November 26, 2009

A History Lesson: Microsoft Windows - "Exit, Stage Right"

Greetings. For this holiday weekend, a quick trip into the Time Tunnel (watch out for the sparks and smoke!) as we explore the ascendancy, and decline, of some key technologies. Students of technology history are quick to recognize that when viewed retrospectively, most major technologies exhibit an arc of innovation, widespread deployment, then decline and decay. The precise shape of the curve in any given case will vary, but particularly in the communications and computer fields, there are some pretty obvious patterns -- with declines sometimes hastened by poor decisions, a lack of insight -- or both.

The telegraph once reigned supreme. But when Western Union was offered the early chance to buy Bell's telephone patents for virtually a song, WU passed, considering the phone to be a faddish device without serious utility. Today Western Union's only remaining obvious existence is as a money transfer service frequently abused by Internet "419" scammers.

AT&T and its kin later suffered their own cloudy vision in not recognizing the enormous potential of the Internet until relatively late in the game -- leading directly to today's various efforts by telcos and cable companies to regain control of the Internet ecosystem through attacks on Net Neutrality principles.

Large mainframe computers saw their time in the sun, and still exist today, but by the 70s were being rapidly replaced in many applications by minicomputers of various sorts. Especially notable was the Digital Equipment Corporation (DEC) PDP-11 series, from which Ken Thompson, Dennis Richie, and the rest of the motley crew at Bell Labs Murray Hill Unit 1127 brought forth the proprietary Unix, leading ultimately to today's ubiquitous open-source Linux.

Yet DEC also suffered a lack of vision, exemplified when their founder Ken Olsen suggested in 1977 that nobody needed a computer in their home. Now DEC is but a corporate memory, twice removed.

And so we come to the arc of Microsoft, based on the rise of the personal computer and the Windows OS. This curve can be traced back to the original IBM PC in 1980, which itself had roots in the hobbyist 8080 CP/M OS world. The IBM PC spawned a vast landscape of clones, largely due to the publication (in the famous "black notebook") of the full source code for the hardware's BIOS, and lack of litigious attacks on clone-makers (in contrast to the later and ongoing behavior of Apple, one might note).

Bill Gates and Microsoft thus provided the software "stuff" that kept the PC architecture running, though not without the use of some business techniques that draw the ire of regulators and consumers alike.

Microsoft's own insight initially proved deficient in a manner similar to that of the phone companies, in their failure to recognize the importance of the Internet early on, then trying to rapidly play catch-up later, sometimes through anticompetitive actions.

Microsoft's various Windows operating systems have reigned supreme over the vast majority of PC for many years now, but that arc seems to be heading inexorably downward -- as various forces converge to relegate the very large, proprietary operating system model to the pages of technology history.

This really isn't about XP, Vista, or Windows 7 per se. I'm in the camp that never saw Vista as being so abominable as many detractors claimed, especially when pre-installed on new hardware. And Windows 7 does clean up some of Vista's rough edges (though some of those edges were pretty straightforward for knowledgeable users to smooth out even within Vista).

The really big problem is basically a matter of complexity. Windows' ubiquity has resulted in an immensely complex mass of software to support a vast range of hardware.

Since Windows is proprietary, piracy is understandably a major concern to Microsoft -- resulting in ever escalating and what many observers would characterize as increasingly intrusive and objectionable anti-piracy authentication mechanisms.

Windows security patches and other updates occur in complex environments, with complicated dependencies that may result in legitimate users being unable, for example, to install new security fixes after their repeated attempts to install other Windows update packages have failed for any of a multitude of possible reasons.

Complexity is part of modern computing, and hiding much of that complexity from most users, without unreasonably compromising the user experience itself, is both a science and an art. But complexity combined with proprietary operating systems can create a "Catch-22" situation for those users who might be able to figure out why they're having problems, but can't even begin to do so since relevant source code isn't available. What does Windows update KB04822284582 really do? Only Microsoft actually knows for sure.

Microsoft is still selling lots of Windows copies of course. And there certainly are alternatives to Windows. Apple's current OS is partly BSD-based (another open-source Unix-like OS family tree), and their tight control over hardware has reduced the complexity of their driver support issues and associated problems. Linux as an open-source, standalone OS, e.g. in modern incarnations such as Ubuntu, is both powerful and flexible, but has had difficulty penetrating the mainstream consumer desktop marketplace due in part to perceived support and applications availability issues.

All of which leads us to the current arc with the positive derivative. "Cloud computing" -- especially linked with relatively compact, typically Linux-based open-sourced operating system environments -- may hold the promise of solving many of the "complexity" problems that have both led to the bloat of Microsoft Windows and held back the mass acceptance of desktop Linux.

Amusingly, cloud computing in many ways brings us full circle back to the
Isaac Asimov "Multivac" model of computing that was de rigueur back in early Unix minicomputer days, with most processing power and storage on central facilities, accessed by users via relatively simple terminal devices. Our "terminals" today are typically far more powerful than the central computers back then, but the comparison remains valid.

Cloud computing can only be as good as the service providing the cloud, and this is an area where Google is rapidly staking out territory while the rest of the industry plays catch-up. And though Microsoft Windows seems most at risk right now, there's unlikely to be a long hiatus before similar dynamics significantly impact the Apple ecosystem as well.

Proponents of cloud computing point to resources availability, data reliability, security benefits, and a host of other positive features. Detractors express concerns about Internet connectivity, security risks, and privacy. It would be foolish in the extreme for anyone to dismiss out of hand any of these factors.

Viewed from both historical and technological standpoints, it's difficult to escape the sense that the arc of the Big, Proprietary OS is on a decided decline. And it appears that largely open source, cloud computing models are but beginning what is likely to be a rapid rise that I hope -- and believe -- will ultimately stimulate the solving of challenges that are currently associated with this computing model in some quarters.

In particular, I don't see any insoluble aspects to these issues, given the will to deal with them -- and there are some fascinating research facets involved.

Getting cloud computing right is especially important given the development and support issues associated with the alternative "PC-centric" model with its Big OS dilemmas and proprietary operational structures -- increasingly unpalatable to users of all types.

None of this is going to happen overnight. Technological change, especially in computing, is more of a process than an event. But as we review the arcs of technology reaching back a century and more, it appears that the Windows OS ecosystem -- that has played such an important role in the advance of personal computing for so many years -- should be preparing to take its final bow and then exit, stage right, with head held high.

Time, and technology, marches on.

All the best for the Thanksgiving holiday.

--Lauren--

Posted by Lauren at November 26, 2009 07:05 PM | Permalink
Twitter: @laurenweinstein
Google+: Lauren Weinstein