Nothing Is Permanent

In an article I saw today, Steve Ballmer claims that, not only are we still in the Windows era, we will always be in the Windows era. A bold claim, to be sure, but an overly optimistic one. I do agree that Windows, in some form, is likely going to be around for a very long time. But I do believe that, at some point, its dominance is going to fade. It is inevitable. The needs of people and companies change, and Microsoft has been very slow to adapt.

But why will it be around for a while? Simply because there will always be some people or companies that won’t want to give it. Look at the mainframe computer as an example. For much of the late 1960’s and into the early 1980’s, the mainframe ruled the computing world. It was far and away the most prevalent form of computing. IBM seemed convinced that the mainframe would always hold the crown for core computing functions in corporations. But, the rise of the UNIX workstation and server, followed by increased use of Windows and Linux servers, coupled with a move to desktop and distributed computing, put the mainframe into what appeared to be a death spiral. Companies everywhere were shutting down mainframes, and decommissioning the few they had bought, and returning the leased units without a replacement. When I was at Bell-Northern Research in the late 1980’s and early 1990’s, our primary goal was removing the mainframe as much as possible, and it was a big day when the first lease unit was shipped back.

Is the mainframe gone from the computing landscape? Not hardly. It was down, but not out, and experienced a resurgence toward the beginning of the 21st century. The zSeries mainframe still represents a healthy piece of IBM’s revenue, and technology from the mainframe made its way into other IBM products. Early virtualization technologies, cluster technologies and some of the scalability of DB/2 owes its origins to the zSeries. For some database computing tasks, the mainframe is still one of the best choices you can make.

Even the UNIX server has seen some evolution. For a brief period, the landscape was dominated by some form of Motorola 68k, running a custom version of UNIX based either on BSD or System V (either release 2 or 3, depending on the vendor). The 68k was eventually supplanted by a variety of bespoke RISC chips such as SPARC, PA-RISC and RIOS. Those in turn have diminished, replaced largely with Intel and AMD processors. Even UNIX itself has faded some, replaced more and more by some form of Linux. Yes, HP-UX, Solaris and AIX are still around, but only HP-UX seems to have any meaningful traction or growth. Their biggest competitor: Windows servers. They integrate nicely with Windows notebooks and desktops, and there are a lot of people trained on how to setup and maintain these systems. The UNIX era certainly isn’t over, but it has been morphing into the Linux/Windows era for workgroup and mid-range servers.

Yes, some technologies are gone, never to be seen again. Operating systems like AmigaDOS, GEM and CP/M have come and gone. MS/DOS only exists as a virtual environment under Windows, and at some point, it will likely disappear. Even the Windows we use today isn’t the same Windows that saw life in the mid-1980’s. It has far more in common with Windows NT that it does with the original Windows 1.0 or 3.0. But Windows has become incredibly widespread, not just on desktops and laptops, but in the server room. I fully expect to see Windows servers, likely for the sole purpose of running MS Exchange, in use and being actively supported for many, many more years.

We are years away from seeing Windows dwindle into the background. I don’t expect the Windows Phone to make much noise, even with the emergence of the new Nokia Windows phones. Android continues its dominance, and I expect a lot of that is simply because there are more choices available and they are far cheaper at the entry level. It would appear that smartphones are now starting to replace feature phones in the mainstream part of the mobile market, and Android has cheap phones that are critical for this part of the market. It will be interesting to see if the free iPhone 3GS makes any sort of dent in the numbers when the Q4 2011 data emerges. But I really don’t see Windows on the phone making much of a dent. It has been downhill pretty much since it’s introduction, with no meaningful uplift in sight. I also don’t see Windows 8 doing much in the tablet space, mainly because you still have to contend with the fact that the OS really isn’t built for mobile devices, and the vast catalog of 3rd party software is largely too big and resource hungry to work well on limited devices. There are too many assumptions about available memory, storage and screen space, and if you’re going to effectively redesign and rewrite them anyways, why not move to a more popular platform? Economics are going to matter has much as technology in that discussion.

So, while technically I expect there will be some form of Windows around for a long, long time, the dominance and relevance of Windows is likely to diminish over time. At some point, the Windows era will come to an end, much as the mainframe era did before it. That doesn’t mean it will disappear entirely. But it will probably cease to be a meaningful force in the industry. It won’t be tomorrow, and it won’t be next year, but it will be within my lifetime.