It is important to remember sometimes that this personal computer industry is only 23 years old. The telephone industry was that old in 1899. The automobile industry was that old in 1919. The television industry was that old in 1962. And for all the success and wealth that the PC industry today enjoys, it is still a very young business fraught with all the problems of any young industry, especially a lack of professionalism.

Thank goodness.

Lord, save me from professionalism. The mainframe computer industry is about 50 years old and thinks of itself as very professional. And what did that professionalism get us? The millennium bug, Y2K, the Year 2000 Problem, whatever you call it -- a bunch of mainframe programmers conspired first to make it happen, and then, not to fix it.

In the PC industry, this kind of thing wouldn't happen in the same way. Oh, you can look at a similar date problem in IBM compatible PCs, but is it really a problem of the same magnitude? Of course not, because these are throwaway boxes running throwaway applications. Nobody is writing PC applications for the ages, because nobody thinks any particular PC platform is going to be around for all ages, despite what Microsoft likes to say about Windows everywhere.

A lack of professionalism can often be a real strength. For one thing, professionals aren't supposed to make mistakes, which means that professionals don't often get anything done. I am about to let you in on a little secret about the PC business, starting in some guy's garage and running right up to this very moment in the holiest of holies at Microsoft: Almost nobody writes product specifications. They just start coding. And almost nobody really tests their code: If it compiles successfully, they ship it. This is the reality of PCs and I think it is good, because it makes change happen oh so much faster.

Want to know a PC operating system for which there was a specification and plenty of testing? OS/2. And what did all that work and $2 billion get IBM in the end? It is a very fine operating system, but not a factor in the market. IBM keeps going with OS/2 mainly because Lou Gerstner's compensation plan can't afford a $2 billion write-off.

Here's a quintessential story of unprofessionalism. Back in 1985, when the first Compaq Deskpro 286s were appearing, everyone was wondering how to exploit the chip's protect mode. The 80286 offered for the first time on the Intel platform a chance to address tons of memory and really run multiple applications at the same time without those applications causing problems for one another. Alas, Intel had divined that moving applications into protect mode was a big job. It wasn't a matter of just shifting 8086 code into extended memory, they said. Everything had to be rewritten from scratch.

At this point in the story, I have to admit a little uncertainty about where to attribute unprofessionalism. Maybe it was unprofessional to accept Intel's word on this matter so easily. Or maybe it was unprofessional to NOT accept Intel's word. Whichever way you view it, Microsoft believed Intel and Novell did not.

In less than one week in 1985, a handful of programmers at Novell -- led by Drew Major -- figured how to recompile the 8086 version of their Netware network operating system into 80286 protect mode. Netware 286, as they called it, was effectively written in less than a week because Intel was wrong. The result of this week's work was Novell's total domination of the PC network operating system business for the next decade and the creation of immense wealth. Netware 286 simply beat the daylights out of just about every competitive product thrown at it. Microsoft didn't have anything competitive for 10 plus years. Some people might argue Microsoft still doesn't have anything competitive.

And for years, Novell kept waiting for Microsoft to wake up and smell the compiler. Eventually it happened -- Microsoft finally got a clue -- and the result was Windows 3.0, which took away from IBM control of the PC software standard. What if Microsoft had figured how to do a protect mode version of Windows in 1986 instead of 1989? What would that have meant to IBM and OS/2?

Without unprofessionalism, the Internet as we know it might not exist. Certainly Cisco Systems wouldn't exist. Cisco got its start because Len Bosack and Sandy Lerner wanted to exchange e-mail between their offices at Stanford University. Stanford in those days was a veritable campus of Babel, though, so Lerner and Bosack were on different -- and totally incompatible -- e-mail systems. Anyone else back then might have switched to using the telephone or maybe got a couple MCI mail accounts, but Len Bosack isn't anyone else. His answer was to invent the multiprotocol router, and then install a network of these gizmos linking 5,000 computers on the 16 square mile Stanford campus without permission to do so and without a budget. The modern Internet was invented at Stanford so Sandy could remind Len to feed the cat.

Other universities wanted multiprotocol routers, too, but Stanford was paralyzed and unwilling to give Len and Sandy permission to fill this need. Having failed to grab even a single share of stock in Sun Microsystems, Stanford was determined not to make that mistake again with this router thingee. But the university couldn't decide how to proceed in anything less than a year or two, so they tried to freeze the router business for that period of time. It was preferable in the minds of the decision makers at Stanford in those days to have some other company gain market advantage in routers than to have Cisco gain that advantage and embarrass the university again by effectively denying Stanford a portion of the booty. Remember, there is rarely any real money to fight over at universities, so all wars are over turf.

Supremely unprofessional, Lerner and Bosack started Cisco anyway, without Stanford's permission. They built the routers in their living room and financed the company with credit cards. They sold their product direct to customers via the Internet -- a clear violation of National Science Foundation rules. And by the time they were selling $500,000 worth of routers per month in this manner, 77 venture capital firms had rejected them.

Maybe it's because this industry is so young, but I have to wonder if anything significant really happens except under the influence of unprofessionalism. Look at Apple Computer today. Steve Jobs is running the place like a little kingdom and getting away with it. There are no checks and balances at Apple. Steve can do anything he wants with the company. He can sell it, change strategy, shut the place down, or give it away and nobody will stand in his way. The Apple board just offered Steve a $270 million package if he'll take the CEO job full time, which I don't think he'll do. That would be too professional.

Frankly, I admire him.

Then there is this nasty little legal battle over Java between Sun and Microsoft, and the possible supporting role in that of Hewlett-Packard. This is the week of the JavaOne show, so I have to write something about Java. Hewlett-Packard surprised the world this week by announcing that it will be shipping a version of Java for imbedded applications. The HP version is supposed to be a classic reverse engineering exercise containing no Sun code at all, which is important because HP decided to implement only a subset of the Java features, which would seem to violate Sun's Java license agreement.

Actually, Sun is okay about this new HP Java because it is a true reverse engineering, and there is nothing Sun can really do about it. HP isn't even going to call the language "Java." It's really flattering, in one sense, that HP thinks so much of the Java concept that its loyalty extends past the Java hype. Doing a Java that isn't called Java is telling the rest of the world that Java is important.

But it still creates a problem for Sun, because right now Scott McNealy and friends are madly suing Microsoft for making an incompatible Java version. But where Microsoft's effort is different is in it being based absolutely on Sun's Java SDK, and on Microsoft's insistence on calling it Java. Still, Sun is feeling uncomfortable about not taking firm action against HP (despite the fact there are no grounds for such action) simply because it might further encourage Microsoft. And for Sun to encourage Microsoft would be very unprofessional.