BY ROBERT VAHID HASHEMIAN

If choosing a CTI platform was merely a matter of taste, I’d go with OS/2. This operating system offers many advantages, not the least of which is its efficiency. Thanks to its low component overhead and its thrifty use of system resources, OS/2 is the Miller Lite of operating systems. As in: "tastes great, less filling." So why is OS/2, a mass market contender if ever there was one, increasingly relegated to microbrew status?

The answer has little to do with technical merit. Rather, it is a matter of marketing power, the significance of which, in my view, has been exaggerated. People are so afraid of being on the wrong side of the Microsoft juggernaut that they can’t scurry toward the Windows NT platform fast enough. In all this haste, people sometimes fail to see that OS/2 would be just fine for them, and that they could use it without penalty. Indeed, if I may be permitted another beer campaign reference, I’d say that OS/2 could benefit from the Beck’s slogan: "Don’t be afraid of the dark." The dark, in the case of OS/2, being potential, but unlikely, compatibility problems.

A BELATED DEBATE?
While I prepared this column, lots of people wondered aloud why I bothered taking on the subject of OS/2. Why raise a dead issue? Was my judgement impaired?

Such questions didn’t phase me. I’m stubborn. Besides, I like to root for the underdog. I decided to go ahead and write my OS/2 column. I started my research, and I soon learned that the number of OS/2 copies shipped in 1997 was around 1.5 million. Of course, this figure is nowhere near the corresponding number for Windows 95 and Windows NT (combined, about 80 million copies). Nonetheless, it indicates a significant number of people use the OS/2 platform.

I must emphasize that just because relatively few people use OS/2 doesn’t mean that the operating system is technologically inferior or that it should be ignored. And just because OS/2 has been a marketing failure doesn’t mean the people who use it should rush to abandon it. Yet the people who support OS/2 are sometimes portrayed as zealots or eccentrics. Personally, I doubt 1.5 million people picked up the latest OS/2 release out of misguided loyalty or an urge to be quaint. These people use OS/2 because it helps them work faster and more effectively. These people just happen to work with applications for which OS/2 is a great platform. And some of these applications just happen to be CTI applications.

OS/2 MEETS CTI
Proponents of CTI knew that its best chance of success was to move beyond its base in Unix. The idea was to break into the mainstream, by supporting the omnipresent PC. Hence, the OS/2 connection. Indeed, the CTI industry was one of the earliest adopters of OS/2.

OS/2 was meant to succeed DOS, which was the preeminent operating system for PCs. In fact, many CTI startups introduced products based on DOS. DOS-based products ranged from voice mail systems to IVRs to call control programs. However, these products were limited because DOS was limited.

As most of us know, DOS wasn’t exactly a robust operating system. It could address limited memory with poor memory management, it was a single-tasking (as opposed to multi-tasking) operating system, it was text-based (as opposed to having a GUI), and it was difficult to set up with hard-ware devices (each program required its own driver). In addition, products based on DOS could run only as stand-alone proprietary programs requiring a dedicated PC.

CTI’s horizons broadened considerably with the introduction of OS/2, which eventually offered a GUI and 32-bit support in subsequent releases. CTI developers responded enthusiastically, creating a host of OS/2-based voice and fax products. Given the number of products, it is safe to assume that developers were bullish on OS/2.

With the arrival of OS/2, many CTI products were free to interoperate with other applications as well as other nodes on the net-work. In addition, they could maximize their usage of hardware, pre-sent attractive GUI inter-faces, and provide for easier installation and maintenance. These improvements were implemented with the assistance of the telephony board vendors, who were quick to write OS/2 device drivers and standardized APIs to access their respective boards. For OS/2 in the CTI space, life was good — until Microsoft decided to crash the party with its Windows operating system.

CLOSING WINDOWS OF OPPORTUNITY
A few years ago, IBM ran an anti-Windows NT ad. A monumentally large NT was paired with more modestly proportioned letters to spell out Not There. The intent was to show that Windows NT hadn’t reached the level of performance, manageability, and integration that OS/2 had enjoyed for years.

While Windows originally showed little promise for CTI applications, it improved slowly but surely over the years. In addition, it specifically addressed CTI issues with initiatives s u c h a s TA P I , t h e Telephony Application Programming Interface. Today, no one suggests NT is Not There. Windows NT is here, there, and everywhere. Windows has become the platform of choice for telephony applications.

OS/2’S ABIDING ADVANTAGES
Although developers can’t ignore the market success of Windows, they should still ask themselves this question: Is Windows (specifically NT) a superior product to OS/2 for CTI? I’d say no. Here are my reasons:

• OS/2 is thriftier with system resources. It can run much faster than Windows NT with a given amount of memory, hard disk capacity, and processor speed. Consequently, OS/2 leaves more resources at the disposal of CTI applications.

• OS/2 carries a lighter burden in terms of component overhead, and it is relatively uncomplicated by compatibility issues. Windows NT, on the other hand, keeps adding functions and features to broaden its appeal. NT 5.0, for example, introduced support for Active Directory and enhanced security models, which certainly adds value. However, such additions also make the product heavier and more complex. Thus, it is harder for applications to support or at least be compatible with the operating system’s numerous components. OS/2 offers a much simpler environment for CTI product development.

• OS/2 has a better track record with respect to support and fast installation of hardware modules. Windows 95 has tried to address these issues with Plug and Play, but in many cases driver and interrupt problems continue to confound users.

• OS/2 has always been more stable than Windows; that is, with OS/2, crashes are less common. OS/2 has had at least a two-year head start on Windows as a 32-bit operating sys-tem. While a longer product avail-ability in itself may not offer an advantage, IBM has used this extra time to fine tune and improve OS/2, yielding a more stable product.

• OS/2 has a good history of pro-viding backward compatibility. With Windows, however, many APIs become obsolete as Microsoft releases new versions. As APIs become obsolete, vendors assume the burden of staying compatible. In contrast, product maintenance cost on OS/2 has been minimal since OS/2 has provided much better support for older APIs as new versions are released.

• OS/2 provides native Java support. If Java telephony turns out to be a big hit (although I predicted it would not in one my previous columns), OS/2 would be able to support it right from the operating system. Windows NT doesn’t provide built-in support for Java. Instead, Java programs need to run within a supporting application.

• OS/2 is less expensive than Windows NT.

OS/2’S MARKET FAILURE
While OS/2 enjoyed a head start in the CTI arena, its standing as a popular CTI platform eroded as OS/2 lost ground in the broader computer market. Exactly why OS/2 failed to win a large share of the market is still debated. Here is my take on OS/2’s market failure:

• OS/2’s reputation suffered because OS/2 was associated with PS/2, which failed thanks to its closed architecture. From the beginning, IBM had attempted to link OS/2 with PS/2, touting the entire package rather than marketing two separate components.

• OS/2 was ahead of its time. When OS/2 became a 32-bit operating sys-tem, and began offering advanced capabilities such as multi-threading and multi-tasking, the PC market lacked the maturity to absorb these benefits. When the PC market finally became more sophisticated, OS/2 was old news and Windows had positioned itself just right to grab market share.

• Microsoft, with its murderous marketing machine, runs roughshod over competitors, and it is particularly aggressive when it comes to promoting Windows. So it is hardly surprising that people side with Windows, even when the choice of platform makes no difference. A case in point: A friend of mine chose an inferior voice mail system running on DOS instead of a superior system running on OS/2. Why? He was afraid OS/2 wouldn’t fit within his strictly Microsoft environment. That was two years ago. Today, the voice mail system is still a standalone box in the closet. How could this standalone system possibly interfere with the rest of the computers?

• Microsoft’s close relationship with Intel. It is no secret t h a t Microsoft and Intel, both giants in their respective markets, have a very tight (anti-competitive?) relation-ship. In fact, Microsoft has participated in some of Intel’s development efforts. Such cooperation reinforces Windows’ position as the operating system of choice on Intel-based PCs.

• IBM, in its dealings with application developers, tends to be arrogant, particularly with the smaller developers. In addition, IBM tends to favor its large customers, and to slight its small customers. (A definite turn-off for the small guys.) Microsoft, to its credit, has always been open and receptive to all of its customers and partners, small and large.

• Last but not least, IBM shipped its own PCs and laptops with Windows pre-installed. Customers had no option to choose OS/2 as the operating system.

Whatever the reasons for OS/2’s marketing failure, it is abundantly clear that Windows is by far the most popular operating system in the broader computer market. This basic fact has implications for the CTI industry. That is, CTI vendors have no choice but to place their bets on Windows NT.

According to Jim Jonez, director of marketing at RightFax, "Our first fax server application was written for OS/2 several years ago, and it is still avail-able for OS/2 platforms. Now most of our development goes into the new enterprise fax server for Windows NT because the market has moved there." Jonez adds that "any technical differences were not factors in the market transition to Windows NT."

This attitude was con-firmed by RightFax’s parent company, Applied Voice Technology (AVT). According to Joe Staples, vice president of world-wide marketing at AVT, "OS/2 allowed us to take advantage of the multitasking capabilities that don’t exist with DOS." Staples also notes that OS/2’s cost of ownership is less than that for Windows NT. All the same, Staples indicates AVT has to recognize market reality: "We still have a healthy demand for the OS/2 version of CallXpress, but the market has clearly gone with NT. Therefore, our strategy at this point is to focus mainly on Windows NT. No new products will be developed on OS/2; however we will continue to enhance our current OS/2 products." Staples also expresses dissatisfaction with IBM’s support for OS/2 developers: "Microsoft is very proactive when it comes to its partners while IBM does not apply the same level of commitment to us — this applies to both marketing and development resources."

I tried to contact IBM to get their views on OS/2’s declining fortunes, but I received no reply. I have noticed, however, that IBM has slowed its OS/2 development, cutting the OS/2 development and support team to a minimum. So, it appears IBM is staging a dignified withdrawal from OS/2, just like everyone else.

CONCLUSION
Whether it was ever necessary to choose sides in the CTI platform war is almost beside the point. People ended up choosing sides anyway. I suppose these people may have been overawed by Microsoft’s marketing prowess. They may have been driven by exaggerated fears over potential incompatibilities. In any case, I wonder about these things. When I see people abandoning a product as fine as OS/2 faster than people fleeing a burning building (or a building they think is burning), I get the uneasy feeling that group psychology has more to do with the choices we leave ourselves than independent assessments of technical merit. Do you have thoughts on this issue? Let me know by sending e-mail to rhashemian@tmcnet.com.