Something I said over the weekend before Microsoft’s Build 2011 conference in Anaheim kicked off: It’s not about the experience; it’s never about the experience. Computing is a process, and what’s important about it is what gets done. This week, a big transition process was started, and a lot got done. And in a few cases, we heard what we’ve been needing to hear for some time: “Message received.”
Let’s go back to the Keys to the Conference from the start of the week, and see whether we picked up on anything close to what we were expecting:
Metro vs. Aero, and ne’er the twain shall meet. The borderline between Metro apps, or “Metro-style apps” (speakers this week were explicitly told to include the word “style”), is big, bulky, and prominent. Calling the sharing of Web apps world and Desktop world “seamless” is like calling Laurel & Hardy indistinguishable. This is a huge design problem.
But we expected it, and perhaps we should give Microsoft a break for not having solved it in the ten or eleven weeks it gave developers to engineer this preview build. There are a number of things to appreciate and truly like about the underlying concept of WinRT. While the .NET Framework has provided useful foundations for Desktop apps, and a necessary degree of abstraction from the underlying system, WinRT gives Windows something it’s needed from the very beginning: a taxonomy. Developers can address the system thoroughly and directly.
So that fact, along with the benefits it provides, is almost a separate issue from the design problems Windows 8 faces. The flaws there now are about sensibility, and the need to address everyday users where they live. It’s the implementation that doesn’t work yet. And here we can comfort ourselves in the knowledge that this isn’t even the beta. The developer’s preview is really “Windows 7.1,” if you will, with a funky front-end attachment. Its problems derive completely from the front-end attachment being an attachment; there’s a missing level of integration that will simply take more than eleven weeks to work out.
Where do all the HTML5 apps go? Um… Do you think we got a clear answer on that one? Should I send the press folks a request for clarification? Anyone out there go deaf after hearing the answer to that question blasted out the loudspeakers?
The answer is, of course, front and center. Windows 8 plans (at least for now) to put all the HTML5 apps in your face. The problem (and it’s a big one), as exemplified by this pair of somewhat different looking apps lists, is that the new system appears to put the old system on a second-class footing. It’s not that you can’t install a Desktop app on the Start Screen – you can. But at least in the Developers’ Preview, the tile that shows up for Desktop apps can’t exactly be called a “charm.” And in the Apps search list on the bottom screenshot above, looking up traditional applications becomes like trying to find the name of an ancient e-mail contact from six years ago on a BlackBerry.
How much of Windows 8 will be in the cloud? A sizable chunk. There’s a level of integration being planned there that truly is seamless; there isn’t some tile or icon or charm that links to a Windows Cloud Portal Access Information System Professional Plus 2012. Most importantly will be an authentication service that enables Web apps to get the identity and trust they need without bartering, and without making the user log on a dozen times. This is the biggest hurdle facing Web apps today, and Microsoft is definitely tackling that problem head-on.
But the problem, like a virus, is evolving; the identity mashup on the Web today is changing literally by the week. Although Microsoft’s demos of Access Control Service work very well today, there is a real chance they might not work next week. Think about it: The whole point of ACS is to enable users to log on through Facebook, without going through Facebook services in the browser. Why would Facebook want that?
Will there ever be a competitive Windows app store? Quite possibly. There’s an important thing about the app store (tentatively entitled “Store,” but surely that’s not the final name) that most observers here have missed, but most professional developers here have not: It has to do with installation, and how it changes the architecture of the program.
Up to now, Windows apps have had to come bundled with installers. They have to be capable of unpackaging themselves and placing themselves where they need to be on any user’s system. This is a tremendous headache for many development shops. Well, “Store” from a developer’s perspective is an automatic installation launch pad. It is the installer, alleviating the burden of having to design the components of programs around how they will be shipped rather than how they work. I’ve been out in the cafeterias and hallways and restaurants in-between sessions at the show, and I’ve been listening to what the developers are reacting to. This very fact – that “Store” is the installer – is among the top three, if not the #1, piece of news for these folks from this week’s show.
Does Windows Server really need the Windows user interface? Windows Server took a bit of a back seat this week, which is undeserved because there was actually some big news there. Server truly is moving away from its dependencies on client-side graphics libraries, and truly becoming more component-ized. I’ll talk more about this development next week, because it should not go unmentioned.
The experience thing. Maybe the Microsoft press folks advised the speakers here at the show to be careful saying the word “experience” around Fulton, because he probably has a clicker in his shirt pocket and he’s counting the number of times he hears it. But it was delightful not to hear “It’s about the experience” blasted at a thousand decibels.
At the same time, though, Microsoft’s UX engineers will need to realize that seamlessness is more than a marketing metaphor. When developers get it in their head that “it’s about the experience,” oftentimes the component that delivers the most graphical pizzazz ends up being bolted onto the underlying architecture, and given the role of “experience layer.” The Start Screen, at least in the Windows 8 Developers’ Preview, is an experience tacked onto an architecture. Observers this week from other publications got the idea that, if you don’t like the Start Screen experience, well, you don’t like Windows 8. And that means (get this!) you don’t like technology itself, and you’re just a fanboy, and you probably have issues with your parents, and so forth.
Listening to the real-world developers here at Build 2011 this week, there indeed has been considerable enthusiasm about WinRT, and about developing real Web apps using HTML5 and JavaScript and CSS3, that get installed without you installing them. That’s what’s getting folks interested, even more so than the touchscreen. To be frank, the Start Screen was a best an intriguing new feature, and at worst a temporary setback. Meaning, you can still dislike the Start Screen and not need psychiatric help, and you can like it without wearing a Klingon costume in public.
Some people here truly do like the Start Screen, and even like the duality between Metro and Desktop. They’re human too. I disagree with their assessments, but certainly those folks saw me coming. But it was not the deciding factor in whether Windows 8 will work or not, because as the developers realized right away, nobody has seen Windows 8 itself yet. This is what Windows 8 will look like; right now, it’s Windows 7 with WinRT attached. It should only be judged at that level. The jury hasn’t even been convened on Windows 8; it’s far too early for anyone to be sentencing fanboys to the Agony Booth.