Welp, in case you hadn’t heard yet – the market for Universal Windows Platform apps has pretty much dried up. So much so, Satya decided to redirect two thirds of the Windows team towards (hopefully) more fruitful endeavors. This doesn’t catch me by surprise; last year’s Windows 10 S variant (Windows RT 2.0) was a bit of a dud.
But we all knew what 10 S really was; a hail Mary for the Windows Universal Platform. I thought it was a neat idea; Microsoft needed to compete with Chromebooks (Netbooks reincarnate). 10 S under the hood was closer to a mobile OS than a desktop OS so it’d be great on those low end machines. The developer tools are just as good as iOS and Android. I was really looking forward to the Qualcomm powered laptops I kept hearing about with hyper long batter life that could run Win32 applications via a reverse engineered x86 architecture. Like I said – neat!
The Achilles heal to Windows 10 S? It didn’t have a browser. Well, that’s a bit of an exaggeration. It had Edge. But more specifically, it didn’t have the browser, Chrome. Robust (ie: useful) web apps have been optimized for Chrome. Most Windows 10 S users got their sweet new premium priced Surface Laptop, tried to load their web app in Edge, couldn’t, and hit the big blue button to install regular Windows 10 so they could use a ‘real’ browser.
Let that sink in for a moment. Users let their browser preference dictate their OS. There were some great benefits to 10 S; mainly significantly longer battery life, easy security and solid performance. But none of that mattered when the browser fell flat.
5 years ago, I opted to invest my skills and precious time into the Windows RT/UAP/UWP tech stack because developing web apps was, well, a royal pain in the butt. There was a lot more fragmentation between the major browsers, each with 20%-35% market share. They all handled HTML, CSS and JavaScript differently. Security was a risk. Windows Apps were a cozy, tightly controlled sand box with easy UI, managed security and that zippy native performance. The app store made distribution a snap for a one person dev shop.
Today, the landscape looks very different. Chrome enjoys a ~62.5% market share on both desktop and mobile. The next closest on desktop is IE/Edge with 12%.
The developer experience for web apps is also far superior compared to 5 years ago. Chrome comes with top tier tools built right in. Web standards have rightfully been adhered to a little more by everyone. JavaScript libraries have gotten faster, reliable, and nicely handle most remaining browser idiosyncrasies. (Though I still think JavaScript is a silly language that we all should love to hate).
The result? Web apps have gotten really good. As good as their native counterparts – at least on the desktop.
So why are native app stores alive and well on mobile, but not on the desktop? First, and probably most importantly, developers have always been able to distribute desktop applications without a store; users are conditioned to get their apps directly. It’s pretty tough to close down an open platform without a PR Disaster. If Windows 95 had shipped as the same closed platform that iOS had, UWP would be killin’ it today.
Secondly, there are some aspects of the mobile use case that make native apps a little better a tool for the job. For what ever reason, it’s still pretty clunky to point a mobile browser to a given url or book mark, especially compared to the desktop. Data connections and processors are likely significantly slower so it’s prudent to download just the data, rather than the app+data every time.
All this being said, Windows as a platform isn’t dead. Don’t mistake the recent shuffle at Microsoft as an admission that Windows has failed. I will still celebrate Windows 10 as a success. It adopted lessons learned from Windows 8 to become an effective, touch friendly, productive operating system. It moved much closer to a service oriented business model. And don’t forget, the heroics it took to move such a large engineering team from a biennial to a biannual release schedule. It did all this while doing what Microsoft does best – support legacy. It’s depended on by millions of people every day and will be the example of a productive OS for years to come.
But Windows as a Platform hasn’t seen much innovation. And frankly, that’s okay. This industry relishes disruptive innovation – but some things just don’t need it. And really, how much more innovation can you expect to squeeze out of a 32 year old piece of software? There are many well known, well loved successful products that are long passed their point of innovation. Products no one would ever call ‘irelevant.’
But let’s go ahead and acknowledge the elephant in the room. Edge. Windows 10 goal of running UWP on all the form factors was admirable, and Internet Explorer probably did need to be rewritten from the ground up. Windows 10 shipped and Edge ended up being good enough – I like the word ‘fine’. And the team moved on and never really made Edge the awesome browser it needed to be. That was the true mistake here. The browser had become just as an important platform as the OS it lived on.
No doubt Microsoft has now realized the importance of the browser as a platform, and is pouring resources into it. The real question though, is it too late? I’m confident in a year or two, they’ll have it at feature parity with Chrome. But that’s only half the battle. They have 22 years of Internet Explorer’s lack luster reputation to dig their way out of. Could this be Windows Phone all over again?
Edit, October 21, 2022: Since the original post, Microsoft did re-release Edge with the Chromium engine underneath. It’s now bumped to 5.83% of the desktop market share. Steady progress. We’ll check back in in another 5 years and see how they, and Firefox are doing.