I'd have to agree. Our tiny startup is interested in adding some HTML5 structure using Modenizer etc but the main problem is the associated effort for not much return yet. It's really a balancing act for time - unfortunately IE7 and IE8 [and gasp IE6] are still hugely prevelant - so it's still 'the same old same old' by having to detect browsers and support multiple scenarios for HTML4 or HTML5. At this stage, until IE7 and IE8 die out "more" - which we certainly hope is in Q4 by end of this year - we will seriously look at integrating HTML5.

It's a real - "improve existing server-side code / optimisation and feature bundle" vs. "add HTML5 support where needed". For us, and this is only IMHO, the former is currently winning. By Q4 or even Q1-Q2 2012 - we'll look at this.

Thoughts?