Three years browsing without Flash
It was three years ago today that I decided to stop using a Flash player. Since then, I have not installed (or have immediately uninstalled) Flash on all computers I use for more than one hour per year. I define it this way because I want to clarify that I make sure Flash is not on any computer I use regularly (including the computers I use at work owned by my employers), which is more than the computers I own. But I think it would be a bit much to force a friend to uninstall Flash if I’m only using my friend’s computer for a couple minutes to check email.
As a result, browsing for the past three years has been a very pleasant experience. My browser rarely crashes, it doesn’t consume all my computer’s resources for long periods at a time, and I don’t have to worry about whether I’m vulnerable to any of the 166 security flaws discovered in Flash over the past few years.
Why it doesn’t matter (anymore)
Back in 2008, one had to make a conscious decision not to use Flash. Most machines that shipped with Windows also shipped with a Flash player, as did most Apple computers (even System76 preloaded Adobe Flash Player on new Ubuntu machines). Mobile browsing was still in its infancy, as the iPhone had barely been out for a year and Android less than a month.
These days, many people will never use a Flash player because their computer or mobile device doesn’t ship with it and might not even allow it to be installed. Though most machines shipping with Windows still include a Flash player, Apple decided to stop shipping Flash on new Macs over a year ago. Mobile browsing has grown considerably since 2008 and now accounts for 5.5% of all browsing (if you believe the stats). And virtually none of these browsing devices come with Flash: iOS devices disallow it entirely while Android devices allow it but very rarely have it preinstalled.
Based on these trends, it’s not hard to see the number of users with Flash installed falling below 90% soon (if it hasn’t already). This is an important threshold, as web designers are much less likely to assume that enough users will have particular software installed or be likely to install it to view their web site if less than 90% have it already. So fewer and fewer web sites are depending on Flash, which is great for the open web.
The past few years have seen most of my predictions in “Why Flash is doomed” come true. The user base for Flash is indeed dwindling as I subscribed above. Flash is no longer needed for video in very practical ways now. Over 6 months ago, over 99% of views on YouTube were available in WebM without Flash. Only a modern web browser that isn’t made by Microsoft or Apple is required. And though Adobe has managed to build a Flash player for mobile devices whose manufacturers let them, people rarely factor Flash into their purchasing decisions (despite a valiant effort by RIM to make people pay attention to their PlayBook Flash ad), since virtually all of their browsing can be done without it and they care more about the apps anyway (more on that later). Even if they do get a device that supports Flash, it’s unlikely to do so out of the box.
Adobe’s Flash Player continues to be slow-moving due to its proprietary nature, even if I didn’t get the precise ways in which this would affect it three years ago. I recommended to Adobe that they adopt royalty-free codecs in Flash Player 2.5 years ago, but even though they promised VP8 support a year after, there is still no word on when we might be seeing that in Flash Player, nearly 18 months later. The Flash format remains a long way from even qualifying as a Draft Standard according to RFC 2026, since Adobe has still not bothered to create or even promote a second compatible implementation. In fact, Adobe actively seeks to destroy implementations that aim to be compatible with the Flash format. Since few people can fix problems in Adobe’s proprietary Flash Player, it continues to crash a lot, so much so that browser makers built special sandboxes for Flash that prevent it from taking down the otherwise-stable browser.
To sum up, Flash is becoming less and less relevant as fewer and fewer people have a Flash player installed and fewer sites use Flash for anything important. Whether I as an individual continue to browse without Flash is increasingly less significant as more and more people browse without Flash, oblivious to the fact that might be missing something, since what they’re “missing” doesn’t matter anymore.
Beyond Flash: new challenges for the web
While Flash’s diminishing importance is good for the web, there are still hurdles to overcome in bringing us closer to a truly open web. Many designers still view the web in a device-centric way, making sure their sites work on particular devices, but paying no attention to how they might look in slightly less popular environments. For example, some web sites present video in a standards-compliant way when they detect someone is using iOS (which doesn’t have Flash), but fail to present the video in that way when someone with a compatible non-iOS browser visits the site, assuming instead that they have Flash installed.
The rise of mobile apps has demonstrated that programs custom-made for particular OSes are (in the short run) better for some tasks than web apps. Standards bodies have adopted important features from OS APIs, such as geolocation and local storage, integrating them into new web standards so more apps can be made for the web. But OS APIs will remain the forerunners in providing device functionality to mobile software developers so it is important that these APIs and the software landscape they encourage remain as open as possible.
Given Apple’s role as an instigator in many of today’s mobile patent fights, it is fairly clear that Apple does not want anyone creating an iOS-compatible operating system (or non-Apple devices that run iOS). So apps made for iOS will continue to run only on Apple devices. Android is in a slightly better situation because anyone is free to build and install Android on their device, but Google still insists that vendors wanting to use Android Market get approval, which makes it harder to make an Android device. Furthermore, most software developers choose to make their apps available only in the Android Market and not also via a separate .apk download, keeping the non-approved Android devices at a serious disadvantage, as I witnessed myself when using Android on my FreeRunner.
While mobile app repositories are somewhat tangential to the web, many people are using mobile apps instead of a browser for the majority of their online activity so it is important that these be as open as possible while the standards are being hammered out. Web-based app repositories, like the Chrome Web Store may move more software developers onto the web, but it’s still unclear how successful these stores will be.
Though YouTube’s adoption of WebM is promising indeed, it will still be a while before all video only is available not only without Flash, but also without patented codecs, like H.264. More efficient WebM decoders, especially for mobile devices, are being worked on, but manufacturers continue to prefer H.264 due to its larger installed base of efficient decoders. Additionally, few devices produce WebM video by default. Most higher-end digital cameras and camcorders still use H.264, for example.
As Flash disappears into irrelevance, we should be constantly wary of these and other new threats to the open web. A world where we can view any information from any web-connected device is not yet within reach, but with more positive developments like the new web standards and royalty-free codecs we have seen recently, it might not be as far off as we think.