What is never really can grasp is the conclusion that no new versions equals a dead platform. In the case with WPF(and Silverlight to some degree) the platform is so mature, that one got all the features that is needed to create rich applications, while having a nice developer experience. It`s just done, no need for future versions and fixes, so one can start build things with it instead.
I have(as many others) made the transitions from building WPF/Silverlight applications to HTML/JS and im astonished over how eager the community is it solve the same problem over and over again. It seems like everyone just goes ahead and reimplementing their own solution instead of improving whats already exists.
Every week someone releases a framework, tool etc. thats exactly the same as the five alternatives already existing and often the only difference is that it`s "implemented by us".
And I will not start with NodeJs, 40 frameworks for spinning up an http-server. Congratulations.
So my two cents, long live WPF, the mature framework for creating Windows applications for many, many years to come!( at least in enterprise ;) )
> all the features that is needed to create rich applications, while having a nice developer experience
I wouldn't say that is true. For example even renaming a Window/UserControl in Visual Studio doesn't work correctly. There are a ton of things that could be improved about WPF, not least (off the top of my head):
* Use Direct2D instead of DirectX9. The renderer used by WPF has been shown to be very inefficient compared to Direct2D.
> Use Direct2D instead of DirectX9. The renderer used by WPF has been shown to be very inefficient compared to Direct2D.
Is that why the capability of a user's video card had an appreciable impact on the performance of a WPF-based app that's little more than a text viewer and editor, at least as of 2009 (see http://community.logos.com/forums/t/6200.aspx)? Is the WPF renderer still suboptimal on today's graphics hardware? I'm thinking in particular of Intel's integrated graphics starting with Sandy Bridge. And does Direct2D have a better software fallback?
One should not confuse the IDE with the technology, but I'd agree that the Visual Studio integration with WPF (although amazing in theory) feels constantly buggy - even after years of work.
Also: regarding multiple selection, this series of blog posts explains the problems: namely that there is no built in way to do MVVM multiple selection with virtual list boxes, a common use-case at least for me.
Yes, I'll accept that some of them may be issues that aren't encountered too often, but they are things that I'm battling with right now so are on my mind ;) But for saving images, why isn't there a Save method to BitmapSource! This to me seems endemic to WPF where the API isn't optimized for common tasks.
I'd say the reason is because bitrot is happening faster than ever before, and if something is not being maintained and you are depending on it for a long-lived application you are going to be burned sooner or later. Granted with Microsoft this is much less of a problem than in the open-source world, but given their scattershot strategy over the last decade it's no wonder that people get jittery at the smell of something withering on the vine (especially after Microsoft stopped being up front about killing things).
You're right, and this opinion is technology-agnostic. Textmate vs Sublime would be a good example, in that Textmate languished while Sublime flourished. Now Sublime has an uncertain present/future, Atom [1] and projects such as Lime [2] have appeared. This still doesn't negate the fact that the original Textmate is a perfectly serviceable editor, and neither Sublime nor Atom truly offer a new, compelling day-to-day feature which would make a developer want to switch.
I'm not innocent of this - I moved from Textmate to Sublime in the hope of exciting new features, and because many people in the community did the same, I have a healthy ecosystem available to me. The point still remains that there is very little wrong with Textmate, in the same way that WPF may still be completely viable.
The new cool is the new cool. People go where the work is and go where their peers are.
I don't disagree that WPF is a good framework for making Windows/Windows Phone applications. Nor do I disagree that the JavaScript/Web community has reinvented many of the same ideas. It's hard to explain to an AngularJS fan that WPF had many of the same ideas long ago. But let's be honest, these ideas (bindings, uni-directional data from from the model layer) aren't really that amazing.
Now, WPF may be ahead of the HTML/JS from a technical point of view right now. But the community is a completely different thing. That's what differentiates things today (not to mention the bag of cash Google is pushing into web standards) and really is the only thing that matters when it comes to the success/failure of a technology.
I think you're right that WPF will last a good deal longer. There's too much software already written for Windows/.NET, but "many, many" years is a stretch.
I have(as many others) made the transitions from building WPF/Silverlight applications to HTML/JS and im astonished over how eager the community is it solve the same problem over and over again. It seems like everyone just goes ahead and reimplementing their own solution instead of improving whats already exists. Every week someone releases a framework, tool etc. thats exactly the same as the five alternatives already existing and often the only difference is that it`s "implemented by us". And I will not start with NodeJs, 40 frameworks for spinning up an http-server. Congratulations.
So my two cents, long live WPF, the mature framework for creating Windows applications for many, many years to come!( at least in enterprise ;) )