Facebook Abandoned HTML5 -- Should You?
We've written here before about the battle between the browser and the native app and we understand developers have very strong opinions about this. That's why it was so interesting that Mark Zuckerber speaking in an interview at TechCrunch Disrupt yesterday admitted that going with HTML5 and open web standards was a huge mistake.
Zuckerberg believes his company lost valuable time while focusing on an HTML5 development path. "It took us 6-8 months to build FaceWeb [the mobile HTML5 version of Facebook], another 4 months or so to decide it wasn't going to do it after we kind of committed to that, and then we had to start over over and start rewriting everything to be native," he said.
He added, "We burnt two years and it's really painful, and I think we will probably look back and say that it was one of the biggest mistakes, if not the biggest strategic mistake we could have made."
That's quite a revealing comment. Zuckerberg and his company had committed to an open approach using HTML5, then found it simply couldn't keep up with the the kind of demand that Facebook users had. Anyone who has used the iOS app before the update can tell you that it was painful to use. It's better now, but still too slow in my view (maybe I'm just impatient).
Glenn Gruber, avp of travel technologies and mobile solutions at Ness thinks Facebook's might have taken this strategic approach with HTML5 because of a desire to pursue a purity of vision about about write once, deploy everywhere, which HTML5 represented. But when it came to executing, it turned out HTML5 couldn't handle Facebook's tremendous performance requirements.
Gruber believes that for Facebook, HTML5 was great in theory, but turned out to be not ready for its needs in practice. "What I think that they missed is that while performance of HTML5 might not be too bad for some apps, the cumulative amount of time that people were spending on Facebook made the performance challenges very visible and then downright unacceptable," he said.
And that could be the strategic mistake Zuckerberg was referring to. I'm still left wondering why Facebook didn't simply pursue a parallel development path, but if as Gruber suggested they had bought into the idea of write once, use many times; once they committed to that approach, it took time to realize it wouldn't work.
It's a valuable lesson for all developers. When you commit to a given approach, you have to live with the results and sometimes there are unintended consequences. Most companies don't have the luxury of Facebook to do an about-face without suffering financially.
Purity of vision is great in theory unless it comes back to bite you in the butt in the end. That doesn't mean you shouldn't pursue HTML5 because not everyone has Facebook's performance requirements, but it's something you need to consider before you decide how to proceed.
Photo by Andrew Feinberg on Flickr. Used under the Creative Commons License.
blog comments powered by