Despite constantly being touted for some time as the future for mobile gaming, new research has shown that HTML5 remains some way behind Flash in terms of performance.
Software architect Sean Christmann used his benchmarking tool, GUIMark, to assess the current disparity in performance between HTML5 and Flash on nine different mobile devices.
In three tests – bitmap drawing, vector graphics and computation – Flash outperformed HTML5 every single time, often achieving double or triple the framerate. While Christmann openly admits that his work was funded, through his employer, by Adobe, the figures speak for themselves: HTML5 is still far from an ideal platform for mobile gaming.
At least, not yet. Much of that, Christmann tells us, is down to browser makers. "They haven't been able to keep up with the expectations of web developers that want to build applications on the web," he explains. "They have made it so the simple things are simple to do, but in many cases the complex things are impossible."
Giordano Contestabile, senior direct of mobile product and business strategy at PopCap, agrees: "The only platform where [complex applications] can run flawlessly is the latest implementation of Google Chrome, which isn't out of beta yet."
If the technology required isn't there yet, why all the hype? Earlier this week reports emerged that Facebook was at work on Project Spartan, an HTML5 platform that will see the social network's apps and games playable through the iOS Safari browser. "It stands to reason that companies that got their start on the open web would be looking to HTML5," says Contestabile, though Christmann notes that "Facebook is about to make a big splash with a native iPad app, which may signify that HTML5 features aren't [yet] good enough for what they want to do."
"What is required, quite simply, is more time. Another hardware and software cycle is probably required," Contestabile explains, adding that proclamations of support from Microsoft, Apple and the like "need to coalesce into a standard implementation across every device and browser to make the HTML5 experience seamless and widespread. I'd argue we're probably 12 to 18 months away from that."
So what should developers do to get their next game to as wide an audience as possible? Contestabile suggests that, for the time being, native apps are still the way to go. "Adoption is much higher, and there's a proven market with highly effective distribution channels," he says. "I'd also advise to make sure that the technology is in place to adapt those games to HTML5 when the time comes."
There are other options: Christmann suggests that HTML5 remains a viable option depending on the complexity of the game in question, and drag-and-drop game development engine GameSalad was this week updated to support HTML5. Part of that support allows developers with apps for sale on the iOS App Store to release playable HTML5 demos of their games through GameSalad Arcade, a welcome halfway house until devices and their web browsers get up to speed.