CIO Mag article: the End of the Vendor Era

Scott Bertinato of CIO Magazine had an article in the December 11th issue entitled, ".Net, Web Services, and the End of the Vendor Era".

Ok, I know CIO is a business magazine and not very technical, but this particular piece is all about stuff near and dear to my heart: interop and .NET.

There's a lot Bertinato gets right, but I totally disagree with the lead-in, which reads:

...in the process of becoming far less than Microsoft had dreamed, .Net has become much more than CIOs had hoped for and is pointing the way to a new definition of the CIO role, creating a world in which vendors - including Microsoft - matter less and less.

First, I disagree with the premise, which is, that .NET has become less than Microsoft dreamed. .NET adoption is ramping up nicely. A study by Evans Data, reported in the news, said that .NET adoption ramped from 40% to 54% of developers, from fall 2002 until fall 2005. Microsoft's own double-blind internal worldwide developer tracker shows .NET usage ramping up steadily. And a study that analyzes mission-critical app platform usage worldwide, a study funded by Microsoft and performed by IDC shows that .NET usage leads, and continues to rise. So slagging .NET with "it is less than Microsoft dreamed" is inaccurate and misleading. Ok, Ok, I know I am a bit touchy about this. But I feel that it's wrong to state it that way.

Having said that, it is fair to say that the term ".NET" is being applied much more narrowly, to fewer things, within Microsoft. And this is a good thing. In the beginning, there was the .NET Framework. That name debuted around July 2000 I think. Immediately within Microsoft it became apparent that the .NET Framework was going to be big, was going to garner lots of attention. As a result, marketing people within the company attached this .NET name to many products that were unrelated to, and did not rely on, the .NET Framework itself. "The .NET Enterprise Servers". "Windows Server .NET". Those were bad ideas, and we have since corrected them.

Even in cases where the .NET Framework is being used by a product, we have de-emphasized the name. Look at Visual Studio: the 2002 release was officially called "Visual Studio .NET 2002", although everyone just called it "Visual Studio .NET" or VS.NET. In the .1 release, the official name was "Visual Studio .NET 2003", and everyone called it "VS2003". In the v2 release, which we released just in November, the official name is "Visual Studio 2005". The .NET was totally dropped from the name. The .NET Framework is still in there, is still the baseline application framework. But the name has been dropped. All this is good.

That takes care of Bertinato's faulty premise, or at least the unclearly stated premise. Now onto the conclusion, which is "vendors matter less and less." This is also wrong, and belies a fundamental misunderstanding of the situation.

Bertinato observes that XML and web services have enabled more and more interop between heterogeneous systems, and he jumps from there to the conclusion that vendors don't matter. This is incorrect. Just because now it is easier than ever before to interconnect disparate systems, does not mean that the systems in question are commodities, easily replaceable or substitutable. Talk to any company that uses SAP, and ask them if they think SAP doesn't matter. No way. Of course SAP matters. For some of those comapnies who've deployed SAP ERP, they run their business on SAP apps, and just because they can more easily interconnect ERP modules with satellite business logic built in .NET, doesn't mean either end of that interconnection is irrelevant.

You could say the same thing about any situation involving interop between two enterprise systems. Interconnecting them doesn't lessen their value or lessen their importance. In fact, according to Metcalfe, those systems become more valuable as they are more interconnected.

So that does it for Scott Bertinato's article. Next time he should check with me before writing this stuff.