures.

The single most significant difference between VMS and Unix, however, was the fact that VMS was a closed-source, proprietary system, while Unix, in its various implementations, has staunchly remained an open-source system, despite the attempts of many vendors to capture it, close it, and call it their own.

VMS and its closed-system approach ultimately fell to the onslaught of Unix and open systems in the early 1990s. No surprise there. Digital's proprietary operating systems have always caved in to the relentless drive of open-source systems. In the mid-1970s, one watched as Digital's RSTS and RSX systems were gradually replaced by Berkeley Unix. In the early 1980s, Digital's TOPS-20 system saw its turf invaded by AT&T's System V Unix.

In this chapter we have compared Unix and its design philosophy with those of other systems in computing history. We have seen one operating system philosophy that expects user interfaces to be like tunnels through solid rock. We have seen another system become popular because of the herd, the mass of people who follow whatever's catching on simply because it appears that everyone else is heading in that direction. And finally we examined an operating system that refined the "tunnels through solid rock" philosophy and brought it to its ultimate conclusion in a complex, monolithic, industrial-strength system.

Unix and its tools approach has outlasted all of those operating systems. Perhaps it is because the Unix philosophy is an approach to developing operating systems (and software in general) that constantly looks to the future. It assumes that the world is ever changing. It acknowledges that even if we know everything about the present, our knowledge is still incomplete.

Each of the systems discussed here has succeeded in delivering a measure of computing technology that works in some cases. But none of them addresses the needs of the broadest range of operating system users. The Atari Home Computer operating system restricted user choices to the point that the users went elsewhere. MS-DOS, while successful for a while because of the herd, eventually saw itself replaced by an operating system with a GUI because it wasn't powerful enough to stand on its own weaker command set. And VMS, while it captured the server market for a while, couldn't compete very well when open-source systems showed their first glimmer of value.

By now, however, you must be wondering where Microsoft Windows is in this discussion. Why did we not compare Linux and Unix to Microsoft Windows? Isn't Windows the only other operating system that matters? Why talk about old operating systems that hardly anyone uses anymore?

I could point out that one reason for the success of Windows is that it limits choices for new users, much like the Atari Home Computer's operating system did. Wall Street knows that Microsoft has replaced IBM as the darling of the herd. And a day doesn't go by where you don't hear about how Windows is the ideal platform for "complex, industrial-strength computing."

Can you see the continuity here? Good. You're catching on. For those who don't get it, wait a few years before you install Linux. You'll enjoy it much more after we've put up the lights in the tunnels.

^[3]Digital later renamed VMS to "OpenVMS" in an attempt to fend off the aggressive marketing efforts of "open" (as in "Unix") systems vendors. Calling a closed system like VMS "open" was perhaps one of the ultimate ironies of its long and successful history.