Is WPF dead: the present and future of WPF

http://pragmateek.com/is-wpf-dead-the-present-and-future-of-wpf/

by anilmujagic • 12 years ago

97 84 12 years ago

pragmateek.com

0x0• 12 years ago

Page not found?

Cache link: http://webcache.googleusercontent.com/search?q=cache%3Apragm...

dotsamuelswan• 12 years ago

Same topic. Different Author. http://greglevenhagen.com/is-wpf-dead-no/

ripter• 12 years ago

"I was just in a meeting with Microsoft and they’re definitely working on the next version of WPF. Obviously, I can’t disclose specifics, but the fact is they are gathering feedback and actively working on WPF."

He doesn't say it was a meeting with actual developers. Just because a sales/management guy says they are working on it doesn't make it true.

Microsoft has a history of this. MFC, Winforms, WPF, and now WinRT

mu_killnine5• 12 years ago

As a developer for my company's internal LoB tools, I really enjoy WPF more than Winforms. I feel very empowered and, as a Microsoft shop, it seems to be really great for meeting a specific aesthetic and function for my users.

Could I do it in a web app? Probably. But it would take a lot more work for me and I'd be battling compatibility with IE7 (yes, we're still on IE7) as the corporate web minimum spec.

WinRT felt half-baked when I used it about a year and a half ago. Furthermore, I doubt Win8 is on the docket for quite some time.

WPF hasn't been the 'cool thing' for quite some time (ever?), but both I and my users seem to really enjoy what it's doing for us. And that's good enough.

TheMakeA• 12 years ago

If you have the choice to deploy a WPF app, couldn't you ship something using chromiumembedded to get the benefits of a web app without worrying about browser compatibility? In the past, I've even dropped a CEF WebView into WPF and had the UI in HTML/CSS/JS.

mu_killnine5• 12 years ago

Perhaps. I am not familiar with the project.

But if you've already got a captive audience able to run anything .NET 4.0 and above (incl. async/await via Bcl.Async libraries) with the rich experience of WPF, why bother shoehorning a web application as a solution?

Our users are in a manufacturing environment, on a well-known environment. It's practically like developing for a gaming console. And in that light, WPF seems to be a good fit.

jenscow3• 12 years ago

This is what I'm doing.

We're able to produce a standalone desktop application and a web app using the same code and have it looking/working exactly the same, on different platforms.

robbyking• 12 years ago

I actually wrote an application using WPF, and what's worse is it was relatively recently. The project was an audio/visual control application, and our CEO (not CTO) made the decision to use WPF because "the Mac fad was in its final days." (Never mind the obvious, platform-neutral rebuttals.)

Truth be told, though, the development process was pretty straight-forward and I don't have anything negative to say about that aspect of the project (though our application was pretty straight forward). The only downside was we deployed the application at a university, where about 90% of students and staff were running OS X and unwilling to install Silverlight on their machines, which made sense; the iPhone 4 was already out, and responsive, cross-platform access was already considered MVP by most product managers at the time.

mu_killnine5• 12 years ago

I am writing a WPF app now, I don't think that's necessarily a bad thing. The issue I see with your example is that WPF as a platform went against just about everything your user needed.

   * Alienate users with a platform that's not compatible with their OS by default (OSX, Linux)

   * Alienate users with a platform that's not compatible with the multiple platforms you may be targeting (Tablet, Phone, Laptop, Desktop)

   * Require an extra step (e.g. Silverlight plugin) for users to run the app at all

That's why the LoB example is such a powerful testament to how WPF may stick around. You've got 'trapped' users (for better or worse) who will likely be forced on the Windows platform.

zbowling• 12 years ago

I thought this pretty apparent a few years ago? Has everyone still been on the bandwagon and still developing in WPF heavily? I sure Microsoft is not going to kill it but it's not the active area of focus with RT. It's unfortunate. Friends that develop components have had to go from VB6 (native ActiveX/OLE C++ components) to WinForms to WPF to Silverlight (then back to WPF) and now RT. No one UI system to rule them all. Always uncertainty that it will change. Hell it changed a lot from CE 6.5 to WinMo 7 to WinMo 8.

On Mac we had classic (carbon) and then AppKit (which is arguably the same UI toolkit since the 90s on NextStep but still doing ok.. although feeling a little dated compared to UIKit).

Linux is a cluster of UI systems but at least X is stable.

wolfgke• 12 years ago

WinAPI is stable.

malkia• 12 years ago

All I can say is - Sticking to Qt5 for now.

maxxxxx4• 12 years ago

If I had to start a new desktop app now I would also go for Qt. I am tired of MS rolling out new frameworks and abandoning them a few years later.

Scuds• 12 years ago

In Visual Studio 14 CTP4 you can still File -> New VB / C# Project -> Windows Forms in same as you could ten years ago, but with the next generation CLR in the new IDE.

It's still _supported_ and there are a ton of vendors still shipping components. Abandoned/EOL/Discontinued is something like FoxPro or Apple's Rosetta emulator. It's not even deprecated.

steven777400• 12 years ago

WPF is dead in the sense that Microsoft doesn't want to be promoting a non-cloud, non-mobile technology. WPF is alive in the sense that it is the best solution for meeting certain customer needs today.

For example, we have a very heavy desktop application that needs to run on Windows 7 PCs with IE 8. When we started developing this application a year ago, the cost to do it in HTML5/CSS/JS given the maturity of the tools at that time, compared to WPF, made WPF the obvious choice. In a few years when local web databases and synchronization frameworks have improved (we just saw a really neat one a day or two ago here on HN) and all the machines have been replaced with Windows 10 and IE 11, an offline web app would be the right choice. But we can't (or, don't want to) delay development until that time. So, for now, WPF is the right answer.

But anyone who thinks WPF is still a first-class citizen or that Microsoft is going to come back in marketing support of it is probably kidding themselves.

_random_4• 12 years ago

Enjoy it while it lasts, Bootstrap/HTML and JS are so lame after XAML and C# :(.

jenscow3• 12 years ago

In my opinion, Dart makes JS less lame (via dart2js). It's the closest you'll get to C# in the browser.

tonyedgecombe• 12 years ago

I just started a new WPF application, I don't think it is going away any time soon.

_pmf_• 12 years ago

WPF is dead in the sense that Microsoft doesn't want to be promoting a non-cloud, non-mobile technology.

Which is strange, since this is the only place where they still dominate and offer tangible advantages over practically all other fat client solutions. Java desktop UIs are dead (except from Eclipse, which is a walking zombie despite their e4 efforts), and there are not many other options if you want a solution that supports bidirectional data binding.

The recent trend of using web applications inside a Webkit/Chromium container has a strange attraction, as convoluted as this solution is.

ZanyProgrammer• 12 years ago

A lot of the criticism I've heard stems around the steep XAML learning curve. OTOH I've only done WinForms and ASP.NET stuff, so I haven't had first hand experience.

mu_killnine5• 12 years ago

That's a pretty fair criticism. But once you've gotten it, it's very powerful (incredible data binding!). And I think we need to be sure to separate WPF from "XAML".

XAML, for certain, is here to stay for quite a while. I recall talking to Carl Franklin about how Windows 8's UI layer is written in XAML. Given that, I think it's fair to say that the time you invest in XAML won't be wasted if you're a Windows developer. WPF may fade into the background, but WinRT will continue to mature and broaden as a platform (esp. as Windows 10 opens the door to cross-device compatibility)

maxxxxx4• 12 years ago

Good luck finding a company that uses Win 8 so you can actually deploy a WinRT app.

And wait until they lose interest in WinRT too like they have done with every new technology quickly.

mschaef• 12 years ago

XAML by itself is actually fairly easy to work with because it doesn't actually do that much: At it's core, XAML boils down to an XML-based mechanism for constructing objects. Squint your eyes just right, and it starts to look like the markup for an IoC container:

http://www.codeproject.com/Articles/17188/Inversion-of-Contr...

Where things start to get complicated are the objects that you create with XAML... (which is where the WPF framework comes into play.)

Five or six years ago, I prototyped a system using XAML/WPF to build an HMI system for an industrial process control applications. The system served up XAML from a server that got loaded by a client to construct UI objects. It then inspected the Binding objects that the XAML created and started subscribing to the appropriate server-side data points. It came together within a couple days and really ran quite well.

_random_4• 12 years ago

Half of Angular is pretty much a WPF/MVVM rip-off, so it's a good experience.

codereflection2• 12 years ago

It will be interesting to see if the editor in Visual Studio is re-written again (it has been WPF based since VS 2012). If a hint of that comes out, then I think we can safely declare WPF as dead.

bdavisx• 12 years ago

I think it was WPF based starting with VS 2010 actually.

codereflection2• 12 years ago

You are correct, it was 2010.

bithush• 12 years ago

I miss the editor from pre-VS2010. The WPF based one has so many problems with fonts other than Consolas and Courier New. Such a pain in the ass.

CmonDev2• 12 years ago

The Visual Studio 2020 might be completely in-browser unfortunately :(.

amaks• 12 years ago

WPF is another victim in ever changing technological landscape in Microsoft. Who uses or even remembers these days ATL/WTL, MFC, OLE DB, ADO, DAO, ODBC etc. etc.?

fleitz• 12 years ago

Now you know why MSFT has the largest developer community in the world.

There are endless lists of technologies that do the exact same thing.

_random_4• 12 years ago

Will anyone be mad enough to adopt the next living corpse of tomorrow - WinJS?

mpthrapp2• 12 years ago

I still use ADO and ODBC (oh, and the BDE, remember that one?) on a regular basis. I wish I didn't, but I do.

dvere• 12 years ago

Raises hand, still have a VM with Delphi 7 that is opened once a week or so....

pjmlp2• 12 years ago

Who uses or even remembers these days QuickDraw, QuickDraw 3D, WebObjects, Carbon, Qt for Java, Java Bridge, Nettalk, Hypercard, NewtonScript, etc

andwaal2• 12 years ago

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 ;) )

grokys6• 12 years ago

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.

* Multiple selection is a hack.

* Can't bind data grid column properties.

* BasedOn="{StaticResource {x:Type Button}}" -- eurgh

* How fucking difficult it is to save an image?

* How fucking difficult it is to get an image's pixels (how do you work out Stride?)

peetle2• 12 years ago

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.

mwcampbell2• 12 years ago

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?

andwaal2• 12 years ago

The form renaming im absolutely with you on, horribel. Once you tried it you learn to stick with your original name.

Direct2D vs DX9 is probably a valid point, I have never faced any issue with this.

The rest of them I don`t think I follow you on. Two trivial image tasks, one ugly syntax and two issue I can`t seem to have encountered.

It should be noted that I always have used telerik`s awesome UI components together with WPF. Which both features and documentation are top notch.

jsmeaton• 12 years ago

* terrible WebBrowser control

dasil003• 12 years ago

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).

peetle2• 12 years ago

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.

colinramsay• 12 years ago

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.

[1] https://atom.io/ [2] https://github.com/limetext/lime

grokys6• 12 years ago

Personal plug here: I've been tinkering with writing a GUI toolkit along the lines of WPF here: https://github.com/grokys/Perspex/. It's obviously very early days and this is as much for my curiosity as to how things could be done differently as setting out to replace WPF, but it may interest some of you...

logfromblammo• 12 years ago

Given enough constraints on the development environment, WPF can turn out to be the best choice.

Those constraints are usually as follows:

- Our developers cannot use anything not written and distributed by Microsoft, and can use only our existing licenses.

- The end users will not have full administrative control over their workstations. They might not be able to install a Java VM, Flash, Silverlight, or even a browser other than the version of IE that came with their OS.

This sort of environment is depressingly common in the government contracting sector. When your choices are WPF, WinForms, ASP.NET, Excel macros, or Access database, the first option seems pretty attractive. And this is especially relevant when the documentation from Microsoft promotes a model-view-viewmodel design pattern for WPF applications. That's a great way to get people who haven't read a single paragraph about best practices since graduation to step boldly into a new decade, even if it is not the one the rest of us currently live in.

_random_4• 12 years ago

I am a Microsoft fan but my thinking recently is like this: why WPF => why Windows => why Microsoft? If crappy HTML5 and JS are their agenda, then what's the point sticking to their ecosystem?

emodendroket3• 12 years ago

So you can leverage your/your team's existing knowledge of C#/the .NET Framework, which seems like a good enough reason. Lots of Web sites are still built on ASP.NET MVC.

Also, one unfortunate aspect of the .NET community is a lot of people know next to nothing about anything outside the MS ecosystem and have a blindly dismissive attitude toward it (at least until MS latches onto it like they have with Angular and others).

jenscow3• 12 years ago

With asp.net vNext able to run on Linux (officially), the question still remains: Why Windows => Why Microsoft? (apart from leeching their tech)

GlennS• 12 years ago

I was told 3 or 4 years ago by a Microsoft MPV that WPF was effectively in maintenance mode with only a single guy working on it.

A shame, because it's probably the best UI framework I've used. I liked it much better than HTML.

ygra2• 12 years ago

Given that HTML is probably the worst UI framework that ever existed, that doesn't say much. However, WPF is so far ahead of every other UI framework I used (WinForms, Swing, Qt, Android, MFC) that its descent into maintenance mode is a bit sad. Also that no one bothers to look at what WPF did right when starting new UI toolkits and frameworks and we get more stuff that awkward and clunky to develop for that looks like all other UI APIs from the 90s.

moron4hire2• 12 years ago

I've been doing Windows development for over a decade now. WPF was too different from WebForms (or less often, WinForms) to get most of the other developers on every team I've worked on to agree to switch to it. I wish we had switched way back when, because it is a much better design than (Web|Win)Forms.

That said, now that I'm on my own, I'm not switching to it. The legacy windows-only apps I support I'm sticking to WinForms, because I don't want to spend effort on learning yet another Microsoft ecosystem API. I'm actively getting out of MS-only work now.

Unfortunately, there are no UI toolkits that I would call "good". They all have some weird idiosyncrasy to them. It's like they all start out with a great idea, and then beef it in the face halfway through completing the full set of controls that would be useful to most people. It necessitates 3rd party control libraries of vastly variable quality, and makes it extremely difficult to tell clients "yes, we can do that, and it will be done by XYZ date".

So, lacking a better alternative, I'm going all-in on HTML5. It's the closest to "ideal" of what I want for a UI toolkit. I've about decided that I really, really don't want my UI to be anywhere near my application layer. A node-webkit-style deal might be bothersome to some people, but for my ideals, it's exactly how I want to arrange my applications.

I'd just like it very much if Node weren't my only readily available option. I like Node for small things, and I use it in a few projects, but there are some other projects where I'd really prefer to be writing the backend in C#, or C, or anything with a wide variety of libraries that aren't mostly just a weekend project for a CS grad somewhere like half of everything on NPM often appears to be.

rwallace3• 12 years ago

Well, one nice thing about using JavaScript and HTML 5 for the front end is it does, as you say, give you a lot of decoupling, which means you don't actually have to use Node for the back end; you're free to use C++ or Python or whatever you find best.

rmsaksida2• 12 years ago

Why is Node your only readily available option? There's nothing stopping you from writing the backend in C# (or anything else) and the client code in HTML and JavaScript.

moron4hire2• 12 years ago

I know, it's just that wiring up the two is significantly easier for me right now than for the other languages. Well, it's mostly just a matter of not having had the time to build out a workflow in C# yet.

cnahr• 12 years ago

Is that for server-client or fat client applications? For the latter, Java today offers the WebView class (part of JavaFX) that's simply an integrated Webkit browser. You can build your UI for HTML5 and easily hookup with the main Java program. I wrote a brief overview last year, along with other options for fat client HTML5 frontends: http://news.kynosarges.org/2013/12/29/webview-the-other-java...

joseacta• 12 years ago

We make a software for licensing WPF applications, http://www.licensespot.com, and it's been growing pretty steady in the last few months so I guess there are still a lot of people developing and selling apps in WPF.

It just works. Some niche markets still prefer to use desktop applications.

ternaryoperator• 12 years ago

Not even so niche. Businesses are still primarily PC/desktop oriented.

doczoidberg• 12 years ago

WPF is not dead. It is mature as Winforms is.

Microsofts now is more open to the open source community. That does not mean that 'old' mature frameworks are dead. It's all about choice.

rwallace3• 12 years ago

If I was writing line of business software these days, I'd either go the web route and use JavaScript and HTML 5, or if it was something that preferred to live on the desktop, I'd use Qt.

Now, if you've only heard a little bit about Qt, you might be tempted to think you only need it if you need cross-platform, but in my experience it's actually better than any of the Microsoft frameworks even purely on Windows. (A simple test: write a Hello World program in Winforms, WPF and Qt, then try doing some simple drag operations with the mouse, moving the window around, resizing it. Last I tried, Qt was the only one that updated the window correctly, without flicker. This remained true even after I tried the recommended solutions in Winforms and WPF to reduce flicker.)

The fact that Qt gives you the option of going cross-platform if you ever need it in the future is just a nice bonus.

mwcampbell2• 12 years ago

I think the best thing that Microsoft could do for developers right now is to backport a reasonable subset of WinRT to Windows 7 and provide it as a redistributable, even if we have to pay for it. I'd suggest backporting the XAML-based UII toolkit and at least some of the other APIs, but not the app container or store-based distribution.

triangleman2• 12 years ago

I'm planning on developing a collaborative helpdesk application for my small team (perhaps 30 users), having only previous experience in standard HTML/CSS/JS/PHP/SQL development.

I was thinking of going the web application route but I get the sense that a Windows executable is the way to go here, with its faster/more responsive interface. Major features needed: Embed and modify web pages in tabs (recreate a web browser), social "ticker" or feed, lots of cross-connections between different types of content (KB articles and tickets, for instance).

Should I go with a WPF application? I would have to teach myself XAML and C#. It would be limited in scope so that's good but it could still take forever.

Make a browser plugin in HTML/CSS/JS?

Make a full page Angular app?

In all these cases I'm going to have to learn something new, learning about Windows development may help me get ahead at my company more than other technologies, but if this project takes too long then it won't be worth it.

Any thoughts?

rwallace3• 12 years ago

Given your stated requirements, I'd use Qt. There's a learning curve, but no worse than with WPF or anything else.

But those are hefty requirements for a small in-house project! Spending a lot of extra programming time on production values is what you do when you're trying to appeal to the mass market. If nobody outside your team is ever going to use the program, that kind of effort doesn't pay off. Who decided the feature list and designed the user interface? Do they realize the cost and schedule implications?

triangleman2• 12 years ago

Maybe I oversold the feature list... I'm really just trying to do an incremental thing, a "single-page app" that helps tech support agents keep track of their tickets and what's going on with the rest of the team. Because the ticketing system is web based I figure I need to embed the web pages into the app. Then I need to make DOM calls to the page to get some info out of them, or perhaps make various GETs and POSTs to get the info I need, displayed alongside the webpage.

Not a crazy feature list, and like I said I am learning how to do this from scratch so it's kind of a 20% time hobby that I hope turns into something usable in a reasonable amount of time (3-6 months).

goalieca• 12 years ago

WPF died to me because of its complexity. I absolutely hated my last job doing development in that environment.

The complexity of the design was overwhelming and yet pointless. Architected to space it was. Take a look at the signature and inheritance diagram for the "Rectangle" class.

http://msdn.microsoft.com/library/system.windows.shapes.rect...

How can they possibly extend upon that without making it uglier!? Just leave it in maintenance mode like with all the other failed MS GUI frameworks.

forgotAgain• 12 years ago

There is another question beyond WPF being dead or not.

Is there any single UI framework that a dev can invest in for Microsoft's various platforms and have confidence that they won't be moving to some other UI framework in 12 months?

They seem to be in disarray moving forward in this area.

rmsaksida2• 12 years ago

I've had limited exposure to WPF, but I found it absolutely delightful. When you work with web development for so long, working on a desktop app feels very refreshing. The toolkit lets you focus (choice is hard); data binding is awesome; components are all readily available; there's rich IDE integration which mostly just works; you can easily offload almost everything from the backend; and with some good sense, you can drag & drop yourself to a reasonably good looking and usable interface in record time.

danbruc• 12 years ago

You can get quite far with the simple equation XAML + Control Library = WPF = Silverlight = Windows Phone UI = ... some of the runtimes are more powerful, some are less powerful, but they are really similar to develop with. I can see runtimes come and go but the way you build UIs on top of them does not look like going away soon. And - at least to me - this is a good thing, it is a great way to build UIs.

pjmlp2• 12 years ago

WPF is only dead for those that cannot think out of the box.

What matters is XAML and data binding and it is pretty much alive.

Just because the XAML and .NET classes used by WPF are a little different than the XAML and .NET classes from Silverlight or XAML and .NET classes from WinRT, the concepts are always the same ones.

CmonDev2• 12 years ago

What is the out-of-box solution for adorner decorators, then?

randunel• 12 years ago

I felt a similar betrayal that this author felt when Adobe started to drop support for Flash, stopping any development. Their claim of supporting every device with AIR went south not by dropping AIR support (which happened when they stopped Linux support), but by not developing the Flash platform as new devices and OSes poured into the markets. I have since successfully moved on, but I did feel weary for more than a year, unsure where I was heading.

Consider this a mistake, and learn from it: don't depend on any vendor's altruism, always look to diversify your knowledge and options, avoid vendor lock-in at all costs. I believe Apple will be a much bigger drama when this happens.

Is WPF dead: the present and future of WPF

Related Stories