Web Development
Strategy & Industry

Windows 10 Ships Edge, and the Long Goodbye to IE Begins

Windows 10 arrived July 29 as a free upgrade to a huge installed base. Buried in its feature list is the ending we have covered toward for nine years. The default browser is no longer Internet Explorer.

Microsoft Edge, built on a new engine with evergreen self-updates, takes the default slot. IE 11 remains in the box strictly for compatibility duty. Its engine is frozen. Its future is officially maintenance. And the separately announced support policy puts a January 2016 end date on IE versions below 11.

Assemble the pieces, and the story that opened with our first browser audit in 2006 reaches its structural end. The IE7 betas. The hack inventories. The conditional-comment quarantines. Microsoft has joined the evergreen era. The frozen-old-IE problem now has end dates. And the browser-support tax that has inflated every web budget this agency has ever written is heading, at last, toward rounding error.

The operative words are ‘heading toward,’ because the goodbye will be long in exactly the sectors we serve. The practical advice is about managing the interval, not celebrating the announcement. Both halves follow.

What actually changes, and how fast

Edge itself warrants brief practitioner notes. It is a modern, standards-focused engine that renders well-built sites correctly. Our checks found the customary nothing, which remains the standards work paying off. It updates on its own, like its rivals. And its existence means every major browser is now evergreen by default.

The consequential math, though, is in the transition speeds, and they are two different curves.

Consumer machines will move quickly. The free-upgrade offer plus evergreen updates means the home audience converges on current browsers at a pace the IE6 era never dreamed of.

Institutional desktops will move at institutional speed, which is to say procurement speed. The hospitals, agencies, and campuses in this audience run change-managed shops where browsers are deployed, not adopted. Line-of-business applications certified against specific IE versions hold the schedule hostage. That is the intranet-dependency problem our enterprise coverage flagged as far back as the Acid2 era.

January’s end-of-support for old IE will move the laggards’ IT departments the way compliance duties do. But ‘moved to IE 11’ is the realistic 2016 destination for many, with Edge adoption following as Windows 10 rolls through refresh cycles. Translation for support planning: the tail is real. It sits in exactly our readership’s audiences. And the policy tool for managing it is the one we built in 2006.

The support-matrix policy, final edition

  • Read your own analytics before dropping anything. The tiered-support policy from our IE7 coverage, ninth year in service. Measure each browser’s actual share of your traffic and conversions. Set tiers accordingly: full support, functional support, graceful degradation. Revisit quarterly. The temptation this season is developers deleting old-IE accommodations on announcement day. The discipline is deleting them when your data, not Microsoft’s roadmap, says their users are gone.

  • Budget the deletions as real work. A decade of conditional comments, polyfills, and workaround CSS is scattered through mature sites. Removing it cuts weight and maintenance surface, and it is worthwhile scheduled work, not automatic. Our sites are logging cleanup tickets against the analytics thresholds now, which is how the tax actually shrinks. Deliberately, with records.

  • For organizations that run old IE internally: January is your compliance date. The other side of the matrix. If your own staff desktops run IE below 11, the end of security support makes upgrading a duty with an audit trail. The intranet apps holding you back just became formally riskier than the upgrade they were deferring. The vendor talks our enterprise notes have long recommended acquire a deadline.

What Edge’s arrival asks of the testing matrix, precisely

The operational paragraph for whoever owns your QA checklist.

Edge joins the matrix now as a first-tier target for consumer-facing properties. Its share will track Windows 10’s, which the free-upgrade offer makes brisk. Its evergreen rhythm means version-pinned testing gives way to current-version testing, per the practice the other browsers taught.

IE 11 remains first-tier for institutional audiences through at least 2016, per the procurement-speed math above. Its frozen engine means the accommodations it needs are stable rather than growing, which is the tolerable version of legacy support.

The below-11 tiers move to the graceful-degradation floor on January’s schedule, with the analytics check as the gate. Our client base’s IE 8-through-10 shares have thinned to fractions that make the floor honest. But sectors vary, really. One healthcare client’s patient portal still sees a stubborn IE 9 cohort from a specific partner, which is exactly the finding that justifies checking rather than assuming.

And the matrix’s write-up should note the simplification underway. For the first time since we opened, every browser in the first tier updates itself, which means the matrix’s future maintenance is subtraction. Few sentences in a QA standard get to say that. Enjoy writing it.

The era, closed with its lesson attached

Nine years deserve an accounting. The IE era taught we most of what it knows about the economics of standards.

A browser monopoly with no competitive pressure produces stagnation everyone else pays for. Competition disciplines the leader better than any petition. Firefox’s rise, WebKit’s spread, the V8 race. Sites built to spec inherit every convergence free, while sites built to bugs pay rent forever. And support policies grounded in measurement, not sentiment, keep the tax honest during the long tails.

Every one of those lessons transfers intact to whatever the next dominant platform is, which is why this goodbye is satisfied rather than nostalgic.

Edge ships. The old engines have end dates. The evergreen era is unanimous. And the correct amount of emotion about Internet Explorer’s long goodbye, after everything, is none. Measure the tail. Schedule the cleanup. Thank the standards bet one more time, and reallocate the tax savings to work that faces forward.

The web’s longest-running line item is finally sunsetting. We are glad we wrote it all down. A suggestion for the freed budget, since the question will come. The accessibility and performance programs that always competed with compatibility work for the same maintenance dollars. The tax’s payout should go to the users who paid it longest.