Title: .Net Conference 2001

Anyone who attended this year's DevWeek conference in London will have quickly realised that the focus of most of the seminars, workshops and talks was Microsoft's .Net technology. OK, so there were some seminars on Java and COM+ and even 'normal' C++ (thanks Kevlin Henney!) but the excitement of .Net overshadowed everything else for most of the delegates.

In fact the main thrust of the conference was supposed to be COM+ technology, with a heavy appearance by XML and Java, along with a few introductory talks on .Net and associated topics like C#, Managed C++ and Visual Basic 7. Notice anything about this? Don't let anyone tell you that DevWeek is an independent conference - it is mostly about Microsoft technologies and platforms, and from what I gather, always has been.

Anyway, I had some opportunities to talk 'off-line' to one or two of the speakers present (Jeff Prosise and John Robbins in particular, both of whom are well-known authors for Microsoft Press) who, when asked to submit their material late last summer, had chosen to focus on COM+ and programming Windows 2000 using it. Come the day, and COM+ is dead (already if you believe all the marketing) and .Net is the future. In recent months both Jeff Prosise and John Robbins have been working exclusively in .Net and, by their own admission, no longer feel comfortable programming COM and COM+ applications.

Don Box, 'COM Guy', gave the Keynote talk at the start of the conference, and mostly exercised great restraint in not talking about .Net (his talk was about XML anyway) but the gleam in his eyes when he could wind it into the topic was evident for all to see. A workshop at the end which I didn't attend, titled 'In depth COM+', became a .Net forum I gather; Don polled his audience to find out whether they wanted to hear about COM+ or .Net. The result - a .Net workshop instead of a COM+ one. Of course, it is obvious from this alone that Don really wanted to be talking about .Net not COM+.

So where does this leave the humble COM+ programmer? Well, still a COM programmer. I overheard some of the delegates berating Don after one seminar on .Net complaining that he had decimated their hard-earned skills in 90 minutes. His reply? "I bet you all have a bunch of COM code to go back to on Monday". The excitement about .Net blinds people to the fact that it ain't going to happen overnight, especially the porting of COM code to it - which may not happen ever. Even Don Box (even?) sees this clearly enough.

As a C++ programmer - as opposed to a COM or ATL or even Windows programmer - I'm less worried about my skill set. I've been presented with the "C++ is dead, Java is dead - .Net is the future - C# is the future" spiel, my immediate reaction is I've seen it all before. If C++ were an aristocrat it would have been assassinated several times in the past five years or so, with the advent of Java and its like. Now we have C#, the latest in a long line of would-be killers, but I don't see it. .Net is very likely (I hesitate to say definitely) the future of Windows programming, but the truth is that .Net is Microsoft, which is Windows, and C# is .Net only.

So we will soon have Kylix, the Delphi and C++Builder for Linux, and Borland would be crazy not to support .Net, which opens an immediate possibility for .Net on Linux; but does Linux need it? The main reason Microsoft need .Net on Windows is because COM, and to a slightly lesser extent COM+, are in such a mess. Anyone who has had to debug an ATL application will testify to this. .Net, and C# in particular, take most of the pain away from COM programming, but Linux doesn't have COM, and to a large extent, doesn't need it.

My fears of being 'just' a maintenance programmer are allayed also by the fact that .Net is a real hog. It's not just a resource eater, it's the biggest footprint yet in a long line of big footed Microsoft applications. In Microsoft's own admission, for anything which needs pure processing power (the example I heard was a high-end web-server) C# and .Net just don't have the necessary horsepower. So we will have to 'resort' to C++.

It's funny (well I think so) that people like Don Box, Jeff Prosise, Richard Grimes and John Robbins, are now saying that Microsoft's most important technology - COM - is now dead. As recently as a year ago, with the release of Windows2000 and COM+, with .Net on the horizon, COM was the most important thing about Windows programming; so important that almost the entire operating system is based upon it. DevelopMentor, Don Box's outfit, and Wintellect, John Robbins' and Jeff Prosise's company, all make money from training people in Microsoft's sexiest technologies. So now it's .Net. What is next? And more importantly, when? These people are saying .Net is the last thing, the final frontier, that it's the whole of the future. The frightening thing is that people are still buying it!

So, I continue to concentrate on my core skills - C++ development, design principles like OO and Generic Programming. I move forward in my use of STL and object-based development in the full knowledge that I can apply these principles at any level of programming. Maybe now that COM is dead I can concentrate my efforts on real application development without the need to worry about getting the framework to work properly. Give me yesterday's stable technology over the future any day.

As far as C# and .Net go, I've got a copy ready to install, as Microsoft might put it "any day now." I actually think that C# will find a niche, particularly in ASP+ development, the Internet side of Microsoft's technologies, and that Managed C++ will have a role in interfacing normal C++ to .Net code. Whatever Microsoft, DevelopMentor, Wintellect and the others tell you, remember that they make their dosh creating and supporting new technologies. They can't afford for people to get too good at them. But then, that's the cynical view.

Call me a Luddite and I'll call you a Sucker.