I think most of the growth from an industry perspective will probably BE outside PC's for the next few years.

Here are some of my predictions:

1. Desktop sales will fall precipitously as notebooks become the predominant computing source. Most people like to take their computers with them. If a standard monitor/speaker/network/keyboard/mouse cradle could be build for all notebooks and PDAs, I could even see a situation where you carry around all your computer work on a PDA, then "plug in" in one of the those standard cradles. Your certificate in the PDA would tie you in to the "net" (hopefully not .NET) and connect you with your work/home/whatever server. Your standard config is loaded on the PDA and you can work anywhere one of these cradles exists.

2. Software sales will also die on the vine because of sales of services. Most companies would rather rent software than own it, unless it's free. If maintenance on BEA WebLogic or DB2 is 20% to 25% of the purchase price EVERY YEAR (and it is or more), then I think most companies would rather buy a service, and someone else pay for that. It also fits with the SLA, service agreement, sue for non-compliance world we currently live in.

3. Finally, there is more and more thought that computers should be "invisible", just like electric power. They should just "fit in" to whatever someone naturally does. This is especially true among younger people. There is nothing special about computers, the "specialness" of them is gone. With that thought, I can see more and more that people just use them for what they need. They care less and less about Windows and more and more about doing what they need to do. I think Windows isn't Microsoft's premier brand. It's Office! Word and Excel documents pretty much rule the marketplace and probably will for the next 100 years.

4. With that, I think places where Small-talk architecture fits, in graphic interaces, in heirarchical structures, in fast development/deployment cycles, Small-Talk makes sense, and is probably light years ahead of even Java (which IMHO is MUCH BETTER than C/C++, with respect the amount of time it takes to build something). Most companies are very afraid of missing their market window of opportunity. Thus, they like the rapid deployment aspects of Java, Python, PHP, and even SmallTalk. I think most of these languages can do 80-90% of the computing, with C/C++ doing a much smaller percentage than it does today. The problem is that a manager/director who did it in C/C++ doesn't want to give up his "power" in the structure to do it in another language he doesn't understand.

5. The younger developers of the new generation coming online will have to take the industry from C/C++ to Python, PHP, Squeak, SmallTalk from C/C++ Java, just like my generation took us from mainframe COBOL to C, then C++, then Windows, then Java. I think as the productivity numbers become proven, first developers, then companies will change languages.

But, that's part of what I love about this industry. I started in mainframe COBOL and Assembler, with some microcomputer ( Lotus 1-2-3, Paradox, BasicA, Turbo Pascal) skills. Then I learned C and C++, then Windows, then Unix, then relational databases, then Tuxedo middleware, then Java.

Now, I'm at a place where I'm back doing C/C++ (and hating it), but making a living until I can do something more interesting. I just hope (SERIOUSLY) that our industry hasn't calcified to the point where things can't change anymore, and where companies won't let you learn. If we get there, then I'll have to find something more interesting to do.

Glen Austin