UWP really is dead (Edge is now Win32 app?)

52

I was actually excited about Edge switching to Blink because I assumed Edge would remain a UWP app but it would get updated via the Store for once, it was going to be awesome. But it appears they’re just scrapping Edge altogether. The new Edge is a Win32 app with a blink engine. I think this is just as big a story as EdgeHTML being scrapped. UWP really does seem to be dead. If Edge is now a win32 app that uses blink, this is painful gut punch to Windows

Comments (53)

53 responses to “UWP really is dead (Edge is now Win32 app?)”

  1. Paul Thurrott

    I don't believe that Edge is a "real" UWP app today. But yeah ... UWP has gone nowhere regardless.


    Also, there's no reason that the Windows 7, 8.1, 10, and macOS Edge apps couldn't all be different.

    • sevenacids

      In reply to paul-thurrott:

      Well, technically it is - for the most part. I mean, it still works on Windows Mobile, although it's a pretty old version of it by now. But this is basically just the shell we're talking here - EdgeHTML was never limited to UWP from a technical standpoint. It was just artificially limited to it.


      I always wondered why they didn't turn EdgeHTML itself cross-platform, but I guess the answer is easy: This thing is so heavily based on COM and DirectX, two technologies only implemented for Windows, that it would possibly be easier to start it over from scratch, which would be an immense undertaking. They should have done it when they split it off from MSHTML, but I think that happened at a time when the people in charge still believed that the world will remain dominated by Windows, despite the strong existence of iOS and Android.


      I have mixed feelings about this Chromium-decision as this will further narrow down the browser engine landscape, and make the web more dependent on a single technology that might be open source on one hand, but on the other hand only a few people will be in charge to make the rules where this thing goes.


      Last but not least, supporting Windows 7 seems a bit odd to me given that it's going out of support in less than two years. What's the point in this? Are we going to see another XP-era? Maybe Microsoft sees a lot of money in it given they will support it for people who are willing to pay.

      • Chris_Kez

        In reply to sevenacids:

        Bringing this Edge replacement to Windows 7 opens up the possibility that it could be a default browser (or at least an option) for organizations who have not transitioned to Windows 10. In some ways, this could ease the transition at least a little bit, both for organizations and for Microsoft.

      • driftsk

        In reply to sevenacids:

        I think making New Edge compatible with Windows 7 makes a lot of sense, especially as the OS is guaranteed to be the next XP: massive userbases across business and corporations that will take years to migrate. They need a browser as that is all that matters - the only real fallback - when the rest of the system moves to legacy and then obsolescence.


        Take Surface RT and RT 2 for instance: they can still be used today, albeit slowly, because the browser still works while the native app ecosystem has faded away into abandonware and the RT Store is a wasteland.


        In other words, it's a good thing to have a modern browser because that can keep an entire system afloat. Microsoft knows there are Win7 users out there and they can't really be thrown under the bus - they tried that in the past and failed repeatedly, perhaps this time they learnt the lesson

  2. madthinus

    Considering the sunset period of Windows 7, why are they bringing Edge to Windows 7? Edge will probably take 6 months to a year to port accross. That leaves very little time in market for it.

    • longhorn

      In reply to madthinus:

      Most mainstream browsers will support Windows 7 far longer just like they did with XP. Chrome and Firefox probably Windows 7 EOL plus 2 years at least. It's a massive user base and it is moving slowly if at all.


      • christian.hvid

        In reply to longhorn:

        Also, it's highly dubious if Microsoft will in fact sunset Windows 7 just thirteen months from now. The market share will, at best, have dropped to 25% at that time. And even if this is an opportunity for Microsoft to extract additional support fees from enterprise customers, there will still be millions of consumers who continue to run Windows 7 but won't be paying for patches. This leaves Microsoft with a hard choice: let all those computers become a malware target, or continue to support Windows 7 for free until the market share is negligible. I have a feeling they're gonna go for the latter.

    • hrlngrv

      In reply to madthinus:

      What Win32 software can run under Windows 8.1 but not Windows 7? Windows 8.1 has 4+ more years of support. Edge will be available for Windows 8.1, so also for Windows 7.

  3. christian.hvid

    The definition of what constitutes a UWP app has been somewhat in flux for several years now. Originally, it meant that the app was written against the Windows Runtime API:s (WinRT), but these days the term is applied to just about anything you can download from the Microsoft Store.


    But WinRT is definitely not dead - it's the very foundation of the Windows 10 UI. If I recall correctly, WinRT was specifically designed to be powerful and performant enough that systems developers within Microsoft would agree to dogfood it (which was never the case with .NET Framework). And for Windows 10 specific work, it's simply unbeatable.


    That said, WinRT/UWP never became the mainstream application platform that Microsoft had hoped for. One of many reasons for this is possibly that Microsoft never made an effort to train developers. Jerry Nixon and Andy Wigley did a couple of tutorial videos, and of course Bob Tabor did an excellent series. But that was years ago - today, there seems to be absolutely no current training material available for would-be UWP developers.

    • maethorechannen

      In reply to christian.hvid:

      But WinRT is definitely not dead - it's the very foundation of the Windows 10 UI.


      And yet there's also the backwards compatible, open source version, WinUI with that "native view embedding" feature. Which doesn't exactly inspire confidence, especially given Microsoft's track record.


      Maybe it won't be thrown under a bus like Silverlight, but it doesn't exactly scream guaranteed way of the future either,

      • christian.hvid

        In reply to maethorechannen:

        I was under the impression that WinUI is nothing more than the UI components of the Win10 platform packaged as a standalone library. The purpose of WinUI is to allow developers to target the latest version of the UI without regard to which version of Win10 the user is running. The fact that WinUI is open sourced is - if anything - a guarantee that it won't go away.

        • maethorechannen

          In reply to christian.hvid:


          The way I look at it, once people start targeting a standalone library it becomes a lot easier for Microsoft to swap out what that library is sitting on top of or switch people over to another library that is API compatible. So if at some point they do to WinRT what they've done to other products then they won't face much developer backlash if all it means is changing a couple of lines in a csproj file.

    • hrlngrv

      In reply to christian.hvid:

      . . . WinRT/UWP never became the mainstream application platform that Microsoft had hoped for. . . .

      With the exceptions of MS Office and nearly everything from Adobe, every other software title I use has Windows, Mac and Linux versions. Mac and Linux markets are tiny compared to Windows, but only Windows 10 runs UWP software, and Windows 10 is only now matching Windows 7 usage. Windows 8.1 and prior, Mac and Linux are a much bigger market still than Windows 10.

      Maybe ISVs' indifference to UWP can be explained that simply. If so, there's a chance ISVs will do more with UWP after Windows 7 reaches EOS.

      Or it could mean there's some software for which UWP offers nothing. I use GNU R, a statistics system which by default is character-mode. I doubt UWP would offer any benefits for that. The GUI front-end I use is RStudio, which is based on Qt and runs under Windows, Macs and Linux. I doubt there'd be much benefit to developing a UWP version along with the Qt/multiplatform version. As for other alternatives, MSFT distributes R Open, it's own version of R, yet MSFT hasn't released a UWP version. If UWP is just so darned wonderful, why hasn't MSFT developed more of its own higher-end software using UWP?

      At the very least, if MSFT shows no urgency for using UWP, why should any ISV?

      • christian.hvid

        In reply to hrlngrv:

        I agree completely. Microsoft has put UWP to good use on the system side of things (like the Win10 UI and supporting apps), but completely ignored it for general app development. This is partially due to the corporate infighting at Microsoft (see Brad's book), but for the most part they just made the same rational calculation as the ISVs: a Win32 application runs on 100% of the Windows PC:s out there; a UWP application runs on substantially less than 100%.


        If your application only needs to target Windows 10, UWP is absolutely the way to go. But that's a rare case, and it won't be more common in the future given the cross-platform requirements that most apps have today. In other words, UWP will never be the right model for mainstream app development, but nor will any other native technology. Apart from already existing applications, classic Win32 will become just as niche as UWP. But neither will die.


      • behindmyscreen

        In reply to hrlngrv:

        They prefer you use VS Code for R Open?

  4. rob_segal

    The problem goes beyond. Windows native app development is in a coma right now. There seems to be more app engagement in macOS there is with Windows users. Markdown apps, email and calendar apps, media creation apps, all seems to be active in macOS. Even the LastPass app there has fingerprint login. The Windows LastPass app hasn't been updated since Windows 8. I haven't been able to find a decent email and calendar app that supports Google on Windows. There are several really nice options on a Mac. Markdown apps on Windows is unimpressive.


    It doesn't matter what the platform is (WinForms, WPF, UWP). Developers are not building new, good Windows applications anymore. Leveraging web apps might be the only way to possibly get new, high quality apps in the Microsoft Store. If that doesn't work, I'm not sure what else Microsoft can do.

  5. Tony Barrett

    Windows app development is indeed in a stall at the moment, win32 or otherwise, and UWP app development is dead, no doubt. All developer focus seems to be Android, iOS or cross platform web with PWA likely to grow massively in the coming years. Why should devs write 'locked in' apps these days for an OS that really isn't going anywhere. Don't get me wrong, for legacy stuff, Windows is still great, but it's going forward that there doesn't really seem to be anything 'native' going on. Windows used to have all the exclusives - it's where the action was. Not any more.

    • skane2600

      In reply to ghostrider:

      Developers are just as "locked in" to iOS or Android if they develop for those platforms as Windows devs are into Windows. Using cross-platform tools result in less optimal apps on each platform, but nothing is stopping developers from including Windows as one of the platforms should they choose to take that approach.

  6. behindmyscreen

    If they switch to Blink, that means the windows store can open up to any browser that wants to use Blink as a rendering engine (a-la Chrome)....making Windows 10 S mode more realistic for people.

  7. JimP

    UWP was dead when they killed Windows Phone (Mobile).

    • Tony Barrett

      In reply to JimP:

      That's the truest thing said in this thread. UWP was developed for mobile, but just happened to work in desktop Win10 as well. When MS killed Windows Mobile, UWP, and it's reason for existing, died as well.

      • skane2600

        In reply to ghostrider:

        Starting with Windows 8/RT, Windows was designed to accommodate mobile, but it's not as if running on the desktop was just luck. It was planned to have that dual operation from the start.


        But it's true that UWP only provided value on Windows mobile because on the desktop it at best would offer redundant capabilities and in fact only provides a subset of legacy Windows functionality.

  8. Ron Diaz

    When was UWP ever alive anywhere but in Microsoft’s fantasies?

  9. willr

    Yeah this is all really depressing. Edge's touchpad scrolling is amazing and now it's going to be like Chrome's. Same with touchscreen scrolling

  10. ggolcher

    Not necessarily. I would look at the Skype example: there's a Win32 version of Skype, an identical UWP version of Skype for Windows 10, and an identical macOS version of Skype. All three look and behave the same but undoubtedly have different platform-specific code.

  11. shameermulji

    Like yourself, to me this signals that UWP is eventually dead. Long-term, the app story on the Windows platform will be PWA's (ie: Electron) & containerized Win32 (ie: Centennial) apps.

  12. waethorn

    "If Edge is now a win32 app that uses blink, this is painful gut punch to Windows"


    It's not really a gut punch to Windows, nor its users. It's a gut punch to developers that towed the Microsoft line over UWP being the future of app development.

    • hrlngrv

      In reply to Waethorn:

      How many new & exciting development technologies had MSFT introduced then axed before it got to UWP? Any developers who put much time an effort into UWP weren't paying attention.

      • karlinhigh

        In reply to hrlngrv:

        MFC, COM+, ActiveX, VB6, .NET Forms, WPF, Silverlight... am I missing any?


        No, I'm not wishing for those things back. Just maybe pieces of them here or there, where Microsoft completely nailed a certain feature, only to move away from the near-ideal thing and never return.


        What I want to see in a development platform is something like what SQLite has.


        "The intent of the developers is to support SQLite through the year 2050... Our goal is to make the content you store in SQLite today as easily accessible to your grandchildren as it is to you."

        <https://www.sqlite.org/lts.html>

      • christian.hvid

        In reply to hrlngrv:

        Fair point, but unlike many other development technologies that Microsoft has forced upon the world over the years, UWP/WinRT is the native app model of Windows 10. This means, essentially, that it's going to stay put regardless of whether it gains any traction in the developer community. Killing UWP would mean killing a host of apps that are actually part of Windows 10, and that's very unlikely to happen.

        • hrlngrv

          In reply to christian.hvid:

          . . . UWP/WinRT is the native app model of Windows 10. . . .

          Granted, but by the end of 2015 Windows 10 had less than 10% desktop/laptop PC user share. By the end of 2016, maybe 25% share, by the end of 2017 maybe 38% user share, and by the end of 2018 maybe more user share than Windows 7 but still not quite 50%.

          Why would developers go all out for UWP until it reached 50% user share?

          I realize MSFT hoped ISVs would be credulous enough to believe MSFT's oft-repeated claims of 1 billion Windows 10 devices in 3 years (this summer just past), but sadly for MSFT it seems ISVs have learned to be rather skeptical about MSFT's public prognostications, especially the blatantly self-serving ones.

          Maybe UWP will finally catch on now that Windows 7 is nearing EOS and Windows 10 has caught up with its user share, but I'm not holding my breath.

          • curtisspendlove

            In reply to hrlngrv:


            Maybe UWP will finally catch on now that Windows 7 is nearing EOS and Windows 10 has caught up with its user share, but I'm not holding my breath.


            The problem is that it is expensive to rewrite an app. That includes dropping a new interface on top of an existing app.


            Microsoft tried to make this easier and cleaner. But if you are going to do *any* rewrite, it makes sense to target a more general platform (web technologies), not yet another Windows-specific framework.


            By far, as long as W32/64 is supported, it is silly to do any major rewrite form most applications.

          • christian.hvid

            In reply to hrlngrv:

            We don't disagree. I just wanted to point out the difference between "UWP is dead" as in "UWP is dismissed by the mainstream developer community", and "UWP is dead" as in "UWP is no longer supported or technically possible to use". The former happened from day one, the latter will probably not happen in our lifetime.

    • lethalleigh

      In reply to Waethorn:

      What developers would that be exactly? It was developer apathy that doomed both Edge and Windows Mobile to an early grave. Don't get me started on UWA's either, where were these developers creating vibrant, fully featured must have apps then?

  13. FalseAgent

    Web browsers have no business being built on another platform - web browsers ARE the platform, increasingly. I bet Microsoft is also working on getting all those Electron and React Native apps to work better on Windows instead of the puke we have now.


    I think UWP still has its place, but UWP is a Windows thing, and developers don't want native. They want cross-platform. Native is only happening on Mobile. Those interested in building a native windows app that works with modern features like touch and whatnot will still want to pick UWP. Or will we see a resurgent Win32? I don't know. Who the hell knows what developers are up to.

    • shameermulji

      In reply to FalseAgent:

      "developers don't want native"


      Depends on the app. If an app is just an extension of a cloud service then you may be right but if it's a tool for higher-end needs (ie: Office, Photoshop, AutoCad, etc.) then nothing beats native, be it native on mobile or desktop.

      • lvthunder

        In reply to shameermulji:

        You haven't been following Autodesk too closely. They rebuilt AutoCAD to run headless so they could run it in the cloud and turn AutoCAD into a web app. It's still early days for the UI since they don't have all the commands yet, but it runs a whole lot better then I thought it would.

        • skane2600

          In reply to lvthunder:

          I doubt that many professionals are going to use a web-based version of AutoCAD. It would be slower and wouldn't offer any compensating advantages.

          • robincapper

            In reply to skane2600:

            Take a look at Autodesk Fusion 360. More complex than AutoCAD. Data & UI in cloud or light local install.

            • skane2600

              In reply to robincapper:

              Autodesk categorizes Fusion 360 differently than AutoCad. They say the former is primarily an engineering tool for mechanical design while the latter is a general purpose drafting tool. Since they are entirely different products it's impossible to compare their performance although it's hard to imagine that a web based tool could outperform a native one given the extra round-trip delay over the Internet.


              These days companies seem to prefer the steadiness of income that comes from a subscription model over the higher performance of a downloadable native app. It's not a technology-driven strategy.

          • behindmyscreen

            In reply to skane2600:

            They can also sell an on prem version. It still saves time and money to support from a server room with a cloud front end in the enterprise.

            • skane2600

              In reply to behindmyscreen:

              There's a lot of considerations that vary from one enterprise to another. Running on a local server would certainly decrease the round-trip time but it might be decreasing the performance by requiring the computing to take place on a single server rather than taking advantage of the considerable processing power of each Windows workstation. Vendors switch to a subscription model precisely because they think it will provide them with more money, not less which means the customer pays more. Support costs don't necessarily drop significantly if you place an application on a server. It depends on the application in question and specific aspects of the operation.

  14. Bob Shutts

    In the Mac arena all apps will have to be 64-bit after Mojave. This seems to be a mixed blessing as a lot of devs are saying "screw it, we're not going to the bother of re-coding."

Leave a Reply