I came across this article via my network on linked-in and couldn’t agree more. There was a panel at Cable-Tec expo at Orlando and these remarks were made by Albert Lai, Innovation architect at Brightcove. Everybody is tremendously excited about HTML5 today but there’s tremendous amount of fragmentation today with respect to browsers, platforms features and what not. I often spend lot of time educating customers about these shortcomings and we plan on writing a white paper in the near future.
Take for instance WebGL. Now common perception is that as long you have a browser that supports WebGL, it’ll automatically work on any platform and you’ll get excellent performance. Wrong. (I am talking about embedded systems here - not X86 desktop). The browser will need WebGL implementation adapted for the OpenGL (or other) hardware implementation available on the platform. Till that’s done, applications are running in software and the performance will tank.
Then there are limitations on video playback. Take YouTube for example - one of the limitations that is listed is ‘Pressing Fullscreen button will expand playback to fill browser window’. So you’ll still see top and bottom bar even though you are playing video full screen. Also notice that the ad insertion is not supported today in HTML5 - technology used by Google to monetize YouTube.
DRM is the other area where HTML5 is lacking today. Granted, there are proposals, but it will take some time for them to get approval and reach maturity.
Browser specific features is another murky area. What’s up with -webkit style switches? What happens to non-webkit browsers?
In essence, HTML5 is far away from the problems that Flash tried to solve - ‘write once publish everywhere’. It is perhaps the best approach to solving this problem but there are significant roadblocks on the way (Remember Mark Zukerberg said writing HTML5 app was a mistake?) Having an open standards body is great but could also slow down things as everyone is looking after their own interests and agreeing on something may take longer.