The title is all wrong. He is not in charge of Silverlight. Pure speculation.
My reading of directui is that XAML is staying but Silverlight libraries are a dead end. If you are a component vendor, you will stop writing additional components for Silverlight. This also heralds the end of any ecosystem for Silverlight. Yes, there may be SL die hards, but they are going to be like Visual FoxPro developers today. Abandoned.
Developers over the next few years are going to have to migrate their work over to DirectUI for tablet development while retaining Silverlight for Windows 7 OS. Oh Silverlight, we hardly knew you.
There is indeed some speculation on my part...but it is based on diving into what is publicly available. There are threads where people have dissected the leaked builds and shown how to create apps using the same apis that you use today for Silverlight. I personally have not downloaded any of them (though I'm significantly tempted).
True, Silverlight apps may not port directly, especially if you break out of the sandbox and do things that aren't supported within the platform. But because of the fact that Silverlight compiles down to IL...all Microsoft needs to do is bring the Silverlight VM over and boom it just works.
Basically what I've done is looked at the current state of the ecosystem and what has been revealed publicly for the future and connected the dots. If you look at the demo video and observe some of the animations and rich interaction being shown, it's obvious that it's not just HTML on display there.
Yeah, I've been following the naiveuser Jupiter dissection. There is no Silverlight runtime. A new library has been developed on top of DirectUI, similar to Win32 on WinNT. It is not WPF and it is not Silverlight. It will probably start as a subset and end up as a superset. Some parts will not map directly from Silverlight. It looks like the animations have a new model. You will have to port your SL app for this to work.
From its initial introduction Silverlight was Microsoft's response to Flash. It's development was a business strategy rather than the next step in the evolution of a programming paradigm.
If you accept that very few if any mobile apps are being developed/sold now with either Flash or Silverlight then you have to accept that the future of those technologies are in question.
The only thing that we can be certain of is that any decisions on the future of Silverlight will be based on a business strategy. The repercussions to developers from any decisions made will be of little or no significance to those making the decisions.
Microsoft has over 20000 apps on the Windows Phone 7 market place...I don't know the exact mix between Silverlight vs. XNA (the two frameworks that you can target WP7 with) but I'd feel safe betting that at least 75 percent are Silverlight. 15000 apps is hardly what I'd call very few.
First are the apps that are built in to the device. The Zune Media Player coupled with a Zunepass lets you download almost any song in the library directly to the device. I've been able to consistently name a song, look it up on the zune marketplace and listen to it 30 seconds later.
The Office hub integrates with Sharepoint Online allowing you to collaborate on documents from your phone. Make annotations etc. Included in this is OneNote which becomes a portable notebook.
The Games hub has full Xbox live integration with games from the major publishers made specifically for the phone.
Facebook and Twitter both have multiple clients both "Official" and third party.
Netflix, Kindle, Youtube, and a number of other major media outlets have native clients.
4th and Mayor is a foursquare app that was written as an example app by a member of Microsoft's platform evangelism team (http://www.4thandmayor.com/)
So far there hasn't been a case where I wanted an app to do something on my phone and couldn't find multiple examples both free and paid. The ecosystem is very robust. Microsoft just released tools to simplify porting apps from iOS and Android so it's only going to grow from there.
This strategy worked out really well for C# as a response to Java. At first it seemed just like a spoiling move but somewhere a long the way c# turned into the better language.
So you can't blame them for giving the strategy another whirl
Funny story, the original plan was to use Java as the premier .NET language. Then Sun got all litigious and ish. So MSFT made their own. They have the benefit of having full control of C# good thing too. Look at what Oracle is doing to Google over Android.
I think there's a very good chance that the Silverlight "brand" is going away. People think of it as a browser plugin and, whether rationally or not, popular opinion has turned against browser plugins.
Just like with Zune, the brand going away doesn't necessarily mean that the actual technical components are. (Did you know that the development platform behind the Microsoft "SPOT" smart watches initiative from a while back is still being actively maintained - they just released a new version with significant new features a few days ago? It's been repurposed and rebranded as the .NET Micro Framework, a competitor of sorts to Arduino, but the API namespaces still begin with Microsoft.SPOT!)
I'm not sure if they'll call it Silverlight (although I think it would be smart if they did), but the skills used in developing Silverlight applications will translate directly to the new platform.
I think the panicky response is unwarranted based on the demo alone, which I thought looked as though the metro UI was still implemented in XAML.
I think the main issue is that Microsoft has a history of delivering mixed messages and ill-fated solutions that disappear within a couple years. I remember seeing a MIX video (can't find it anymore) from when they announced Silverlight where Microsoft made fun of themselves for doing this to developers so consistently in the past with things like Hailstorm.
Alternately, it could just be a convenient excuse for developers who might have already been looking at platforms with more buzz like Android & iOS.
The problem here is they made this announcement to the Wall Street Journal. Had they done this at a developer conference it would be "just another Microsoft sending mixed signals" thing. But by saying "HTML5 is our development engine" to the largest business newspaper in the world they put MS developers in a tough spot. Because CEOs and CFOs read the Wall Street Journal and suddenly want to question every budget item that contains the word Silverlight.
After watching the demo, I kind of wondered if they expected this initial reaction. Or rather, wanted to see the community's reaction. From there, they could have taken a pure HTML5+JS path if there was nobody upset about dropping .NET. Alternatively, having an upset community shows how much passion and investment there is for XAML.
Good writeup though, it is interesting that XAML has exactly the same mindset as HTML5+JS. Only under MS's control.
In the post you say "Effectively the past 5 years has served to wean developers off what came before and start learning the future of Windows development."
I really wonder if this comes from the same planet I'm on.
I'm a software developer at work and our products don't currently depend on anything newer than .Net framework 2.0. A huge percentage of our customers are running XP SP2 or Server 2003r2. I can't imagine how far into the distant future that anything announced for Windows 8 and above is going to be an acceptable dependency for commercial software development.
As someone who's seen people's careers peak and die as MS promotes then abandons one generation of APIs after another I hardly agree that this kind of speculation is "little more than pure hysterics". It's an industry survival skill.
I am in fact the author. I do understand the conservative approach toward adopting a new framework from Microsoft. However, as I pointed out in the post, Microsoft has been moving this way since Windows 95.
Right, and when they told us Active Desktop was the future of the Windows UI and we should bet the success of our applications on it some of us knew to ignore it back then too.
The lesson here for developers is not to do what MSDN advises you to do. Look at what MS is actually doing for its own application development. In some cases, like Office, that even means having a portable codebase.
Very good point. But Microsoft has made a STRONG commitment to declarative UIs starting with WPF continuing with Silverlight. Windows Phone 7 was a total reboot of their mobile platform centered around Silverlight. There's no pussyfooting when it comes to Windows 8. They're severing the ties and going all in.
Microsoft has made a STRONG commitment to declarative UIs starting with WPF continuing with Silverlight.
What does that even mean? Talk is cheap, show me the code.
If a DLL shipped in Vista, I can expect it will be available in Windows 8. So I might consider using it (for app code that is intended to be Windows-only) once none of my customers are running anything earlier than Vista, or maybe after MS stops providing security fixes. We recently dropped support for Win2k for example.
So new APIs, even for old ideas, aren't usually interesting to me. Wake me up in 5-10 years if you're still committed to it then.
Is this one of those considerations that cloud computing (e.g. Azure) is somehow going to make go away?
My reading of directui is that XAML is staying but Silverlight libraries are a dead end. If you are a component vendor, you will stop writing additional components for Silverlight. This also heralds the end of any ecosystem for Silverlight. Yes, there may be SL die hards, but they are going to be like Visual FoxPro developers today. Abandoned.
Developers over the next few years are going to have to migrate their work over to DirectUI for tablet development while retaining Silverlight for Windows 7 OS. Oh Silverlight, we hardly knew you.