The Web browser you’re probably using to read this article is a small marvel of engineering. It can be broken down into many discrete parts—a URL bar, a set of bookmarks, perhaps a built-in video player. But the most important component is hidden: the “rendering engine” is the specific part of the program which ingests the hypertext-markup language, or HTML—the invisible stream of code downloaded when you arrive at a specific Internet address—and turns it into a visible Web page that can be drawn and displayed on the device’s screen. Without the rendering engine, nothing else would matter; rendering engines determine the shape of the Web as we perceive it.
On April 3rd, just a hair shy of four and a half years after Google unveiled its Web browser, Chrome—now the most popular desktop browser in the world by some counts—Google announced that future versions of it will switch from the popular rendering engine WebKit to a new custom engine, called Blink. (WebKit notably powers Apple’s Safari, along with most popular mobile browsers.) For now, Blink remains an almost-identical copy of WebKit (which is allowed because WebKit’s code is open source), but in the near future it will be refined by Google’s team into a new, lightweight engine that is fast, efficient, stable, and feature-rich. For Google, this will better facilitate the browser’s integration in alternative emerging contexts, like Android smartphones and its new Glass wearable computing device. Blink is expected to start powering Chrome by this June.
Blink is far from the first browser engine to emerge from the shell of another. Mozilla’s Firefox browser, the third most popular browser in the world, and its Gecko engine trace their origins to Netscape Navigator. Initially launched at the end of 1994, it was, for a time, the world’s most popular browser.
One late-summer evening in 1994, a founding engineer at Netscape, Lou Montulli, informally mused with colleagues about the tremendous gap between Lynx, the text-only browser he had created a few years prior, and the much more elaborate Web interfaces which could be displayed inside Netscape Navigator. He noted that the only remotely fancy thing Lynx could do with text was flash the characters on the screen. By the time Montulli woke up the next day—a Saturday—another developer had added the feature to Netscape, which could be triggered by writing as an element within the Web page’s source code. Montulli is now widely known as the inventor of the tag, though he is adamant he never advocated for ``, much less wrote any of the underlying code.
This new element was ostensibly a peer of the and tags used to render fonts in bold or italic face, but it was never actually standardized as an officially accepted feature of HTML. The flashing text it created was far more obnoxious—triggering epilepsy later became a concern, even—which is why it was swiftly and viciously maligned by designers. It did not receive universal support in browsers either; [Microsoft’s Internet Explorer](http://www.insidedhtml.com/tips/html/ts12/page1.asp) has never officially supported. Nonetheless, proved to be an easy and popular animation. During Netscape’s heyday in the mid-nineties, amateur coders and publishers flocked to enormous Web communities like Geocities and AngelFire, which formed proto-social networks that, unlike Facebook’s relatively sterile environment, were almost completely editable and customizable through raw HTML code. Since the rules and best practices for keeping the Internet clean and readable had not been established, they deployed liberally and recklessly. The results were often hideous, no matter which browser was used to view them.
Netscape was bought by AOL in 1998 as an attempt to control its own alternative to Microsoft’s Internet Explorer browser, which had started being tightly integrated with Windows, leading to an overwhelming market share (and a historic antitrust suit). As Navigator’s market share plummeted, AOL’s priorities shifted, and it abandoned the browser, eventually open-sourcing the core code, meaning it could be freely shared, cloned, and used for other projects.
In 2002, Internet Explorer’s market share peaked at a staggering ninety-five-per-cent share of the browser market. By then, its Trident rendering engine had grown extremely quirky—buggy, even. Phantom line breaks were sometimes inserted. Certain images didn’t display correctly. Page layouts broke. Designers just assumed the buggy behavior would be present and coded their sites accordingly; other browsers, even though they performed properly, were sidelined as a result. That same year, a few former Netscape employees had decided to repurpose Navigator’s open-source code for a new lightweight browser. It was called Phoenix at first, then Firebird, and, finally, Firefox. It reignited browser wars. And support for the `` tag survived intact in its code base.
Several years later, Internet Explorer’s monopolistic grip on the shape of the Web had slipped away. Apple made its WebKit browsing engine, which powers its Safari browser—most notably on the iPhone and iPad—open source in 2005. A number of browsers based on WebKit emerged, particularly on mobile phones, like Google’s Android. For the most part, these browsers do not support the `` element.
The rise of Web 2.0 in the mid-aughts—characterized by sites like Flickr and YouTube—and of Web apps, like Gmail, which behave and run like native computer applications, but do so inside of a Web browser, taught Google that it could mine data for customer profiles in powerful new ways. At the very end of 2008, perhaps feeling the same need for control that had led AOL to acquire Netscape nearly a decade earlier, Google announced that it would be releasing its own Web browser, giving it a reliable platform upon which it could always run its Web applications, no matter what Apple, Mozilla, and Microsoft might decide to do with their browsers. It was called Chrome, and it was based on WebKit.
Chrome’s recent move to Blink undercuts the primary olive branch it promised to Web developers upon Chrome’s release in 2008; those developers now need to test their Web sites in an additional rendering engine. But there is an argument in favor of the change: WebKit is now very widely used, especially in mobile devices, in much the same way that Internet Explorer 6 dominated the market and brought a near-halt to real innovation in the look and feel of the Web a decade ago. “Fundamentally, the belief is that having multiple rendering engines—just as there are multiple browsers—will spur innovation and help ensure the long-term health of the open Web,” said Alex Komoroske, a Google product manager.
Explaining Blink’s name, Komoroske said it “evokes feelings of speed and simplicity, which fits with our goals around speed and simplicity of architecture. We also have a tradition of slightly ironic names, and of course we’re aware of the infamous blink tag from the early days of the Web. But just like Chrome is all about minimizing browser chrome and the [Chromebook] Pixel is all about not seeing any pixels at all, Blink will never support the blink tag.”
The maker of Opera, a browser favored by a small contingent of extremely tech-savvy users (and people in areas with limited Internet access), recently announced that it also intends to switch rendering engines to Blink, in its flagship desktop browser. (It had previously announced it would switch to WebKit.) This is significant because Opera’s Presto engine allows the `` element.
Since won’t blink in Blink, Firefox would be the only remaining browser that allows text to actually flash using the element. In the messy world of Internet technology, where the browsers often can’t even agree about the size at which to draw a simple box, that is as clear a signal as one can reasonably hope for: perhaps it’s time to retire ``.
A few hours after Google unveiled its plans for Chrome and Blink, a manager on Mozilla Japan’s internationalization team, Masayuki Nakano, filed a new ticket in the company’s internal bug-reporting system to suggest that Firefox do just that. After a few rounds of discussion, Nakano altered the necessary code in about a day’s worth of work, and submitted his changes on April 14th. Starting with version No. 23 of Firefox, Gecko, Mozilla’s internal rendering engine, will no longer support the `` element.
Firefox intentionally removing support for the `` element draws a hard line demarcating the end of a wildly popular primitive Web animation which preceded today’s streaming videos and humorous GIFs. The change may be bittersweet for a certain weird variety of Internet nostalgist, but in most other senses it’s probably for the best: it’s a fantastically annoying bit of code and shouldn’t exist at all.
Even though no sensible designer has used the `` element in years, its complete disappearance is still disconcerting. Sites like the Internet Archive’s Wayback Machine store old versions of sites for historical purposes, in some cases dating back as far as the mid-nineties, but soon enough it may be impossible to view the sites as they actually appeared at the time. Geocities was finally shut down in 2009, but since it was such an important part of the early Web, the contents of many of those sites are still available via a massive six-hundred-and-forty-gigabyte archive posted on peer-to-peer file-sharing networks.
In fairness, the cultural value of the `` element is debatable. Brendan Eich, the creator of the JavaScript programming language and the current chief technology officer at Mozilla, does not think very highly of it: “It serves as a literal ‘attractive nuisance’ and a cautionary tale from an era when browser market share was unbalanced, and unwise innovations could become de-facto requirements without a standard process,” he said.
For the time being, slightly older versions of Firefox and Opera can still render all those archived sites with blinking text intact. But old browsers only run on old operating systems, and old operating systems can only be installed on old computers, and old computers eventually will be dead computers. Small pieces of information that make up the history of the Internet will be forever lost once the dwindling economy of scale makes it financially impractical for the manufacturers to continue producing a particular chip or component.
In any case, even though the trickle-down effects of Google’s recent browser development and broader strategy decisions will likely soon result in the loss or retroactive modification of our ability to properly recall and view artifacts from a key period of the Internet’s early history, Netscape’s aberrant HTML element was available for a little under twenty years. If you zoom out far enough on the timeline of all the culture and technology produced by our society, that’s practically the blink of an eye.
Vijith Assar is a writer and software developer in New York.
Correction: This piece originally identified Opera’s current rendering engine as WebKit; the company had announced plans to switch to WebKit, but it still uses its Presto engine.
Illustration by Maximilian Bode.
