The company’s browser will still be named Edge and should retain the current look and feel. The decision to switch was motivated primarily by compatibility problems: Web developers increasingly test their pages exclusively in Chrome, which has put Edge at a significant disadvantage. Microsoft’s engineers have found that problematic pages could often be made Edge compatible with only very minor alterations, but because Web devs aren’t using Edge at all, they don’t even know that they need to change anything.
The story is, however, a little more complex. The initial version of Edge that shipped with the first version of Windows 10 was rudimentary, to say the least. It was the bare bones of a browser, but with extremely limited capabilities around things like tab management and password management, no extension model, and generally lacking in the creature comforts that represent the difference between a bare rendering engine and an actual usable browser. It also had stability issues; crashes and hangs were not uncommon.
Microsoft’s own telemetry showed that many users did give Edge a chance, but as soon as a problem was encountered—a crash, a hang, or perhaps a page that didn’t work right—they’d switch to Chrome and never really look back.
In each successive feature release of Windows 10, Edge has been improved. It’s still not the most feature-rich browser, but it’s just about good enough, and, while it has only a limited number of extensions, they can fill some of the functional gaps. The browser is substantially more stable than it used to be, and its site compatibility is nowadays very good, with incompatibilities relatively rare. Much of this compatibility improvement comes from Microsoft engineers investigating problem sites, determining the minor fixes needed, and making Edge itself fix the sites.
But it turns out this isn’t enough to bridge the compatibility gap. If every Edge user were using the very latest version of Edge, it wouldn’t be so bad, but that’s not the case, and that’s because of how Microsoft has bundled Edge with Windows 10. Most home users will end up running the latest feature update to Windows 10 within a few months of its release. But enterprise users are more diverse. An enterprise that has rolled out Windows 10 version 1709 to 10,000 seats may follow that by deploying version 1803 to another 10,000. The systems with 1709 will stick with 1709—after all, it’s still supported—as the enterprise focuses its resources on getting legacy systems onto “some version of Windows 10,” rather than getting every Windows 10 user onto the same version of Windows 10.
The result is that all those compatibility, stability, and feature improvements made in versions 1803 and 1809 will be completely off-limits to the machines running 1709. This means that Edge, already a relatively small target for Web developers to think about, suffers major version fragmentation. Contrast this with Chrome, where within a few of a new version coming out, almost the entire user base is migrated.
This coupling with Windows 10 similarly means that there’s no facility to offer Edge to Windows 7 or even macOS users.
It’s this issue that’s perhaps the lethal, killer blow for Edge. If every Edge installation were always the latest and greatest one, and if corporations could standardize on Edge across their entire fleet—including Windows 7 systems, including Macs—then the browser might just have been a big enough, consistent enough target that it would be sustainable. But Redmond has decided that it isn’t. Microsoft did consider decoupling the browser from Windows 10 so that it could be updated on its own cadence and so that it could be ported to Windows 7 (if not macOS), but for reasons that are unclear, it decided that the work to do this would be too substantial.
Switching to Blink is seen as the easier way to break the Windows 10 dependence, and so not only will Edge’s release cycle be decoupled from Windows 10’s, Microsoft will also make the new Edge available on Windows 7 and macOS.
The old Edge rendering engine is probably not entirely going away; it will be used for embedded Web views in Windows apps.
New engine, new processes
The company wants to become a contributing member of the Chromium open source community. Early priorities include adding ARM64 compatibility to Chromium, enabling smooth scrolling with touch devices in Chromium, and improving the integration with screen readers to improve accessibility. Over the longer term, the company wants to bring unique Edge features, such as running risky content in an isolated virtual machine, to the new browser.
In tandem with this, the company is also having to develop new engineering processes. For its use of Chromium to be successful, Microsoft must be able to very quickly incorporate updates made to the Chromium code, build Edge, and then distribute it through Windows Update. Any delays in this process risks creating security gaps, wherein a flaw is fixed in Chrome/Chromium but still exposed in Edge. Quick integration is straightforward for a pure clone of Chromium, but the more Edge diverges from Chromium, the harder this task becomes. Designing this development process is still a work in progress. It’s not yet known if, for example, Microsoft is going to adopt Google’s multiple development channels.
And, of course, it’s not yet known how well Microsoft and the Chromium project will work together. Microsoft would like things, such as improvements to scrolling performance, to be merged into core Chromium so that the experience even of Chrome users on Windows is improved. But this isn’t guaranteed. Blink is itself a fork of the WebKit rendering engine, created because Google and Apple couldn’t reconcile their different priorities in the same codebase. Microsoft has significant reasons to want to avoid any major deviation, but it would be a shame if Edge users were forced to give up the things Edge does well just so Microsoft can stay close to Chromium.
The need to create that engineering and development process has another consequence: Microsoft isn’t expecting to have any builds of Edge with Blink for several months.