Despite IBM’s stomping all over America this week proselytising for OS/2, Gartner Group Inc, Stamford, Connecticut has bad news for the company: it predicts that IBM will not succeed in its efforts to encourage a majority of users to migrate to OS/2, and by 1995, OS/2’s market share will only be half that of Microsoft Corp’s Windows. This came out at a Personal Computing Conference held by the Group. It reckons that in 1995, Windows will take a 41% market share of new installations of personal computer operating systems, compared with 12% last year; it reckons that OS/2 will have climbed to second place by then, but only just, with 21% of new desktop operating system purchases in 1995, mostly in the large corporate market, up from a miserable 1% last year; hard on its heels will come the Macintosh environment with 20%, up from 10% last year; Gartner sees Unix scoring a low 6%, only twice last year’s 3%, and all other operating systems 2% in 1995 against 5% last year. So what is the problem with OS/2? Gartner does not knock it: the Group reckons that companies that are deploying complex, critical applications on personal computers should use OS/2 – trouble is, this is only a small niche in the market, and most companies will be much better off adopting Windows and following its evolution into a 32-bit system – and, Gartner warns, under any circumstances, people should use high-level development tools to insulate themselves from the shaky future of Presentation Manager. The fate of OS/2’s Presentation Manager user interface is hotly debated, and Gartner reckons that it has been dealt a severe body blow, if not mortal injury, by Window’s announcement. MS-DOS users, it recommends, should see Windows 3.1 as their next step – unless they are on the leading edge of technology implementation, because both OS/2 and Windows will offer similar applications in the long-term, but Windows holds all the aces in the short term because it is as applications-rich as OS/2 is applications-poor. Microsoft has grabbed desktop architecture control from IBM, says Jonathan Yarmis, vice-president, personal computing at Gartner. Though OS/2 is technically superior to Windows, it lacks third-party appeal. Without that appeal, it doesn’t matter how well OS/2 performs. Users buy applications, not operating systems. Can IBM save the situation? Gartner Group makes some recommendations: IBM must stop pretending that Windows doesn’t exist. It needs to deliver its strategic, proprietary extensions – Communications Manager, Data Manager for Windows as well as OS/2, although if it is able to deliver on its promise of Windows compatibility under OS/2, it may well be able to further OS/2 adoption. It should also offer financial incentives to attract software developers en masse to OS/2. It believes IBM will use marketing muscle and financial clout to combat Microsoft. Vendors may be able to pad their wallets or snare a close marketing or co-development relationship with IBM should they decide to look OS/2’s way.
Big Bang
It must rush OS/2 2.0 out the door – but Gartner doers not expect it to ship until the fourth quarter – and must deliver a big bang with OS/2, including many of its delayed initiatives such as OfficeVision, or it will lose even more developer support to the next generation of Windows. Overall, Gartner does not believe that any operating system will again take 90% of the market as MS-DOS did in 1989, and that instead, a single company will have three to five groups of personal computers, each using a different operating system. And it believes that it is a mistake for users to cling to MS-DOS and hope the others will just go away, because on a five year view, MS-DOS will be the most expensive operating system to support. When hardware and software purchase, support, administration, maintenance and end-user trouble-shooting are factored in, it forecasts that the annual cost to 1995 will work out at $6,537 for MS-DOS; $5,167 for Windows; $4,962 for OS/2 and $3,870 for Macintosh. That compares with 1990 figures of $4,716 for MS-DOS, $5
,640 for Windows; $4,740 for OS/2 and $3,096 for Macintosh, so that by 1995 the decision to run MS-DOS will cost at least $1,300 more per user each year. The reason is that as MS-DOS is pushed further beyond its original design goals, it becomes less and less stable, adding new support burdens, increased down-time and user frustration.
