Upgrade lessons learned from Microsoft Vista

Like odd-numbered Star Trek movies, some Microsoft software releases seem best skipped. Is there a pattern here?

So, it's April 2008, and before me sits a classroom filled to bursting with IT people from one of the nation's largest healthcare companies. They're here because their organization's upper management may or may not be pulling the trigger on Windows Vista, and if that trigger gets pulled, they need to be ahead of the curve on Microsoft's latest and greatest.

There isn't much enthusiasm in the room.

During the three-day course, we discuss the features of Vista and the details of its deployment in depth; but more than this, we delve deep into the pros and cons of Vista, exploring the consequences of pulling that trigger. The company that sent this group has more than 50,000 desktops and laptops out there. Lots of time and money at stake.

In the end, they decided to wait for Windows 7, and I'm certain that every single student from that session was relieved. I'm also certain that the same scenario played out in thousands of companies around the world, post-XP and pre-Vista.

Like most of you, I loved the even-numbered Star Trek films and wasn't crazy about the odd-numbered ones. And it's not much of a stretch to say that most IT departments demonstrate the same feelings toward Microsoft releases. We can make a pretty good case for XP as Star Trek II: Wrath of Khan and Windows 7 as, say, Star Trek IV: The Voyage Home, with Vista inspiring about as much love as Star Trek V (the one Shatner directed). Windows XP persists even now with almost supernatural staying power, while just about everyone who experienced Vista is trying to forget it.

Do we see a pattern? We do. If we wander over to SQL Server, we see the hesitations inspired by Vista hovering beyond SQL Server 2005. Vaulting Microsoft into the enterprise database class for real, SQL 2005 was well-received and sold like a John Grisham novel. It was packed with truly innovative advances in the core technology, like CLR integration, the Service Broker, Integration Services to replace the worn-out DTS, and a build-your-own data warehouse kit. It made most of us happy enough that we weren't in any big hurry to jump on board when SQL 2008 came along.

And SQL 2008 was, indeed, a slow mover, more of a replatforming of the 2005 product to stay in step with Windows Server 2008 than a real improvement. Yes, the world was rising to 64-bit systems, but when to go 64-bit was for many a chicken-egg problem. We wanted a SQL Server release that would take us to the next level, as SQL 2005 had done.

And with SQL 2008 R2, we got it. What was the difference? Centralized management of lots of SQL Servers, integration of Reporting Services with SharePoint, Master Data Services, and data tiers in Visual Studio. The next level.

And over in BizTalk, we went from the nothing-much BizTalk 2002 to the awesome Visual Studio-integrated BizTalk 2004 to the prettier but not much more functional BizTalk 2006 to the awesome 2006 R2 with its WCF integration (and built-in EDI!), to the not-even-cosmetically-better 2009 (just another Windows Server 2008 replatforming) to the beyond-awesome 2009 R2 with its performance management, PowerShell, FTPS support, and AppFabric. The next level.

Are any rules of thumb emerging?

If it says R2, buy it. The R2 releases have tended to be not only solid but packed with meaningful features that can be game changers. With the SQL Server and BizTalk R2s, we got a new level of server management and deeper integration with the other server products. And with Windows Server 2008 R2, we got a whole new security platform and a new layer of Active Directory functionality. R2 has come to mean, most of the time, "the next level."

Not just another pretty face. Why Microsoft, of all vendors, would still be so obtuse as not to realize that users don't like their UIs radically overhauled is beyond fathoming. Whenever something is prettier, it may indeed be easier to navigate and may be simplified and objectively better, but it's different, and that pisses people off. Do we adapt in the long run? Of course we do; we have no choice — we're all still using Facebook, right? But a new, improved look isn't really a selling point in itself, and if it's at the top of the new features list, consider it a red flag.

How long has it been in the oven? This rule may or may not hold up over the next few years, but if we look back over the past decade, the Go/No-Go line seems to be about 36 months. Less than that, and we're getting a replatforming release or something pretty; more than that, and we get meaningful improvements that change the way we build things. There are exceptions; Windows 7 followed Vista by a mere two and a half years. But Vista itself was such a turkey that we can understand Microsoft working faster than usual to clean it up. In any case, the serious upward climb seems to yield a new step worth taking about every three years.

And — as always — wait for Service Pack 1.


Scott Robinson is a 20-year IT veteran with extensive experience in business intelligence and systems integration. An enterprise architect with a background in social psychology, he frequently consults and lectures on analytics, business intelligence...

Editor's Picks