As an ASP.NET developer, I've been eyeing Silverlight from a distance for a while now.
But I've been burned by jumping on the Microsoft bandwagon too early in the past when their projects get radically overhauled or binned.
I've been reading some good reviews about Silverlight 3.0 recently.
So have we reached the point where, as a mainstream ASP.NET developer, it's worth investing the time and effort to get to grips with it?
Any thoughts?
Reviews are one thing, adoption by the masses is another. Is Silverlight as ubiquitous as Flash, which is alleged to be installed in 99% of browsers? Are people commonly installing Silverlight when prompted by a site that it's required, or do they pass it by and go elsewhere? That's what will indicate to me that Silverlight has arrived. It's not about reviewers or developers, it's about customers.
UPDATE: My point is that the market is fragmenting again where no one vendor can dictate the way Microsoft could when they had 95% of the browser market for computers and mobile devices were not as important as they are now.
It really depends on your client base too. I've been working with Silverlight 2.0 for the last year and I absolutely love it! Granted, it has it's weaknesses, but it's only version 2!
I'm really excited about the features 3.0 has to offer and I think, compared to Flash, Silverlight's "momentum" is growing much faster!
The features I'm most excited about is the cascading styles, pixel shaders and 3D (really 2.5D) integration.
Plus, the IDE blows Flash away! Visual Studio? I would MUCH rather code with Visual Studio! Intellisense has me spoiled!
Probably not.
Microsoft forever relegated themselves to the Windows-only world on the web when it dropped the Mac version of IE. Now we have cross-platform Safari, Firefox, and Opera, all innovating faster than Microsoft. If Microsoft still held a large share of Mac browsing, and added a Linux IE, they could have pushed Silverlight right into the browser.
Silverlight and Flash are both fighting last year's battle. The hot growth is in the mobile web. The mobile web makes Ajax development more important. You can spend a bunch of money making the desktop version of your site hot in Silverlight, and then still have a bunch of people frowning when they visit your site on Blackberries, Androids, and iPhones.
So you can certainly do Silverlight, but maybe you also have to do Flash as well. And if you have to do both, why not just do Flash, which is on almost all desktops? For coverage, what makes the most sense is Ajax and/or Flash on the desktop, and then some kind of reasonable Mobile strategy.
Silverlight is an answer to last year's problems, not next year's.
I attended a Microsoft "Fire Starter" on Silverlight 3 a couple of weeks ago, and I can tell you that it really is quite amazing. From a technology perspective, I think Silverlight now has more "out of the box" capabilities than Flash. From a development perspective, Expression Blend 3 has radically improved the ease of developing Silverlight products. One of the most important features of Silverlight 3 is the new 'out of the browser' runtime; yes I realize this capability already exists with Adobe AIR, but it's significant in that .NET Developers can now create RIA applications in other operating systems without using Mono.
Another important feature added to Silverlight 3 is a business templates framework, which makes developing typical in house applications a little more practical (though I still don't think this would be my first choice).
In the end, I think many of the previous posters are correct: Silverlight adoption will depend primarily on the skills and comfort-level of the developers that use it. As the technology grows more capable, I don't think low browser penetration will become much of an issue (users will end up installing it to get access to the services they enjoy [like Netflix]).
Silverlight seems to be hitting a critical mass with v3 very similar to .NET 2.0. The base feature set feels very robust and Microsoft is starting to focus on frameworks to complement it (RIA Services, Azure.*, SL Toolkit, etc). 3rd party control support is also excellent at this point in the technology. Looking out over the next two years, a developer's productivity in SL may be more about how many complementary technologies are in his/her arsenal as SL "baseline" knowlege gets to be more standard. One example might be that companies want to know whether you're familiar with ASP.NET MVC, not just if you "know ASP.NET." I'm an open and verbal proponent for SL so take this fwiw, but I think now is a great time to get up to speed. If you wait too much longer you'll end up needing to learn SL and a slew of complementary technologies to get jobs or solve problems at the same speed as the rest of the herd.
If you want to take a rough look at adoption, check out dice. A quick search for flash jobs returns about 1200 postings, silverlight returns 300. That's not bad for a technology that really came of age 9 months ago.
PS - if you're already an ASP.NET developer the transition isn't even that tough! You've already got the tools so you may as well give it a shot. You may like it ;)
For certain products, the time to take Silverlight seriously was 2.0.
I'd never touched Silverlight (or WPF) until August. Since then I've put together a complex user interface for the agricultural telemetry monitoring product I'm developing for a client. We released it a few months ago to rave reviews.
I'm not an ASP.NET UI expert--I have only about a year of traditional web development under my belt. But I have even less time with Silverlight, and I am certain I would not have been able to finish this particular project to the same degree of quality within a similar timeframe without it.
For those who have never used Silverlight, do you recommend starting with SL3?
I've been using SL 3.0 since May, and it's a much better product than SL 2.0. To take just one example that drove me nuts: web services with SL 2.0 were just barely this side of useless. Among other things, you couldn't pass faults from the web service to the client, nor could you catch web service exceptions. Both of those are fixed (with a little server-side work) in SL 3.0. Moreover, the model for duplex calls in SL 2.0 was unbearably nasty (~50 lines of code per web call). This also has been dramatically simplified in SL 3.0.
On top of that, the toolset for SL 3.0 is much better than for SL 2.0. Blend 3 is the first version of Blend that I haven't found myself swearing at each time I fire it up. Among other things, it finally has Intellisense, and seems to handle XAML errors with a bit more grace.
The beta version of SL 3.0 had significant tool issues (which were mostly left over from SL 2.0). I haven't worked with SL 3.0 RTW long enough to know whether they're all resolved, but I'd say that SL 3.0 is definitely stable enough, and feature-rich enough, that it's worth giving it a shot.