20 Years Of Windows

[Guest Mr. HiTech]

20 Years of Windows

By Staff Reports

November 13, 2005

Tech Analysis: How one operating system came to dominate computing—and what's next.

Has it really been 20 years? Two decades since Microsoft unleashed its graphical operating environment on what was arguably an unsuspecting industry? Say what you will about Windows, its impact has been huge. As eWEEK recognizes the operating system's china anniversary, each of the Labs' analysts examines the effects of Windows technology—both positive and negative—in his or her beat.

Henry Baltazar: Storage

For all the Microsoft haters who say the company never gives away anything for free, the CIFS (Common Internet File System) saga is a compelling example of the company's "largesse." Microsoft's CIFS strategy has demonstrated that the company is more than willing to create technology and bundle it into the operating system at no charge—provided the technology can knock the wind out of a multibillion-dollar competitor.

Microsoft's embrace of the CIFS file-sharing protocol in 1998 was one of the most important developments in the storage industry in recent years because it caused a shift in the LAN landscape. Before CIFS' deployment, Novell NetWare file servers dominated LANs. However, Novell's servers were built on proprietary protocols. With CIFS embedded in every Windows system, Microsoft opened the floodgates for file sharing without user licenses.

Through CIFS, Microsoft was able to bury NetWare, the former LAN network operating system king, and make NT the dominant network operating system. CIFS also pushed the IP protocol into the LAN market, which had been dominated by Novell's IPX in the early '90s. IP was originally popular in Unix and wasn't included in early versions of Windows. (Third parties wrote IP support for early Windows clients.)

With CIFS as a protocol bundled into every Windows-based machine, the doors were wide open for NAS (network-attached storage) vendors and the open-source Samba team to deliver file services for Windows networks free of pesky licenses, picking away at Novell's key source of revenue.

Novell's NDS directory and print services were better than Microsoft's, but file services were Novell's bread and butter—and IP-based CIFS became far more attractive than proprietary Novell File Shares running IPX.

Novell's move to IP and away from IPX came far too late to change IT managers' minds about moving to Windows NT. Microsoft could afford to offer file services for "free" because it could squeeze IT managers on a wide variety of other licenses (for example, NT Server, Exchange e-mail, Office, SQL Server and, later, Active Directory).

With NetWare unable to appeal as a platform for applications, much to the dismay of many a CNE (Certified NetWare Engineer)—myself included—its market share evaporated, and Novell was forced to reinvent itself as a Linux software vendor to escape elimination.

Jason Brooks: Operating systems

Needless to say, the impact of Windows on the operating system space has been dramatic. Windows now powers the vast majority of desktop computers, a sizable chunk of servers, and a large—and growing—segment of smart phones and PDAs.

Regardless of where you and your computing device want to go today, there's a good chance you'll be climbing through Windows to get there, and Microsoft has used its intimate relationships with OEMs to make life difficult for rivals, such as the innovative, excellent and short-lived BeOS.

There's no question that the dominant position that Windows has held—and the aggressive tactics that Microsoft has employed to maintain it—has worked to somewhat constrain our software options.

That said, I think choice and flexibility, at least where hardware is concerned, have been the best things that Windows has wrought. Windows broadened our options by giving us independence from hardware-vendor-specific operating system implementations.

Despite the Big-Brother-battling, "Computing for the Rest of Us" image that Apple strove to build for itself, it was Windows, and DOS before it, that empowered the various cheap, capable PC clones that defined my formative computing years—while other platforms such as Mac OS demanded more costly (but less cost-effective) hardware.

From white-box clones to ultrathin notebook PCs, there's a great deal of cost and feature diversity among hardware systems, and Windows has provided a vital abstraction layer to span these options and help maintain flexibility for consumers.

However, the hardware-commoditizing force that Windows has exerted on the PC world has helped set the table for Linux and the free-software movement, which have begun to broaden platform choices beyond Windows. Linux and free software will force Microsoft to focus more keenly on cost and innovation in the same way that Windows has helped keep hardware vendors on their toes.

Just as the commoditization and fiercer competition that Windows wrought on the hardware level brought us lower costs and more choices in the boxes we run, the period of broadened operating system and platform choice into which we're now entering will mean more innovation—both from Microsoft and from others.

Michael Caton: Collaboration

The idea and nature of collaboration and enterprise applications have changed considerably with the evolution of Windows, but the success of Windows and enterprise applications seems intertwined, from the beginnings of OLE and DDE (Dynamic Data Exchange) through the client/server revolution to now-prevalent Web-based applications.

In any discussion about the rise of Windows, it's next to impossible to separate the success of the applications from the success of the operating system. Server operating systems, including NetWare, Unix and IBM's LAN Manager, introduced most PC users—including me—to the concept of sharing data and applications over a network. However, Microsoft's introduction of Windows NT Server, starting with Version 3.5, brought simplicity to networking, and the common development model made it possible for software companies to write applications so businesses could run one copy as a server and a couple of other copies as clients.

OLE and DDE allowed for the first cross-product collaboration by sharing data in one application with another, but the true granddaddy of office productivity was Lotus Notes. Although Windows (and Microsoft's ability to dominate office productivity software) eventually killed Lotus' first generation of products, 1-2-3 and Ami Pro, Notes allowed enterprises to put e-mail and data-driven applications on every desktop. Furthermore, Notes tied e-mail and the databases running on Lotus Domino servers together in a way that no other vendor did.

E-mail, Web conferencing, and instant messaging and presence have become the primary drivers of collaborative productivity, making the desktop and server platform less important than the tools and data that reside on it.

The Web itself, and the rich Web interfaces available through technologies such as AJAX (Asynchronous JavaScript and XML) and Macromedia's Flex, will further reduce the relevance of the operating system as users get the rich and fast client experience that they used to get from client/server systems in a Web browser—regardless of the platform.

Anne Chen: Hardware

The Windows device driver architecture may have left a lot to be desired when it was introduced, but Microsoft brought to light the issue of handling devices and, along the way, taught many others how to do it better.

Introduced in Windows 3.x, VxD (virtual device driver) was subsequently used in the Windows 95 operating system. The architecture handled software interrupts from the operating system for a computer's hardware devices, such as the keyboard and serial and parallel ports—one of the most buggy areas of an operating system.

True hardware geeks will remember how painful it was to write drivers before VxD. And writing devices using the VxD framework was no guarantee that a system wouldn't still crash. Nevertheless, it cannot be denied that Microsoft vastly improved the experience.

It's true that Microsoft may not have been the first to introduce a framework for writing device drivers, and it certainly was not the best at it. However, Microsoft had the weight to ensure Windows and VxD were used by enough developers to become the standard.

VxD gave way to WDM (Windows Driver Model), the framework Microsoft introduced for device drivers for the Windows 98 and Windows 2000 operating systems. WDM defined a unified driver model by standardizing requirements and reducing the amount of code that needed to be written. It was also designed to be forward-compatible, allowing drivers to be used across all Windows operating systems from Windows 98 and beyond.

Today, of course, WDM is used to provide features including plug and play, which enables users to add and remove devices across Microsoft operating systems without having to download or install drivers. That's a long way from the days when programs such as PrintShop (remember that?) roamed the earth, and the number of disks with printer drivers outnumbered the number of disks for the program by a 4-1 margin.

Thank goodness those days are long gone.

Peter Coffee: App Development, business intelligence, tech apps

The good news and the bad news: It's the standard.

Before Windows, application development suites were like high-end stereo systems. You bought best-of-breed components and plugged them together: an editor from one company, a compiler from another and a debugger from yet a third. Borland's Turbo Pascal and Microsoft's QuickBasic introduced integrated environments, but tool integration wasn't the norm because any developer might be aiming at any of several platforms—like trying to do mass transit in Los Angeles, a city with no single center.

Windows, by comparison, was San Francisco or New York.

Visual Basic 1.0 defined drag-and-drop development, and Windows 3.0's good use of Intel 386 hardware made it clear that OS/2 was fatally handicapped by being originally designed for the 286. The combination of Windows and Visual Basic was a powerful attractor for specialized applications, creating a positive-feedback loop where Windows-based tools built still more Windows-based tools.

Engineering and scientific applications proliferated on Windows for two reasons: Visual Basic made it easy to make the application look good, and BASIC had a long tradition of offering ease of entry to people who wrote programs to do their job—not as their job.

Click here to read Peter Coffee's story "If Windows Had Never Happened."

Any number of difficult tasks became quickly supported by widely available software. For example, my teenage son had a summer job for a professor at USC, turning videotapes of drifting flame balls in a space shuttle experiment on zero-G combustion into numeric data that could be analyzed. We assembled the tool chain to do that in one afternoon of Internet browsing.

The second thing that happened was less benign. In the business intelligence space, data quickly migrated from databases into spreadsheets as Excel became both ubiquitous and surprisingly powerful. BI applications soon had to answer two questions: How did they bring in data from Excel, and how did they integrate with Excel as the tool for end-user analysis?

Without the ubiquity of Windows and Office, enterprise data might have had its primary residence in more disciplined databases and other repositories. As it was, data leakage to the desktop was just too easy and useful to prevent.

Andrew Garcia: Security

From the turbulent days of the early Office-borne macro viruses to today's swarm of self-propagating worms, memory-resident ad engines and mutating spyware strains, Windows continues to be a fleshy and delicious treat for malware.

When the first Windows-specific virus came on the scene in 1992, MS-DOS and the IBM PC were already the most attractive target for virus writers, and Windows has only enhanced that attraction, with the rise of the Internet and high-speed broadband as deadly accomplices. Poor Windows security has given rise to a huge cottage industry of companies vying to shore up its defenses.

I've spent countless hours reviewing the many Windows-based security technologies—anti-virus measures, desktop firewalls, host-based intrusion detection systems and, now, anti-spyware—without hope of ever being able to cover every product that has arisen to clean Windows' (pig)pen.

By chasing the twin gods of simplicity and backward compatibility, Windows continuously leaves itself open to abuse. It's true that Microsoft's dominance in end-user operating systems leaves Windows as the lowest of the hanging fruit for malware authors, but Windows' design and architecture make it a meal that's all the more tasty. Rife with coding errors and exploitable vulnerabilities? Check. Begs and practically requires users to run with the highest levels of local administrative permissions? Check.

The latest Windows desktop incarnations, Windows 2000 Professional and Windows XP, made some small steps toward improving host security, but Microsoft is committed to improving only the most recent operating system iterations. With Windows XP's Service Packs 1 and 2, Microsoft has added a few features (an inbound integrated firewall and restricted executable installation via Internet Explorer) to tighten security. But it steadfastly refuses to offer these features for older operating systems.

With the forthcoming Windows Vista release, I hope Microsoft can solve its security problems—not by bolting on additional protective services but by fixing Windows' underlying flaws.

Jim Rapoza: Web browsers

When Microsoft pulled its famous 180-degree switch on the importance of the Internet, its strategy was essentially based on two cornerstones: IE and IIS (initially known as Internet Information Server and later changed to Internet Information Services).

At the time, if I'd had to place bets on which would become the dominant player in its space, I probably would have gone with IIS. There was a lot of heavyweight competition in the corporate Web server space back then. IIS, which was included free in Windows NT 4.x, looked as if it would wipe out most competition, leaving the high-end Unix side of Web serving to powerhouse servers such as those from Open Market and Netscape, with Apache picking up the smaller Linux-based sites.

As it turned out, most of the competition did go away—of six Windows Web servers I reviewed when NT 4.0 was released, only IIS remains. But IIS didn't become the dominant Web server, instead giving way to the surprisingly mighty open-source Apache, which dominated all sites, high-end and low-end, and has become the de facto engine of the Web.

Because of multiple, serious security holes, IIS became one of the main enablers of dangerous worms such as Nimda, and its appeal to many large Web sites dropped significantly. Now IIS has to settle for being king on Windows.

IE had the exact opposite experience. Pitted head-to-head with the massively popular Netscape browser, it prevailed in short order, due to its own capabilities, mistakes made by Netscape and the fact that Microsoft tied it as closely as it could to Windows.

At first, IE's rise was a good thing because the browser had better standard support than Netscape and heralded the age of free browsers. But by the time it reached 96 percent market share, the rarely upgraded IE was stagnating the Web. It now faces challenges from the children of Netscape, namely Firefox, and has seen its share drop to the mid-80s. This drop is mainly due to mistakes made by Microsoft, as well as the fact that IE isn't being updated on the Windows systems that most people use, and because it has been an enabler of a number of malevolent worms and viruses.

Cameron Sturdevant: Management

Microsoft released SMS (Systems Management Server) 1.0 in 1994, but change and configuration management became a really big deal for Windows, and Microsoft, at the turn of the century.

What began as SMS 1.0 has evolved into SMS 2003, with a second service pack going into beta as we go to press. In the intervening 10 years or so, SMS has grown from a tool for gathering a basic inventory of desktop software and hardware to a more powerful system that includes servers and mobile devices, along with gaining far-reaching configuration capabilities.

The journey hasn't always been easy for Microsoft. When I started covering system and network management tools for eWEEK in 1997, SMS was known for its slow development cycle, clumsy software license metering and awkward implementation. Competitive offerings packaged as families of tools, including Computer Associates' Unicenter, Novell's ZENworks and IBM's Tivoli management products, covered a broad range of operating systems and gave Microsoft a run for its money.

They still do.

SMS 2003 has certainly come a long way, but other products have a longer history of deeply integrating heterogeneous operating systems into a single pane of glass (or fewer panes of glass, at least) in the operations control center.

The introduction of SMS 2003 did away with many of the thorns found in previous versions of SMS, such as unscalable software license metering, and replaced them with nicely crafted, much more practical tools for system management.

SMS 2003 can now reach the wide range of devices that users and administrators must configure and manage, and it therefore does a good job of lowering management costs associated with daily operations. Although management tools from Microsoft will likely stay Windows-centric, the track record now points to a future where Microsoft management tools will increasingly manage diverse environments with greater expertise.

SMS has moved far beyond its early inventory capabilities, which were aimed at desktop systems. Mobile systems such as laptops were hardly at the head of the list of supported devices when SMS first hit the street. The addition of BITS (Background Intelligent Transfer Service) technology enabled SMS to distribute software while taking into account bandwidth constraints and unplanned network disconnections.

Fonte