Cutting through the Nonsense about Windows Vista, Windows 7, etc.

By:  Peter Schay

Date:  14 January 2009

Hot Issue:  What is the optimal timing for rolling-out new versions of desktop operating systems?

SmartSummary

•     Lack of application software compatibility continues to be the number-one reason that businesses have not accepted Windows Vista.

•     Vista represents a major advance in Windows security, but the advance comes with trade-offs.

•     Businesses commonly tie OS upgrades with periodic hardware refresh programs.

•     The time has come to start planning your migration to Vista or Windows 7.

SmartAdvice

With Microsoft’s release last week of the Windows 7 beta, there has been a crescendo of articles and blogs from pundits unmatched since the release of Windows Vista two years ago. Much of it is rubbish, betraying a lack of understanding of how operating systems are engineered, the economics of the software industry, and IT departments’ real decision processes. This SmartTip summarizes our advice to clients for optimizing their desktop operating system selection/migration in 2009 and beyond.

Why has Vista “failed”?

Let’s accept the premise that the adoption rate for Vista, particularly in business, has been far lower than Microsoft had anticipated. The reasons generally cited include the following:

Lack of application software compatibility – We believe this was, and continues to be, the number-one reason that businesses have not accepted Vista as their default desktop platform. It’s important, however, to clarify different aspects of application software incompatibility:

1.     Because of repeated, lengthy slips in the Vista development schedule, many software vendors (including some major ones) delayed testing their products with Vista pre-releases, which in turn delayed the availability of Vista-compatibility patches/releases. This issue was largely resolved, however, by the third-quarter of 2007, which is much sooner than most businesses would have rolled-out a new OS version anyway.

2.     Like it or not, many businesses are running back-level versions of application software products, due to the effort and cost required to migrate to current versions. As such, the availability of Vista support in the current versions of those products is no help. Likewise, all of the usual difficulties with modifying legacy in-house-developed applications apply to modifying those applications to work properly with Vista. Finally, in the case of out-of-production peripherals or embedded systems, the vendors have no motivation to provide Vista-compatible drivers or applications.

3.     A distressingly large number of applications, both vendor products and in-house developed, violate long-standing Microsoft programming recommendations (e.g., don’t store variable information in the Program Files directory). This often requires that end-users operate with administrator privileges (a really bad idea from a security perspective). Moreover, too many existing applications took advantage of known holes in Windows’ security model. Which leads us to…

You can’t have it both ways when it comes to security. – Microsoft has taken a lot of abuse over the years about the security of their products, for the most part deservedly. We believe, however, that since launching their “Trustworthy Computing” initiative in 2002, they really have been trying to do their best to engineer more secure products – given the realities of software engineering and economics. While not perfect (no non-trivial software is), Vista does represent a major advance in Windows security.

Ironically, for all the effort they put into securing Vista, Microsoft still can’t win. First, as noted above, the security improvements in Vista are a major factor in the lack of application software compatibility. Second, one of those improvements, User Account Control (UAC), turned out to be a major end-user annoyance.

The principles behind UAC are simple, and hard to disagree with (at least if one is concerned about security, as businesses should be). First, don’t allow malicious software (whether from e-mail, web sites, or wherever) to surreptitiously install itself or otherwise modify the system. Second, don’t allow end-users to do things that could potentially undermine enterprise security policies.

Unfortunately, the Vista implementation of UAC, which brings up a dialog box requiring administrator credentials whenever a program tries to do anything that violates the principles above, can be highly annoying, particularly if one installs a lot of software or doesn’t have administrator credentials. Of course, if your enterprise locks-down its desktop configuration, then your users shouldn’t see UAC anyway.

Personally, as a Vista user since RC1, I like UAC, having trained myself to stop and think each time I see it, “am I doing something that should be asking this, or is it malware?,” but for many users it was seen as getting in the way of getting their job done.

Hardware requirements – Let’s face it – Vista needs at least a Pentium 4 processor and 1.5 Gbytes of RAM to run acceptably. And yes, the fancy Aero graphics require a discrete graphics processor with 128 Mbytes of video memory – not something that the usual installed-base business PC would have had in January 2007. Who cares? The reality is that businesses don’t typically upgrade to new OS versions on installed hardware anyway – the labor required to do an in-place OS version upgrade is too expensive. Businesses migrate to new operating systems as part of their hardware refresh cycle, and any new PC has more than enough horsepower to handle Vista.

With the current squeeze on IT spending, of course, an obvious and common tactic is to increase the time between hardware refresh cycles. Many businesses are now adopting four-to-five year cycles to stretch their IT budget and resources. This is another reason for the slow uptake of Vista in the enterprise.

User interface (UI) differences – Yes, some things moved around in Vista from where they are in Windows XP, and it does take some time to learn the differences. Yes, Microsoft has a long history of rearranging the UI with each major release, but somehow the persistent perception is that the Vista UI does not offer enough new benefits to justify a switch. Curiously, this obstacle is becoming a moot point, since it is nearly impossible to get a consumer-grade PC with anything other than Vista, so chances are rising that users have experienced a Vista system at home. Moreover, Microsoft Office 2007 has been accepted in business more widely than Vista, despite its significantly different UI, because the general feeling has been that the benefits of the new features outweigh the migration training costs.

What about Windows 7?

Remember those analogy questions when you took the SAT in high school? Windows 7 is to Windows Vista as Windows XP is to Windows 2000. Don’t be fooled by the version numbering – Windows 7 should have been called version 6.2. Why isn’t it? Because Vista has developed such a bad reputation that Microsoft, for purely marketing reasons, needs to leave the Vista brand behind.

For IT managers, the most important thing to understand about Windows 7 is that, with regard to all the reasons cited above to delay migrating to Vista, they apply more-or-less equally to Windows 7 as well. If your application software is incompatible with Vista, then it will be incompatible with Windows 7. While there will incremental performance improvements in Windows 7, the hardware requirements are the same as for Vista. UAC has been improved to have more granular control, and thus be less annoying, but it’s still there (as it should be). And yes, the UI changes yet again, but is much more like Vista than like XP.

Is it time to look seriously at alternatives such as Linux or Macintosh?

A recurring theme in the blogosphere, ever since it learned what the hardware requirements for Vista would be, is that the demise of XP, combined with the lack-of-acceptance of Vista, creates a huge opportunity for desktop Linux and/or Macintosh. We consider this naïve at best, and disingenuous at worst.

Why? Look back at the number-one reason that businesses have rejected Vista – lack of application compatibility. Neither Linux nor Macintosh brings anything to this party. Yes, Linux has Wine, but an application sensitive enough to Windows version differences to not run on Vista is not likely to run on Wine either. Yes, Macintosh can run Windows, as a virtual machine using VMware or dual-boot using Boot Camp, but doing so adds complexity for both the end user and IT, and still leaves the issue of the eventual end-of-support for XP on the table.

TAC SmartGrid

SmartActions

So, what should businesses do in 2009? Sooner or later, Microsoft is going to pull the plug on XP, as they eventually did for Windows NT and 98. Unless you have other motivations to undertake a migration to Linux or Macintosh (either of which would be a much bigger effort than migrating to Vista/7), now is the time to start planning your Vista/7 migration. If you already have a Vista migration project underway, continue it, as Vista and Windows 7 will coexist nicely. If you haven’t started a Vista migration project, Windows 7 availability will be soon enough that a “skip Vista” migration plan makes sense – with the understanding that it isn’t going to save you from the work and cost of upgrading your applications software to be Vista/7 compatible.

SmartLinks

Getting Started with Application Compatibility in a Windows Deployment – Microsoft white paper

About the Author

Peter Schay, TAC Executive Vice President, has more than 25 years of experience as a senior IT executive in IT vendor and research industries.  Previously, he was Vice President & Chief Technology Officer of SiteShell Corporation, an Internet start-up which built and operated a network of 170 radio station websites.  Prior to working for SiteShell, he served 12 years at Gartner, Inc. as Director of Research, Vice President Midrange Computing Strategies service, Vice President & Service Director Client/Server service, and Group Vice President Research Infrastructure & Support.  Before that he was Vice President Marketing at Vitesse Electronics Corporation, Product Marketing Manager for the VAX 8600 and VAXclusters at Digital Equipment Corporation, and a Member of Technical Staff at AT&T Bell Laboratories.