Microsoft’s Bold Plan to Bring PWAs to Windows 10

Posted on October 29, 2017 by Paul Thurrott in Android, Cloud, Dev, iOS, Mobile, Windows 10 with 69 Comments

Microsoft's Bold Plan to Bring PWAs to Windows 10

I recently spoke with Microsoft’s Jeff Burtoft and Aaron Gustafson about the company’s plans to bring Progressive Web Apps (PWAs) to Windows 10. And I couldn’t be more excited about Microsoft’s plans to integrate PWAs with the desktop.

This is a milestone I’ve been looking forward to for a long time. But with the “Redstone 4” release—currently due in March or April 2018—Microsoft will formally support PWAs as a key platform feature of Windows 10, matching and in some ways exceeding what Google has done in Android.

Watching the industry embrace PWAs is exciting on a number of levels. This apps platform is a perfect storm of the right ideas at the right time, a spiritual combination of the cross-platform dreams for Java and the pervasive nature and openness of the web. I’ll be writing soon more about my broader take on the future of PWAs across all client platforms. But for now, I’d like to focus on that most unlikely of platform makers to embrace the future. Yes. Microsoft.

Given how much has changed at Microsoft in the Nadella era, maybe “unlikely” is the wrong word: The once-insular software giant has embraced open source and open standards at a pace that is both dizzying and confusing to folks like me who have followed Microsoft for decades. I’ll get there.

But in the meantime, the new Microsoft is racing forward into the 21st century. And part of its evolving platform strategy is embracing technologies that it did not invent or play any developmental role. It does this to make life easier on developers, which can now target a platform that works on virtually all modern devices, and to make things better for its own customers, who, to date, have been under-served by the Microsoft Store on Windows.

Nothing Burtoft and Gustafson told me is necessarily new information: Microsoft has quietly been trickling out its plans for PWAs over the past year and a half or so, and this is all public information. But aside from me and—OK, just me, basically—few people have really reported on these efforts. Which, by the way, are just about to go public in a big way.

So let’s discuss how Microsoft came to embrace PWAs and how it plans to integrate this technology into Windows 10. This will be news to many, I think.

“Google led the way with Progressive Web Apps, and after a long process, we decided that we needed to fully support it,” Microsoft’s Jeff Burtoft told me. “We weren’t initially sure if it even made sense to align our strategy with theirs, to give it the same priority.”

Of course, Microsoft was already building a set of technologies that might be seen as a PWA competitor, or least an alternative. If you go back to Build 2016, you’ll see Microsoft introducing various “bridges” for developers, ways for them to bring their existing code bases and apps to Windows 10. Among them was something called Project Westminster, or Hosted Web Apps, which lets developers publish existing web apps in Microsoft Store as UWP apps.

“Hosted Web Apps are UWP apps,” Burtoft said. “They’re built with web technologies, hosted on a web server, and run like a website. But they’re really UWPs that can run across Windows 10 PCs, phone, Xbox One, and even IoT. Developers gain access to all UWP APIs, just as they would with a new UWP app built in C# or C++. It’s an immense API layer.”

But just as Microsoft was going down that path, Google started talking up PWAs, a formalized way of thinking about web apps that provided native-ish features like responsive design, security, push notification support, discoverability, and more.

“Google had a similar strategy for web apps,” Burtoft noted. “But from a different angle. Where we were talking about getting web apps into our store, they were talking about browser apps that initially ran only inside of the Chrome browser on Android.”

Google’s initial take on PWAs wasn’t that compelling: The full resources of Chrome needed to load each time a PWA ran, and there was no minimal user interface or runtime. But when Google introduced the notion of service worker, the technological core of what we now know as PWAs, it was a big differentiator. With service workers, PWAs could work like native apps, offering features like offline support, background processing, and more.

“We met with Google about a year and a half ago,” Burtoft told me, “to see if these two things [Hosted Web Apps and PWAs] were really the same thing. And if they were close enough, we decided that we would move forward together, and provide a single way for web developers to build apps that run on all platforms.”

That’s exactly what happened.

As a result, the Hosted Web Apps platform will blend into PWAs on Windows 10 over time, I was told. And Microsoft began work towards that future by implementing service worker and other PWA technologies at the platform level. The foundational work occurred, roughly speaking, during the development of the Windows 10 Fall Creators Update (“Redstone 3”). And it will be turned on, first for Insiders, during the Redstone 4 development cycle.

“We are providing the same support for PWAs in Windows 10 as Google does [in Android],” Burtoft said.

But Microsoft is also making PWAs more native, if you will—native-ish, I guess—on Windows by distributing them through the Microsoft Store. This was a matter of some debate, and it is possible—likely, I’d say—that in a future release of Windows 10, users will be able to install PWAs from a hyperlink in a web browser. For now, however, it makes more sense for the Store to be the conduit for all apps on Microsoft’s platform.

“This is the right approach,” Burtoft noted, after explaining the internal debate. “People understand the Store. You use it to install apps, it manages updates to those apps, and because PWAs inherit real apps features on Windows, they are apps. You can right-click them to uninstall them, and the package sizes are tiny, measuring in KB, not MB.”

There are other benefits to putting PWAs in the Store. Users don’t care “how” an app was made, just that the app exists and is of high quality. PWAs can sit in the Store alongside other app types—“pure” UWPs, Desktop Bridge apps, whatever—and be discovered normally.

Better still, with Google embracing PWAs as well, it is highly likely that the search giant will convert its own massively popular web apps—Gmail, Google Calendar, Google Play Music, whatever—into PWAs, making it a truly universal platform. And that means that these apps, finally, would be coming to the Microsoft Store and Windows 10.

Breathe. This can happen. Probably will happen.

One of the unnecessarily controversial aspects to Microsoft’s PWA strategy is that it is using the Bing search engine to index the web, find PWAs out in the wild, and then put them automatically into the Microsoft Store. Some have alleged that this is Microsoft overreaching, that perhaps developers don’t want their PWAs in the Store, or would like to do that work themselves and perhaps charge customers for their efforts.

Both of these concerns are misplaced.

As it turns out, the way that PWAs are indexed is that Bing searches for explicit instructions contained in a manifest file. These instructions will tell Microsoft whether or not the PWA can be added to the Store. So any PWAs that are automatically added to the Store in this fashion were done so with the OK of the developer. And if a developer wishes to put their PWAs in the Store explicitly and, optionally, charge for it, they can do so.

There are a lot of technical issues to this system, and I don’t want to get too bogged down in that here. But, yes, Microsoft understands that manual curation is required too, and that not all web apps are going to be of high quality. There is a fear that the easy onboarding of PWAs into the Store will result in malicious web developers trying to squeeze their way past the normal Store protections. And Microsoft is sensitive to the notion of its “embrace and extend” past.

“This is not a land grab,” Aaron Gustafson told me. “It’s just another avenue for developers to distribute and promote their offerings. We’re not making money from this.”

Internally, Microsoft is also making its first steps towards a PWA future for its own apps. As you may have heard, Microsoft Teams is being implemented as a PWA so that a single code base can run on any client platform.

This led me to ask about various possibilities.

For example, given that there are what feels like 17 different versions of Outlook out there, would that team consider rallying around a single Outlook PWA code base and sparing us from today’s confusion? After all, one of the beautiful things about the PWA platform is the progressive nature of these apps: They can scale according to the capabilities of the device on which they’re running. That is, on a phone, an Outlook PWA would look and work like a phone app, but on a PC, it would offer more functionality.

“It will happen,” Burtoft predicted. “We are having conversations around the company now [on this topic].”

(It is reasonable not to expect such a solution to happen in the short term, of course. But that Microsoft is discussing this internally is, I think, profound.)

Critics will note that PWAs simply don’t make sense for certain types of solutions. Games, for example. That truly native apps will always “beat” native-ish apps for certain workloads. That’s obvious. But the world isn’t black and white. And just as cloud computing will generally, but never completely, “beat” on-premises servers, PWAs probably make more sense, generally, than native apps going forward.

“There’s always a ‘web vs. native’ rivalry in the press,” Gustafson added. “But I don’t think that fits in this. Yeah, it makes for good headlines. But developers see PWAs as a great option. They may need to go all-in on native code for whatever reasons. But maintaining multiple different versions of an experience—web, iOS, Android, Windows, whatever—is unsustainable. You can build largely the same experience for each by doubling down on the web, by focusing your developments effort there. It’s a huge opportunity.”

PWAs should also have an important and positive impact on such things as Windows 10 S, which can only run Store apps, and Windows 10 on ARM, which will run x86 desktop applications in an emulation mode of which we are not yet confident from a performance perspective. Because PWAs are Store apps, they will run on both of these platforms. And they will help undercut any compatibility or performance complaints about those platforms. PWAs will run great on both.

Great, but what about that schedule? What’s next?

As noted, the foundation for PWAs in Windows 10 was laid in the Fall Creators Update, and developers who are interested in testing PWAs in that release can do so by setting a flag in Microsoft Edge.

Sometime soon, however, this flag will be opened up in Microsoft Edge for users who are in the Windows Insider Program and now testing Redstone 4. That means that Insiders will soon be able to run PWAs and websites that support service worker in the Windows 10 Insider Preview normally.

Microsoft is working with several partners to test PWAs on Windows 10, and while they would not divulge any to me—“they’re names you know”—Burtoft and Gustafson said that the work was centering largely around getting PWAs to look and work well on the desktop.

“Lots of PWAs are geared to mobile today because of Google,” Burtoft said. “But we need developers to think outside of that box. The principles that Google speaks of—responsive design, and so on—are never really illustrated for the desktop. So that is in our hands.”

Once Redstone 4 ships in the Spring, Microsoft will have a “big push” for developers, and it’s reasonable to assume that Build 2018 will be a very public venue for that discussion.

“By Build, those pieces will all be lined up and running,” Burtoft added. “We’re going to talk about it in a wide way.”

Developers who are interested in PWA should check out Microsoft’s PWA Builder website. This will help you build that manifest file that Bing is looking for, and it will point out issues that will prevent a website from becoming a real PWA. It can even generate an AppX file—e.g. a UWP—that developers can submit to the Store. Today, as a Hosted Web App (with no service worker support). Tomorrow as a PWA.

On a side note, one of the most hopeful aspects of this story is how Google and Microsoft are working together. Burtoft and Gustafson spoke often of their collaboration with Google. And I had noted previously that Google referred to their “friends at Microsoft” during a public presentation about PWAs this year.

And it’s not just Google: Because PWAs are all about open web technologies, they are being embraced by all the right players, including Apple and Mozilla. And for its part, Microsoft will let standardization have precedence over native features. For example, Windows 10 supports its own form of push notifications, but if PWA push notifications offer the same rich features, they will favor the PWA tech over the Windows-only stuff.

“Now that we’re moving to a software standard, some of those [Windows-centric] advantages are less important,” Burtoft told me. “Google is following par on this, too, and matching the features we are exposing in PWAs in Windows.”

Some features will, of course, remain platform-specific. There’s probably no reason for Google to support live tiles in Android, for example. But standardization will occur where it makes sense. Which is in most capabilities.

“Developers want to write portable code,” Gustafson added. “When that can happen, it will.”

I didn’t get specifics, but we should hear soon from the PWA folks at Microsoft, both on the Bing indexing front and via the emerging PWA capabilities in the Windows Insider program. And I can’t wait. This is huge, and it’s going to positively impact everyone who uses Windows.

Resources

While I’ve written a lot about PWAs here on Thurrott.com, Burtoft, Gustafson, and others have been spreading the word as well. Here are some PWA resources you might consider for more information.

Yes, That Web Project Should Be a PWA by Aaron Gustafson. Perhaps the definitive take on PWAs at this time. Well worth your time and effort.

Next-generation web apps for education by Jeff Burtoft.

Manifoldjs is becoming PWA Builder, to simplify building Progressive Web Apps by Jeff Burtoft.

Progressive web apps and the Windows ecosystem. Great video session from Build 2017, with Aaron Gustafson.

PWA, HWA, Electron, oh my! Making sense of the evolving web app landscape. Great video session from the Microsoft Edge Web Summit 2017.

Twitter: Jeff Burtoft

Twitter: Aaron Gustafson

Tagged with , ,

Join the discussion!

BECOME A THURROTT MEMBER:

Don't have a login but want to join the conversation? Become a Thurrott Premium or Basic User to participate

Register
Comments (70)

70 responses to “Microsoft’s Bold Plan to Bring PWAs to Windows 10”

  1. MikeGalos

    Combining these two takeaways:

    Hosted Web Apps are UWP apps,” Burtoft said. “They’re built with web technologies, hosted on a web server, and run like a website. But they’re really UWPs that can run across Windows 10 PCs, phone, Xbox One, and even IoT.” and “We met with Google about a year and a half ago, to see if these two things [Hosted Web Apps and PWAs] were really the same thing. And if they were close enough, we decided that we would move forward together, and provide a single way for web developers to build apps that run on all platforms.” and you get:

    Google is building Universal Windows Platform apps as their strategic platform for Android.

  2. jwpear

    This is very exciting. Imagine what we might get if devs can write apps that run and provide a similar experience on any platform and we have some great competition between the app and platform vendors. Consumers win, businesses win, and the app and platform vendors are forced to really innovate to maintain and grow their user base.

  3. davidblouin

    Great news, until Google decides to change the spec and screw things up for everyone else...

  4. adamcorbally

    Really nice read Paul, more of this stuff on Thurrot please!

  5. polymath

    "....And what is the benefit of this on Desktop PC, hmm none."

    the pc is becoming miniaturised, what used to fill a tower now a laptop, soon a tablet or maybe a new form factor.


    local cpu & hard disk is being replaced with 4G LTE , cloud based processing & storage


    its a transition of technology ,, like the gramophone to the wireless, both reproduce sound,,


    radio is more flexible. Later radio has pictures to, then colour pictures,


    then personal communications - the mobile.


    the Microsoft "pc", is going through a metamorphosis


    like Caterpillar to butterfly, maybe.

    do you want to remain an earth bound Caterpillar or take to the sky's like a butterfly?



  6. lwetzel

    Hey Paul,


    Can you suggest some PWA's that exist in the world to look at?

  7. Bats

    Ya know, this is weird. It's weird that this conversation is happening now in Paul Thurrot's website, when it should've happened a number of years ago. LOL...i guess that means, that for years, I have been ahead of the curve. I am reading this Paul's post and I can feel the excitement in the tone. LOL...I guess it's kinda like the time Windows Phone users were so excited for Cortana, when the majority of the smartphone-using population was already using Siri or Google Now.


    Like I've said so many times, the web is kinda like an Operating System. It's been the biggest threat and "silent killer" to Windows. It's even one of the biggest reasons, the PC is faltering. Unfortunately for Microsoft, this is not going to change anything because Google controls the web and the cloud with Chrome and once people are set in their ways, they won't change. Sure Amazon and Microsoft may own the "property", which is the cloud, but Google runs the personal and corporate business on top of it. So, if the PWAs are going thru the store....it won't be a game changer at all.


    So I guess, it's "Welcome to the future Microsoft. For most of us, it's the present. However, just sit back and let the smarter people do their job and you have to do is smile and nod 'Yes'."

    • William Kempf

      In reply to Bats:


      The web is a horrible place for Apps, which is why Paul's excitement here is a bit misplaced. Some of what makes the web a horrible app platform is addressed here... PWAs ditch the browser chrome. Except, those same PWAs still work inside the browser, so the issues there remain issues developers have to deal with. The bigger issue is JavaScript. JavaScript was never meant to be used for applications. It's a SCRIPTING language, and as such just doesn't scale to large application size well, if at all. Cross compilers like TypeScript help here, but it's still an issue. WebAsm and attempts to use existing languages compiled to it promises to solve that problem, but we're years away from that panning out. Finally, there's the problems with HTML as your display technology. Thankfully that's being fixed with things like Grid, but since the emphasis in browsers is on document display, not applications, I'm not convinced we're ready to consider this a solved issue either.


      We've been trying to fit a square peg into a round hole since around 1995, and we're only just know getting to a point where you can see a future for this idea, nearly 25 years later, and still have a few years to go. So, sorry, I can't be excited by this. It may be the future, but it's not the future we should have had, and it seems like a very rocky road ahead for both developers and users.

  8. SherlockHolmes

    And I´m sure MSFT will get this "new" plattform wrong again like with UWP apps.

  9. hrlngrv

    Windows 10 PCs, phone, Xbox One, and even IoT

    Other than apps which duplicate the functionality of radios or kiosks in shopping malls (OK, maybe also Cortana), what sort of apps would more than 5% of users want to run on PCs, phones, Xboxes and any IoT devices? It's probably my own lack of imagination, but I just can't see wanting to run much software both on PCs and IoT refrigerators. And a tangent: which phones would those be? The dwindling number of Windows phones? MSFT hasn't yet sent out a companywide memo that Windows phones are no longer being made and Windows 10 Mobile won't be seeing RS4? Hope springs eternal?

  10. Chris_Kez

    Great article Paul! I love the inclusion of a Resources section as well.

    This is right in line with the kinds of things I was talking about (though I imagined this might be a differentiator for Premium).

  11. polymath

     "........ April 2018—Microsoft will formally support PWAs .."

    nov dec ,,,, jan feb mar apr


    5.. 6 months will be needed to get Mk1 compilers for web assembler in a usable state


    then "desktop apps" can be written which will run in a >>portable<< form in the browser's ... Edge, chrome, firefox, safari ... at native speeds, ie " desktop speeds"


    then with any luck, applications written will run on any processor any operating system at native speed's simply served from a web page link.


    the problem will be, DVD's could have worked world wide, however REGIONS were introduced so you could sell the same disk is many REGIONS, will greed spoil the idea of portable software on universal platforms.


  12. ErichK

    Great article, interesting stuff.

  13. polymath

    this is the text from the talk


    -- this looks interesting...

    Our upcoming release next month (Fall Creator's Update5) will have Service Worker support hidden behind flags in Edge:


    -- this looks interesting...

    One of the big initiatives we are taking is getting your PWAs into the Windows Store with very little involvement from you. The way we are doing that is by using Bing to crawl your manifest.json and relying on the metadata there to create the app that lives in the Store: ....

    We are piloting the entire workflow with a handful of apps (and their developers) right now, and we will start ramping this up as we get closer to full PWA support next year. The goal is to get quality apps into the Store. Coming up with the right algorithms to detect quality is something we are constantly tweaking, but we will detail our quality criteria shortly. You can see where we are leaning towards by referring to Aaron Gustafson's PWA talk6 at Build earlier this year.


    Progressive Web Apps (PWA) and Windows 10

    https:// forum.kirupa.com/t/progressive-web-apps-pwa-and-windows-10/637192


    of the talk.. from P's references

    PWA, HWA, Electron, oh my! Making sense of the evolving web app landscape. Great video session from the Microsoft Edge Web Summit 2017.

    https:// channel9.msdn.com/events/WebPlatformSummit/Microsoft-Edge-Web-Summit-2017/ES04?term=pwa

  14. polymath

    "...Keep in mind, only use PWA if you don't have a >>budget<< for a REAL Mobile App, "


    I think you misunderstand the idea behind pwa's


    you only need to write ONE pwa, it will then run on Chrome, Edge, Firefox & Opera


    there is NO need to write a "REAL Mobile App," for Android, ios, chrome etc,, just the ONE


    combined with WEB assembler, will give a native app experience


    do you understand the IDEA behind pwa?

    • MutualCore

      In reply to polymath:

      PWAs even implemented to newer standards that allow access to things like sensors on mobile devices will never be able to fully access the iOS/Android API for distinct features that are only on those platforms. Native apps will always be king, especially for Snapchat, Apple Music, Google Maps.

  15. Carlos Osuna

    Keep in mind, only use PWA if you don't have a budget for a REAL Mobile App, else the world will know you are a cheap shop with no Native App experience which is trying to lower costs by using inferior PWA apps.

  16. Carlos Osuna

    "a spiritual combination of the cross-platform dreams for Java and the pervasive nature and openness of the web"


    Peyote fueled spirits, I guess, as this is just a rehash of Metro's stupid idea of including HTML/JS on the stack, but then demanding from developers that they develop in a certain way and with JS awful promise API before it was finished.

  17. skane2600

    Despite many decades of failure, the tech industry can't resist the siren call of the "universal platform".

  18. polymath

    this says its a PWA directory


    https:// pwa-directory.appspot.com/


    if so,, take a look



  19. polymath

    is this a web app


    https:// www.theguardian.com/uk?INTCMP=webapp



  20. polymath

    try a PWA now


    https:// lite.twitter.com/content/lite-twitter/en.html


    what do you think?

  21. polymath

    "...The web is a horrible place for Apps, which is why Paul's excitement here is a bit misplaced..."


    in my opinion people have become lost,,, PC's, mac's used .exe programs, later ios and android had "apps"


    if you use YouTube on a web browser, that's an "app" if you like to call it such a thing, it seems to work well, it has text in/out, sound/vision,


    with mobile 4G, soon 5G, you can UPLOAD video at 20..30mb's which is the same speed of current DOWNLOAD speeds, which is the current BOTTLENECK with "broadband internet"


    also, if you have used Skype to chat with some one on the other side of the world, there is no noticeable delay, as the internet now uses optical fiber so packets move at the speed of light around the world.


    its also presentation, what your used to..

    with Chromebooks you can specify "open in window" instead of a TAB, which makes it "..LOOK.." like the windows you get with Microsoft windows or the apple mac.


    local computing "PC"

    Display,KBD --------- wiring------- (RAM,HDD,CPU) << "software"


    Cloud computing "Chromebook"

    Display,KBD --------broadband internet-------- ( cloud (RAM,HDD.CPU) ) << "software"


    what people seem to have misunderstood, is that broadband internet at 10..30mb/s = 3Mbytes/s this is as fast, maybe faster than the resulting link in a normal pc and the "function" of its local cpu/ram+HDD <<-->> Display,KBD ,, as more people come to terms with this, they will begin to understand the transformation that is happening under there noses.



  22. polymath

    "....I've said so many times, the web is kinda like an Operating System. It's been the biggest threat and "silent killer" to Windows. ...."


    Relay?


    the browser, Chrome, Edge, Firefox, safari ,, all need supporting software, or an "operating system",, i suppose


    so the web is hardly a killer of operating systems


    i don't think "..the web is kinda like an operating system .." ,,, it is a packet switched system .

  23. cseafous

    If a site like Southwest.com can turn their existing website into a PWA, they would be crazy not to. This would make my Surface Pro even more useful. Instead of pinning web pages to my start screen, they are just apps that don't take up a lot of space on my drive. There are still a few desktop apps that I need but other than that, I could just use apps from the Windows store. I wish this was a thing when I was still using my 950XL. I have an iPhone 7 Plus but I miss my Windows Phone.

  24. Bart

    I am just trying to comprehend Microsoft is actually at the forefront of this stuff.....?

  25. nbplopes

    These are excellent news. Let’s hope MS seeks it through 200%.


    Cheers.

  26. Breaker119

    Good read! Long ago when working on a major mobile platform we had a discussion about whether platform web apps were the way to go...it seems we were just ahead.

  27. Alin Maior

    If they don't need store verification(even google's lame one), they're a big no for me.

    Also, they can offer no real multithreading, as the core will be javascript.

  28. Jules Wombat

    Why should developers get excited by yet more churn in the Microsoft development story. We get a different story from Microsoft every two years, which is why Developers left the Microsoft ecosystem to concentrate on Google ecosystem, as they still believe in supporting and inspiring Software developers with a clear direction and strategy.

    Google + Android is the future, regardless of anything Microsoft does now.

    • Paul Thurrott

      In reply to Jules_Wombat:

      Maybe because this strategy is cross-platform, open, and makes sense.

      • hrlngrv

        In reply to paul-thurrott:

        Cross-platform to Windows 8.1 and prior?

        • MutualCore

          In reply to hrlngrv:

          WIndows 7/8.1 are increasingly diminishing with every day that passes. Windows 10 is already past 600 million users and growing. Your point about older OS versions is a non-story, especially in 2018.

          • hrlngrv

            In reply to MutualCore:

            Granted Windows 7 and 8.x usage should be declining while Windows 10 usage should be increasing, but Netmarketshare.com currently (OK, for October 2017) shows Windows all versions in total with 90.77% desktop share, Windows 10 with 29.26%, so all older versions combined with 61.51%, or a bit more than 2/3 of all Windows desktop systems in use.

            As for 2018, we'll see how quickly Windows grows. From October 2016 (22.59%) to October 2017 (29.26%), Windows 10 usage grew less than 30% (.2926 / .2259 - 1 = .2953). If its average rate of growth doubles going forward, Windows 10 should reach reach Windows 7's current usage share in a year, still below half of all desktops, though just over half of all Windows desktops.

            I figure Windows 7 usage will remain higher than Windows 10 usage through the end of 2018, but they'll be very close by year-end 2018. Windows 10 won't move past Windows 7 until 2019.

    • Winner

      In reply to Jules_Wombat:

      You were downvoted by fanboys but you do have a point. Microsoft tries things and kills them, stranding developers. They did it repeatedly with Windows Phone. They have a reputation. Let's hope they can actually persist with this or it won't be successful.

    • ym73

      In reply to Jules_Wombat:

      Except that Google is promoting this as well. Even if MS drops this, it will still be useful on chrome. So, were is the risk to developers to implement these features?

  29. chrisrut

    First rate journalism Paul. Thanks. Being copied to my whole tech team.


    BTW, I wonder if this emergent move is related to the failure to allow tabs to be presented as as taskbar-icon'd apps as you've so often requested? Perhaps to do so would have created a situation where you'd deploy it one way today, then find that in conflict with the new technology. If so, that tab-based interface could well come along at the same time - I.e. Redstone.

  30. brduffy

    “People understand the Store. You use it to install apps, it manages updates to those apps, and because PWAs inherit real apps features on Windows, they are apps. You can right-click them to uninstall them, and the package sizes are tiny, measuring in KB, not MB.”


    .... and we can take 30%

    • Paul Thurrott

      In reply to brduffy:

      No. Most PWAs will be free, just as on the web. It's only 30 percent if a developer decides to charge for the app explicitly. "This is not a land grab."

      • hrlngrv

        In reply to paul-thurrott:

        If a PWA developer intends to charge for an app, must they modify their manifest in order for Bing's web scraping to omit them? IOW, if a PWA developer wants to charge for their app AND doesn't want to give MSFT a 30% cut of revenues, MSFT will be sure not to include their app in the Store? That is, is this MUST OPT OUT rather than MUST OPT IN?

        I can see how free, free to try, and effectively shareware could be distributed through the store. I think MSFT would be walking on legal landmines if they sold anything without having physical signed agreements with developers before they made the first sale.

  31. marshalltm

    I would be interested to hear from Paul about what this means for Amazon. Currently, Amazon tablets, but also Amazon TV dongles run on a variant of Android, but without the Google Play Services. This significantly hinders what gets put in their separate appstore. So even though they run android, there is an app gap. PWA potentially closes this app gap, particularly as it relates to Google apps. The echo show just lost YouTube, but PWA might bring it back. PWA with high quality tools in AWS, might end up shifting things heavily in Amazon's direction

    • Brian Mueller

      In reply to marshalltm:

      This is bigger than just Amazon, it essentially tears down the walled garden for software. WebOS, Tizen, Windows phone are all relevant. Throw on Web assembly and that destroys it even further. I don't think MS and Google realize the far reaching implications of this. I predict this will create dozens of new OSes. Many from Google rivals in other countries, Others could be desktop Linux distros or even something completely new. This is really disruptive.

      • MikeGalos

        In reply to Evilsushi:

        Actually, that'd be great for Microsoft. Because of installed base, nobody is running a modern operating system architecture. Windows is tied to the architecture they did in the late 1980s and everybody else is tied to what Bell Labs did in the late 1960s.

        What makes it good for Microsoft is they're the only ones who have been able to build a new operating system architecture in decades despite other companies like IBM and Apple trying and failing multiple times and the falling back on using *ix.

    • Paul Thurrott

      In reply to marshalltm:

      I could see this being a win for Amazon as well. Ultimately, of course, this is really just a win for all of us, regardless of which platform(s) we choose.

  32. matsan

    I'm sure the few developers out there will have one last look at UWP and move on to the next Microsoft platform... Or away...


    No wonder win32 is so prevalent - it was stable for many many years when Microsoft didn't feel the need to invent new platforms every year.

  33. glenn8878

    Pointless without a mobile strategy plan.

    • Jack Smith

      In reply to glenn8878:

      Totally agree. Everything today comes from mobile.

    • jrickel96

      In reply to glenn8878:

      How so?


      On Google's side, they will make sure PWAs work well on mobile. Microsoft is essentially managing the desktop side.


      The kicker is this does help Chrome OS, but it also helps Windows. Andromeda is coming. PWAs that can adapt to your device is exactly what Microsoft is after for Windows AND they also have the long term play on the back end. Azure will be a major player behind the scenes - as will AWS and Google's own cloud services. The larger providers will use multiple cloud vendors. That's what happens now.


      Unlike other MS strategies, this actually CAN eliminate the app gap. It makes Windows 10 in a portable form more palatable as an option, especially for enterprise.


      If you can eliminate the app gap, the main platform question that comes up for business users will be security and reliability. Android has issues on both counts, especially security. While the back end may be secure, people will still store data on their device, so platform WILL matter, but it won't matter so much for the applications available, but for the security and other aspects that improve access.


      PWAs fulfill the promise that MS had hoped for with UWP. They are the same thing, in many ways. Obviously a different platform, but they accomplish the same goal of one app that works across different types of devices - and even different OSes.


      Microsoft has made sure they are everywhere. This makes that easier for them, but it also allows their platforms to have the apps users need.


      If your apps can go with you regardless of platform, it puts pressure on EACH platform to better serve the user. If Android proves to have massive security vulnerabilities and PWAs allow for those apps to go elsewhere - Windows 10, iOS, or some other OS - then users can easily move. The key then becomes setup and creating an interface that users like and want.


      If the same PWAs are on every platform, it increases competitive pressure because it removes that factor from the pains of switching. That puts pressure on the platform designer to give the user the most efficient and enjoyable access to those applications they use.


      It means that Microsoft, Apple, and Google will have to be more responsive to their customers because they will be able to more easily leave.

      • glenn8878

        In reply to jrickel96:

        Again, pointless without a mobile strategy. Closing the app gap does nothing for Windows since no one uses apps on the desktop. What’s the point of downloading web apps when you can just get it from the browser directly.

        • Soundtweaker

          In reply to glenn8878:

          MS has a mobile strategy and that's what ultra mobile PC's are all about. Bringing a scalable desktop OS to smaller devices like next years Surface phone which (thanks to PWA's) will be able to run any app. This is huge for MS and will make Windows the ultimate universal platform.

      • Roger Ramjet

        In reply to jrickel96:

        I see this two potential ways from Google's perspective:

        1/ Maybe this is the ultimate middleware, and Google thinks they use PWAs to rule even over Windows, in which case it no longer matters which OS/platform you use, you will use Google services, and they will monetize that. Essentially, they would be going for (on Windows) what Microsoft has been struggling to do on Android/iOS last 2-3 years, except they start from much better positioning. One could speculate that various things they see in the future (e.g. EU kibosh on Android monopoly, WoARM, even AI/Ambient) are leading them to change strategies. And of course they have won in mobile, an agile company looks for the next thing ...

        2/ If they still care about the Android dominance, why, they could simply not allow Microsoft, Amazon or anybody else they dont want, put Google PWAs into any store. Google services continue to be available online. I guess if they wanted they could go further and restrict device access (a la Echo Show). I don't see facts in these PWA write ups that says Google has decided to enable anything of this sort to any specific non-Google stores, its mostly an assumption that these would be universally available services.

        • jrickel96

          In reply to Roger Ramjet:

          Google really cares about services. I think Android is more of a headache than they'd like and the Play Store does woefully bad at monetizing apps. I get data on that since my company publishes in the various stores. Play Store generates less revenue than even the Windows Store for most developers IF they are there. Now, that's a bit misleading because there are some higher cost desktop apps and games that do skew the Windows Store, but the Play Store generates a fraction of the revenue of the App Store. So Google spends a lot on overhead for the store and really doesn't make that much with their cut. The primary thing they get from Play is tracking users so they can push ads.


          In dealings with Google, that's all they care about. More data about you so they can sell ad space to push them to you. That's the only thing that really makes money for them. I have family that work in medicine and they tried to get whole hospital systems to switch to G-Suite and disregard HIPAA laws because they wanted patient data because they REALLY want to do Big Pharma ads. The hospitals all declined, of course. Seriously, they offered to share ad revenue and cover legal bills and lobby Congress to remove some patient privacy protections.


          To Google, if each app allows you to sign in via a Google account, that gives them a chance at tracking what you're using and who you are, giving them the ability to get data on you and push ads more efficiently.


          Google isn't a tech company. They're an ad company that uses tech to push ads and increase their per ad revenue. They are often willing to break laws and compromise privacy to do so. PWAs would give them a multi-platform in.


          And if most of those PWAs use Google Ads then they lose nothing. It actually opens the door to them getting more detailed data on Windows desktops AND more data on Enterprise users that could be LARGER ad buys. That's what they tried with G-Suite, but it's been a pretty big failure overall. They had hoped to get a lot more marketshare and a lot more data from it.


          But everything they do is about increasing ad revenue. That is all that matters to Google because that's the only way they really make money.


          And look at their last quarter. Big numbers BUT their cost per click numbers are plummeting . Down 18% year on year. This means they need more ads, more intrusiveness to continue to have blockbuster earnings.


Leave a Reply