By almost any measure, the Microsoft BUILD Conference was big. As attendees return home and digest all that Microsoft has presented, I’m reaching out to key experts and capturing their impressions. Today, I sit down with Andrew Brust, founder of Blue Badge Insights and a Microsoft Regional Director out of New York City. Andrew is also a columnist for Visual Studio Magazine and chair of the Visual Studio Live! conference.
Michael Desmond: In your recent blog post about Windows 8, you write about Microsoft serving two masters and both the unique value and concerns that come with it. What is the danger to Microsoft with its two-tone OS strategy with Windows 8?
Andrew J. Brust: The biggest danger is that people find the two sides of the Windows 8 coin (Metro style apps and Desktop applications) to be awkward and confusing, rather than versatile. Some people demand choice; others are burdened by it. Given that Apple offers tablet users very little choice and that they will have had a two-year head start on setting the tablet market's expectations, Microsoft could find itself a square peg going into Apple's pre-made round hole.
If Microsoft does this right though, customers won't be faced with two opposing paradigms discordantly competing for their attention, but rather a spectrum of choices in device types and modes of operation. For example, tablet form-factored hardware would most likely be used primarily in Metro mode, but would offer the Desktop mode for occasional use, much as people [who] use Remote Desktop clients to connect to Windows machines on their iPads today. At the other end of the hardware universe would be (physical) desktop PCs, whose users would stay in Windows 8's Desktop mode most of the time. In the middle of the extremes, we can imagine laptops with touch screens that get roughly equal use in each mode; or dockable slates (such as the one that BUILD conference attendees received) that get used in Metro mode when untethered and in desktop mode with a keyboard and mouse when docked.
I think we won't know which way this will go until the machines hit the market and consumers react. And the reality is that no matter what Microsoft and the OEMs do, availability of compelling apps, or lack thereof, will have great influence over customer acceptance.
Desmond: I wonder if Windows Runtime and .NET Framework might put Microsoft developers in a pickle, as some devs may have to track and work with two runtime stacks. Obviously Visual Studio and shared tooling alleviates a lot of the concern, but any thoughts on the burden posed by WinRT/.NET?
Brust: I think my answer here would parallel my thoughts on form factors and associated usage scenarios. In other words, I imagine that LOB (line of business) developers will still focus on the .NET stack for years to come. Whether it be Silverlight on the client or ASP.NET, WCF, WF and AppFabric on the server/in the cloud, .NET has the chops for business applications and that's not really Metro's target, at least not yet.
I think we'll have "app" developers using Metro for consumer and kiosk development and business/enterprise devs using .NET for LOB. Will there be crossover? Will there be some experimentation with Metro in enterprise scenarios? Sure there will, but given the ability to use Visual Studio plus VB/C#/.NET with WinRT and the general affinity between WinRT APIs and .NET APIs, I think that crossover will be very reasonable. Certainly more so than trying to move to Objective C and all of Apple's tooling.
Desmond: Silverlight didn't get much mention in the BUILD keynotes, though Silverlight 5 is on its way. What did you take away from BUILD with regard to Silverlight's future?
Brust: I am now more bullish on Silverlight than ever; I think it's now the strategic Microsoft client technology for the next 3-5 years. Think about it: it could take a year for Windows 8 to come out. Then two years more for enterprise deployment of that new OS, and then at least a year or so for new application development to standardize on Metro. Meanwhile, Silverlight is here now, it's mature, and it uses the same markup language, async programming patterns, deployment model and abbreviated .NET profile that Metro style apps will use. If by no other means than a process of elimination, Silverlight is the winning choice now.
Desmond: Overall, how did BUILD compare with some of the epic PDCs of years past? Did it live up to the hype?
Brust: They did a very nice job with that BUILD. My closest comparator is PDC 2003, when Yukon (SQL Server 2005), Whidbey (Visual Studio 2005) and Longhorn (before it was Vista-branded and tarnished) were introduced.
That was big. BUILD was bigger. Everyone was there, everyone was nervous about what would be revealed and even if some people had concerns, everyone was excited by what they saw. It probably also helps that everyone got a free tablet with dock, keyboard and a year of AT&T 3G service on it! PDC 2003's t-shirt didn't really pack the same punch.