W3C

Interview: Todd Anglin on the Kendo UI Developer Survey (February 2013)

The Kendo UI survey of 5000 developers released today provides strong evidence of HTML5 adoption. For example, when asked “How will you tackle the challenge of building apps for multiple mobile platforms?” 70% of the developers surveyed answered “adopt HTML5,” and 14% planned one native implementation per target platform plus one “catch-all” HTML app for all other platforms.

The survey results whetted my appetite for more information, so I contacted Todd Anglin, EVP Cross Platform Tools & Services for Telerik, who leads the Kendo UI team at Telerik and played an active role in crafting the recent survey. Some of my questions were not addressed directly in the survey, but Todd offered great insights.

Ian: The survey makes clear that for multi-platform development, HTML5 is the preferred option. But I am curious: how many projects are cross-platform in practice? What proportion of projects focus on one or two target devices?

Todd: We conclude from a variety of studies that multi-platform development is already important. For instance, many companies have embraced a “bring your own device” (BYOD) policy for their employees, and the result, on average, is that enterprise applications must support 3 to 4 devices. Gartner reported last year that the average CIO must already support 3.5 different mobile platforms, and most expect that number to grow. In our survey, we found that about 40% of developers spend time developing the same app or feature for multiple platforms, another clear indicator of active multi-platform development. For those that do this kind of development, 70% of respondents noted HTML5 as their first choice for managing the multi-platform complexity.

Ian: Suppose I want to reach 4 platforms. Which is more economical: 4 native apps or 1 Web app with the tweaks required to make it work well for users on the different platforms?

Todd: Anecdotally, we know that when you take a native approach the costs extend medium-term. For each platform, you need to hire (or train) people with those skills, and you need to maintain those skills for the lifetime of the app. I think the HTML5-plus-tweaks approach will have lower costs in that sense- one set of skills, and likely more shared code.

Ian: With HTML5 you’ll need to maintain the code and tweak it for new devices from time-to-time, but I assume that once you’ve solved an interoperability issue for a device in one app, you can reuse your solutions in future apps. That should reduce costs as well.

Todd: Yes, exactly! One of the indirect benefits of “standardizing” on standards, if you will, is that teams become masters in those skills. Rather than having developers that are “okay” in two or three (or more) different development platforms, you have developers that become deep experts in one. That facilitates faster problem solving, less context shifting, and, as you note, more ability to share lessons learned across development efforts.

Ian: It is my understanding that as more people are using the Open Web Platform, the demands on performance, capability, and interoperability are growing. The previous Kendo UI survey made clear that developers want broad interoperability. Do you find developers more or less satisfied with HTML5 compared to 1 year ago?

Todd: For reaching the growing diversity of platforms, HTML+JS is pretty much the only option today. People from so many industries want to solve the multi-platform problem, and HTML5 shows the greatest potential to tackle the challenge. So interest and investment are very high. We found that 80% of developers found HTML5 useful and 70% found it important. Even where they may be frustrated today, they are optimistic and think it’s just a matter of time for run-times to meet their needs, for tools to appear, and so on. The rapid refresh of mobile devices also helps get improvements to market faster.

Ian: I thought hybrid apps would have been more popular than “pure” HTML5 but the survey says they are neck and neck.

Todd: I was surprised, too. I think we need to dig more into this question. People we speak with about “pure” apps have said they don’t need hybrid, which suggests that what they are building can be delivered via the Web. It seems they prefer to deploy via the Web given the complexity of app deployment, particularly when public app stores are not an option. But what I found most interesting was that the response rates were the same for all types and size of organizations. I expected biases between, say, global enterprise and start-ups. But the answers were consistent across categories. No matter the organization size, survey respondents showed near identical preference for developing apps with HTML5 versus native.

Ian: I also found interesting that people are tuning one version of an app for native and also creating a “catch-all” Web app for all others. Why?

Todd: We were happy to see that strategy validated. Having the HTML5 application insulates you against rapid changes in the market. Think about it: six years ago the dominant mobile platforms were Palm, Blackberry, and Windows Mobile. Today the dominant platforms (iOS, Android) are entirely different. And, as I mentioned earlier, the market for devices is only going to grow more fragmented with a whole wave of new platforms hitting in 2013. Having a HTML5 app is the only way you’re going to scale and be ready to reach users on these new devices. Meanwhile, the validation of this “native + HTML5” app strategy also suggests that the decision is not black and white. You don’t have to be exclusively native or HTML5. A blended approach may be appropriate.

Ian: For the next survey, do you think you will ask about support for devices like television, ebooks, autos, and so on?

Todd: Good point. HTML5 covers such a broad space maybe 2013 is the year to branch out beyond mobile. We certainly believe HTML5 is the fastest way to reach the largest number of users on anything with a screen! And the latest survey seems to indicate that a large portion of developers agree.

Ian: Thank you, Todd!