Silverlight
- Mar 25 2009 - silverlight 3 - after the high ...
I failed to convince my manager at work that sending me and a few members of my team to MIX was a worthwhile expense in this economy. So instead I spent a couple days this sprint with http://live.visitmix.com/ on one screen and visual studio in the other. I have to say, Microsoft did an amazing job with MIX in terms of getting me excited and having me “tuned in”. If you are at all interested in web development on the Microsoft stack and haven’t checked out the keynote I’d recommend it. I really enjoyed Buxton’s presentation and Guthrie was amusing.
So now that it’s been a week, and “the Gu” and all those dancing flashy lights are no longer influencing my opinion… I’m STILL excited about Silverlight 3. Sadly the development tools can’t be run in parallel with Silverlight 2 and we’re near the end of our sprint so can’t afford the risk. Which is really too bad because one of the things our current application is leveraging is the wcf duplex polling module. A lovely little COMET like implementation for server push. The version of the duplex polling that made it into the Silverlight 2 toolkit was a little more bare than your typical Microsoft module. And while it works pretty well, it leaves a lot of plumbing code in the hands of the programmer, specifically a lot of asynchronous channel handling code that is a bit of pain to deal with. (though a bit educational too) Anyways, this is one of those areas that Microsoft is improving on in Silverlight 3, and one of those things I’m excited about. Right next to the simpler duplex polling usage for me is the introduction of binary serialization for web services (including duplex!). When comparing to Flex and the myriad of tools and options for using AMF Silverlight was really behind the ball on this one. When we eventually decided to build our tool in Silverlight as opposed to Flex we basically committed ourselves to rolling our own binary serialization. I’m very happy we’re not going to have to follow through on that. Read more from the web services team :
http://blogs.msdn.com/silverlightws/archive/2009/03/20/what-s-new-with-web-services-in-silverlight-3-beta.aspx
Another great addition in the realm of things-that-were-annoying-but-possible-and-already-in-flex is the new navigation uri support within Silverlight 3. Check out Tim Heuer’s typically great post on all the silverlight changes here. (link specifically to the nav)
Lastly to round out my list of really exciting enhancements to SL3 are the network monitoring API, which gives developers events to subscribe to detect when the network is and isn’t present - as well as assembly caching which is huge, allowing Silverlight to cache assemblies like the toolkit so that once a user has been exposed to it they don’t necessarily have to download it again until a new version is required. This in turns makes XAP’s smaller which is always a good thing.
So to summarize, I think the top five features from the slew of enhancements that I’m looking forward to are :- Binary Serialization
- Duplex polling enhancements
- Network detection API
- Assembly Caching
- Navigation and Deep Linking suport
My perspective on Silverlight is very biased to the needs of our application of course. And our application will live and die on the network, with performance being a top concern in everything we do. Controls are nice but we can buy those from vendors like Telerik, animation and media are cool for demos but likely won’t do much for us in the short term. The out of browser story is huge, but again with a SaaS app that relies on the network we don’t envision a whole lot of offline work happening in the early versions of our app.
Honorable mentions for features go to GPU acceleration (performance) and the SaveFileDialog (control) and Expression Blend 3. I don’t use Blend much myself, but the current version is a huge pain for our team. Maybe more on that in a separate post. - Jan 31 2009 - QWERTY Myth and the entrenchment of Flash ...
This is a great article about the myth of how the best technology doesn’t necessarily win. Granted, sometimes the best technology does not win, but there is a persistent and pervasive sense that the populous often chooses the “VHS” over the far superior alternative. The article addresses the VHS vs Beta debate directly as well as the victory over Dvorak by QWERTY. To encourage you to read the original I won’t reveal the clever arguments made.
http://www.reason.com/news/show/29944.html (Read Me!)
I’m posting this because there seems to be a real sense of fait accompli when it comes to the Flash vs Silverlight debate. Critical mass has already been acheived, why would content producers or development shops choose to target any other platform than the Flash runtime when users have clearly already made their choice? How could Beta possibly make a resurgence against an already entrenched VHS? It would take an entire round of evolution before DVD would come along and supplant the status quo. There are a couple reasons why this article has relevance for Silverlight, and why the VHS / Beta argument doesn’t hold water.- Flash vs Silverlight is about a producer investment in technology NOT a consumer investment. Machines are powerful enough, and installations simple enough that the relative cost of owning both technologies is nothing like owning two peices of hardware.
- If there is a competitive advantage for a producer to be gained via a specific technology they will use it. Any differentiators in a competitive field like software has a high potential of making a return. This is a very different decision process than it is for consumers.
- Consumers don’t really care or even know which technology is driving their rich content. They care that it “just works” (like flash based video in comparison to WMP or Quicktime) and that the functionality they desire is there. Without a right-click most users won’t even realize which is which behind the curtain once they have both installed.
- “Owning” everyone (high adoption) is really not that big a deal when your competition can also have 100% adoption at the same time. This is not like choosing a computer or an operating system. Only Microsoft can prevent themselves from achieving their penetration goals.
- Better technology does win. I’m not saying that Silverlight is necessarily the better technology right now, Flash maintains an edge on some specific rendering speeds it appears, and their designer tools are clearly better… but Silverlight has the benefit of coming at this with second mover advantage. They didn’t start from scratch, they built out a proven technology (.NET) into new ground by largely copying and improving on the entrenched technology. (sure looks copied from my perspective but that’s a different post) The .NET runtime, threading, compiled/managed code and the lack of legacy in Silverlight will all combine to produce demonstrations of browser based technology that will be extremely difficult and expensive to reproduce on the Flash runtime.
- Silverlight does not have to “kill” Flash to win, it only needs to join Flash in the 90% adoption numbers to be a great success.
- Jan 28 2009 - Silverlight controls ...
Silverlight 2 may not have the control set that Flex developers are used to seeing out of the box but there are a significant number of control vendors who are stepping up to the plate to fill the void. It seems as though Microsoft’s strategy has been to get the Silverlight 2 runtime out as quickly as possible (and as lean as possible) always knowing that this type of extension to the framework would exist.
I do still hope to see Microsoft push a little further in controls that are downloaded once and only once with the framework itself thereby making our applications leaner - but it’s a pretty serious tradeoff until the runtime has the kind of penetration that Flash enjoys.
Anyway, here’s a nice post from Tim Heuer that does a good round up of where to find those missing controls.
http://timheuer.com/blog/archive/2009/01/28/comprehensive-list-of-silverlight-controls.aspx