Microsoft and Google engineers are now working closely together
Microsoft CEO Satya Nadella was frustrated.
Microsoft’s Edge web browser, released in 2015, hadn’t made much progress by the summer of 2017. “Satya came to us and said, ‘Hey, I’d like to see us make more progress on the browser,’” says Microsoft’s Joe Belfiore, in an exclusive interview with The Verge. Nadella pointed out where the web was heading, and he wanted more from Microsoft’s Edge browser.
“Google is doing well with web-based collaborative experiences, the Office team is investing more to improve ours, we want the web browser to be better,” says Belfiore, reciting Nadella’s comments. “Edge had a pretty mixed reputation,” he admits, but Microsoft had been spending a lot of its time trying to improve its browser compatibility, all while Google was marching ahead with Chrome, which was solidifying its position as the de facto default web browser on the desktop.
Something had to give. Microsoft had to change its Edge browser in a big way. That meeting with Nadella ultimately led to Microsoft’s huge decision to jettison the browser it built in house and start from scratch using Chromium as a new foundation. The stakes for success couldn’t be much higher: the future of Windows and the web itself could hinge on this project.
This is the story of how Microsoft made that monumental decision and what could happen next.
Shortly after that meeting with Nadella, Microsoft’s browser team started to analyze everything that was wrong with Edge. It was a way to spark a discussion internally about the changes needed.
“We wrote a paper. We wrote the paper for the purpose of having a discussion at an offsite that Terry Myerson [former Windows chief] had in October of 2017,” says Belfiore.
This paper included a bunch of the benefits and drawbacks of Edge. Microsoft picked a different term for the drawbacks, though. It called them “headwinds.” It was a signal that, in 2017, the problems with Edge weren’t just technical, nor were they necessarily insurmountable. They were just — theoretically — the things that were slowing down its adoption.
One of the strongest headwinds might surprise you: distribution. Despite Windows 10’s growing install base, there weren’t enough people using Edge. People were still picking Chrome and other rivals. “Our volume use is low, partly because we’re only on Windows 10,” admits Belfiore. “In the global world of all the devices, although it’s huge, [Windows 10] is a minority.”
Only being on Windows 10 meant that businesses still running Windows 7 couldn’t get Edge for those machines. Even if a company had switched to Windows 10, that didn’t mean they would be willing to adopt new versions of the OS right away. That meant their Edge browsers would only get updated if they took the very latest feature updates to Windows 10, which are released roughly every six months. Given how enterprises like to hold back and test Windows versions, this was a recipe for Edge getting seriously behind the times.
That adoption headwind was nearly indistinguishable from the second big gust: compatibility. Because Edge used a different rendering engine than Chrome or Safari, it meant that it would sometimes have problems on websites. Testing a website against multiple browsers has always been difficult, and because Edge had so little uptake, it meant optimizing for it often fell off the priority list for web developers.
Edge is also built on Microsoft’s Universal Windows Platform (UWP), the company’s previous big push in Windows 8 and Windows 10 to get single universal apps that run across desktops, tablets, phones, Xbox consoles, and devices like the HoloLens. “Our third headwind was UWP. And it’s not that UWP is bad, but UWP is not a 35-year-old mature platform that a ridiculously huge amount of apps have been written to,” explains Belfiore. That meant things like multiple monitor support weren’t always solid for UWP, and the Edge team would have to wait for general UWP improvements. Microsoft had to get Edge back to a real desktop app, available across Windows 7, Mac, and Windows 10.
Microsoft considered a number of different ways to deal with all of those headwinds. It looked at seemingly simple solutions, like getting Edge into the Windows Store so it was updated more often than the core OS. It also looked at trying to make Edge available on Windows 7. Considering moving to Chromium was on the list of ways to fix these problems, too.
“We had this meeting and conversation, and we didn’t decide at that point to move to the Chromium engine. We considered it, we said, ‘No, we think we can get there with compatibility,’” says Belfiore.
So Microsoft threw more engineers at trying to fix its Edge compatibility issues, but it was a whack-a-mole approach: Microsoft would fix one issue in Edge, and another 10 would pop up. The web has been moving at a far more rapid pace in recent years, with Google pushing a lot of new standards and speed with Chrome, and Microsoft didn’t have the structure to keep up.
One of the other big issues for Microsoft’s web ambitions was the way it used to be organized as a company. When Edge began, web platform work for Windows (EdgeHTML) was handled by a separate team than the one building and maintaining the Edge browser itself. Microsoft reorganized its Windows business in a big way last year, with former Windows chief Terry Myerson departing the company, and Microsoft’s web platform and Edge teams converging under Joe Belfiore for better accountability and ultimately a better browser for Windows.
This reorg put an even bigger spotlight on Edge’s problems, especially now that the web teams were aligned to improve the company’s browser efforts. Microsoft started building new browser prototypes on different implementations of open-source technology to see what was possible. One, codenamed “Blade,” tried to take the existing Edge application and add the Blink rendering engine behind the scenes. Another prototype, codenamed “Septagon,” actually fully implemented Chromium. The Septagon browser prototype was the obvious choice. But that switch was such a big decision that it would have to be made at the top.
Nadella’s leadership team meets every Friday, and they have a process dubbed “Researcher of the amazing.” It’s designed so each product team can put forward people who are trying interesting things. Coincidentally, at the point the Windows team was prototyping new versions of Edge, it was their turn to present their own “researcher of the amazing.” The Septagon prototype made it to the senior leadership meeting, and the feedback was good.
Microsoft then spent months doing a serious review of what it would take to even consider moving Edge to Chromium, as it was a big shift in the way the company typically does things. “We did a little bit of a roadshow,” explains Belfiore. “We went and met with Bill Gates, we went and met with Kevin Scott, our CTO who came from LinkedIn, and Reid Hoffman who was on our board.” The team also met with Nat Friedman, who is now the company’s GitHub CEO.
MICROSOFT SPOKE TO BILL GATES ABOUT THE SWITCH
The aim in all of these meetings was to get some outside perspective on switching to Chromium and the challenges involved and what sort of relationship could the company expect with Google. Switching to Chromium, after all, is a big bet for Microsoft and an even bigger adoption of open-source code.
Some of those challenges involved how to have a good relationship with the open-source community, but others were fundamental changes to the way Microsoft’s Windows engineering teams operate. Edge shipped to the public every six months, but the Chromium Canary developer builds ship every single day. Chromium is also a different codebase that Microsoft’s engineers would have to understand and contribute to.
Microsoft spent a bunch of time analyzing that before deciding in September 2018 that it was going to adopt Chromium. Nadella and his leadership team signed off on the change, and Microsoft then spent some time getting ready to announce it publicly in December. It was a big moment.
“We were a bit nervous,” admits Jatinder Mann, a Microsoft Edge program manager. “The fortunate thing is, as soon as we made the announcement, we got a lot of positive responses from Chromium engineers, from other Chromium browser vendors that were pretty excited to see us join this community.” The reaction was mostly positive, and other browser engineers saw this as a way to get more help to improve the web.
Mozilla, the company behind Firefox, didn’t welcome Microsoft’s move. It could mean that web developers will be less likely to code to web standards that would with any browser and more likely to just code for Chromium and Safari. It’s a real risk, and so Mozilla says it will continue to “fight for a truly open web.”
Once the decision had been made and then made public, the real work of switching Edge to a Chromium base began. Google and Microsoft engineers had been used to working together to help design web standards before, but this would be very different. Microsoft’s adoption of Chromium has meant that those relationships have grown even closer over the past six months.
There have even been some moments of having to decide between communicating with Google Hangouts or Microsoft Teams or using Word or Google Docs, but Microsoft engineers are willing to use whatever works best for Chromium contributors. “We recognize that we’re joining this community, so we wanted to collaborate in whatever they were using already,” explains Mann. “We use Google Hangouts when we’re talking to Chromium engineers, just to make it easier for us to collaborate.”
Google’s engineers talked to Microsoft about how contributing to the Chromium open-source community usually works, and Microsoft’s engineers have adopted these processes. “A lot of our principles aligned well with these best practices,” says Mann. Those practices include being part of the Chromium project for the long run, maintaining any contributions Microsoft makes, respecting architectural design, and cross-platform and cross-device needs.
“Even when we talked to Chromium engineers about best practices, they were pretty helpful,” says Mann. Google has even suggested good bugs for Microsoft engineers to go fix to learn the code base. It’s a learning process for both companies, but the engineers involved ultimately want to work together to improve both Edge and Chrome on Windows. For Microsoft, it benefits all Windows users if both browsers work better.
Microsoft and Chromium engineers are now working together to address things like accessibility features, scrolling, and touch controls. Touch has been a big part of Edge, and Microsoft is now contributing changes to Chromium that will make picking a date or time in a web form far more touch-friendly. Microsoft is also working on better support for the Windows touch keyboard in Chromium, including swiping on the keyboard to write.
“The fascinating thing for that area was Chromium engineers had started that project at some point. I think they took a hiatus or some other priorities took over, so we actually picked up the project they had started and finished it off,” explains Mann. “We had a meeting with them where they’re talking about ‘this is on our agenda for next month, next quarter’ and we’re like, ‘Oh, we’re working on that right now’ and there were some delighted faces on the other side.”
It’s clearly early days for Microsoft’s Chromium adventures, and engineers from Microsoft and Google seem to be collaborating well toward similar goals. But Microsoft and Google are still fierce competitors, and we’ve yet to see how that will play out in the browser space. Previously, we’ve seen Chrome-only sites from Google that haven’t worked well in Edge, and a former Mozilla executive recently accused Google of sabotaging Firefox for years, despite having a close working relationship and a search deal to keep Google as default in Firefox.
“Google Chrome ads started appearing next to Firefox search terms,” explained Johnathan Nightingale on Twitter. “Gmail and Google Docs started to experience selective performance issues and bugs on Firefox. Demo sites would falsely block Firefox as incompatible.” Nightingale says there were dozens, perhaps even hundreds, of what he calls “oops” incidents where Google would respond to Mozilla and say the changes were accidental and would be fixed. “I’m all for ‘don’t attribute to malice what can be explained by incompetence’ but I don’t believe Google is that incompetent,” Nightingale wrote.
Google and Microsoft have had some pretty public spats over the years, which have resulted in ecosystem wars and affected customers of both companies. Windows Phone users were temporarily cut off from Google Maps, there was a bitter battle over a Windows Phone YouTube app, and Google surprised Microsoft by cutting off Gmail’s Exchange ActiveSync support for Windows Phone. Combined with Google’s refusal to develop Windows apps, all of these incidents served to undermine Microsoft’s mobile efforts to compete with Android.
YouTube engineers even hatched a secret plot to kill off Internet Explorer 6 nearly 10 years ago, and, more recently, YouTube hasn’t worked well in non-Chrome browsers like Edge, Firefox, or Safari. Given this history, some fear that Google could use its many web properties and services to undermine Microsoft’s move to Chromium. It’s not an unfounded fear, but it is an ironic one: Microsoft could end up being the victim of an “Embrace, Extend, Extinguish” strategy if this relationship goes sideways.
These fears, unfounded or not, became heightened recently when Google Meet suddenly stopped working in Microsoft’s new Edge browser. Google also recently added an “unsupported browser” warning to Google Docs when Edge Chromium users use the service. Eric Lawrence, who worked on Internet Explorer before joining Google to work on Chrome and is now back at Microsoft, revealed on Twitter that Google’s recent changes probably aren’t malicious. Google isn’t trying to purposely block Microsoft’s new Edge browser.
“I think our expectation is that Google is going to compete with us,” says Belfiore. ”What we hope can happen is that we are engineering collaborators on web standards and a web implementation that works well on Windows devices. That seems possible, and that is in everyone’s best interest. And there’s plenty of room to compete in differentiation.”
Google, unsurprisingly, has welcomed Microsoft’s contributions to Chromium. Google engineers invited Microsoft onstage at a recent BlinkOn Chromium event, and Microsoft has contributed more than 400 merges to Chromium so far. “Chrome has been a champion of the open web since inception and we have welcomed Microsoft to the community of Chromium contributors,” says a Google spokesperson in a statement to The Verge. “We appreciate the work we’re doing with Microsoft and the web standards community in advancing the open web, supporting user choice and delivering great browsing experiences.”
MICROSOFT IS STILL FIGURING OUT HOW ITS CHROMIUM PUSH WILL GO
Microsoft now has a lot of work to do until this new version of Edge is ready for all Windows users, and there are still a lot of unknowns. Microsoft also runs the risk of running into the most classic of Windows problems: backwards compatibility holding back future development. It’s likely that a future version of Windows 10 will have to support the existing Edge browser, Internet Explorer 11, and Edge Chromium all at once.
It’s not going to be an overnight switch from old Edge to new Edge, and Microsoft is still figuring out how it’s going to approach that. And don’t forget that web rendering technology isn’t just a thing that happens inside a browser tab, it happens in apps and all sorts of surprising places in an operating system. How Microsoft will address those challenges remains to be seen.
The software giant is demonstrating some new features it’s planning for Edge Chromium today, and a Mac version should be coming very soon. A new “IE mode” for Edge is coming, too, which might help some enterprise users who still rely on ancient internal sites designed for Internet Explorer to switch to something more modern. A new “collections” feature for Edge is designed to improve the irritating experience of collecting information and notes from the web. Microsoft is even adding in granular tracking controls to stop ads from following you around the web. You can read more about these new Edge features here.
Beyond these features, Edge Chromium and Microsoft’s improved relationship with Google are still very much works in progress. It feels like Microsoft is still working out the details on a project that’s very new for the company. It might all be new, but the end result means if you’re using Windows then both Chrome and Edge are going to get a lot better in the coming months. Edge already feels as good as Chrome, and it could easily become the best default browser Windows has ever had. As good as Chrome might not be enough, though. It’s now up to Microsoft to convince everyone why it’s worth ditching Chrome.
You must log in to post a comment.