The Second System of man

3.5 The Second System of man

The Second System is a strange beast. Of the three systems that man builds, it garners the most attention, often going on to become a commercial success. It does present some level of comfort and ease of entry for risk-averse individuals. Depending on the size of its market, it may capture the hearts and minds of thousands or even millions of users. Yet, ironically, in many ways the Second System is the worst of the three.

3.5.1 "Experts" build the Second System using ideas proven by the First System

Attracted by the First System's early success, they climb aboard for the ride, hoping to reap rewards by having their names attached to the system. Everyone wants to be associated with a winner.

This group of self-proclaimed experts often contains many critics of the First System. These individuals, feeling angry with themselves for not having designed the First System, lash out at its originators, spewing forth claims that they could have done it better. Sometimes they are right. They could have done a better job on certain aspects of the design. Their specialized knowledge can prove very helpful in redesigning several more primitive algorithms found in the First System. Remember: The First System's designer(s) had little time to do it right. Many of these experts know what is right and have the time and resources to carry it out.

On the other hand, some of these experts are trying to dump a load of sour grapes on someone's hard-won success. Usually this smacks of NIH, the ever-popular not invented here syndrome. Although many of these individuals could produce the First System, they were beaten to the punch. Rather than join the fun, they hope to receive credit for "having improved the obviously amateurish attempts of the original designers" by seeking to replace the First System's mechanisms with their own.

Such attitudes often invoke the ire of the First System's designers. Occasionally they fight back. Bob Scheifler, a pioneer of the popular X Window System, once responded to critics of his early design efforts in handy fashion:

"If you don't like it, you're free to write your own industry-standard window system."

3.5.2 The Second System is designed by a committee

Whereas the First System is typically the brainchild of fewer than seven people, dozens, hundreds, or, as in the case of Linux, thousands contribute to the Second System's design. The First System's success acts like a magnet. It draws in many people who may have only a remote interest in its thought-provoking ideas. Some of these people have a sincere desire to go further with the earlier designer's ideas; others just want to go along for the ride.

The Second System's design committee carries out its business very publicly. It posts meeting announcements in highly visible network repositories, user-group newsletters, and other well-known information channels. It publishes design documents that proudly display the names of all contributors. The committee tries to ensure that all participants get the credit they are due—and occasionally even that they are not due. If a committee builds a sidewalk, its members all want to write their initials in the cement.

In spite of (as opposed to because of) the activities that occur, some real design work takes place. A few committee members sign up to produce key design pieces and actually deliver quality software. Some act as "gatekeepers" for certain critical areas of the software to ensure that modifications are not only well intended, but well designed, too. Others play "devil's advocate" in helping the committee render an honest solution. Still others offer to launch points for interesting discussions that ultimately serve to clarify the new design.

Unfortunately, the design-by-committee approach involves drawbacks. It is nearly impossible for a group (an organizational body comprising at least two individuals) to agree on all salient points. For participants to feel valued, each must believe he has contributed something to the general design, no matter whether he has any expertise in the design area. It doesn't matter whether such individuals are right or wrong. They argue for the sake of their own justification, for the chance to say to themselves, "I can hold my own with these experts and, therefore, that makes me an expert, too." When you add up all the design contributors, though, you wind up with an elephant when you really wanted a gazelle.

3.5.3 The Second System is fat and slow

It possesses the opposite characteristics of the First System (i.e., where the First System was lean and mean, the Second System lumbers along like an enormous giant). If the First System required a minimum of one megabyte of memory, then the Second System refuses to run in anything less than four. People praised the First System for its high throughput on a one megahertz machine; users bemoan the Second System's turtle-like performance on a one gigahertz machine.

"It's because the Second System has more features," says the committee. "You get what you pay for."

The Second System does have more features. Its impressive array of capabilities accounts for part of its success, but the average user takes advantage of only a fraction of these. The rest just get in the way. As a result, the Second System runs slower because it must deal with its copious "advantages."

Often the only way to make it run faster is to buy more hardware. I've long held a belief that computer manufacturers love the Second System for this reason. In these days of portable software, most Second Systems will run on nearly any vendor's platform, provided it is big enough. To take advantage of the new technology, however, customers often must invest in new equipment. The massive size of the Second System virtually guarantees huge numbers of product sales: faster CPUs, larger disks, higher-capacity tape drives, and plenty of memory chips. This spells big profits for hardware vendors.

So the Second System is a mixed blessing. You get lots of features, some of which you might even use. And you get the chance to convince your boss that it's time to buy a new machine.

3.5.4 The world hails the Second System as a great achievement, with much pomp and circumstance

It makes a big splash in the marketplace. It garners widespread commercial acclaim for its flexibility, broad range of options, and expandability. Vendors praise its virtues: It delivers tomorrow's technology today. It goes far beyond existing systems in every respect. It's the last system you'll ever need.

The herd, overwhelmed by this hype, looks to the experts for answers. The experts willingly accommodate them. Any member of the design committee (which by now has grown to several hundred members) instantly gains respect as an expert. Others climb the ladder of credibility by critiquing the work of the design committee.

Interest in the system grows. It becomes the media's darling. Magazines spring up to track its progress. Books appear to explain its mysteries. Conferences are held for the serious-minded to discuss its future. Seminars help the less enlightened to explore its past. The phenomenon builds momentum as more followers join the ranks of the informed and the initiated.

Once the herd is convinced by all the hype that the Second System is outstanding, it remains stubbornly so. For example, the X Window System incorporates many features beyond its basic windowing capabilities. Most of these are seldom used, and their presence drastically impairs general performance. Still, X survives because it is a Second System. Despite its shortcomings, it has enough momentum to overcome any other window system in the Unix marketplace. It reigns supreme. As a Second System, no other system could take its place—except the Third System.