I buried the lede in my last piece, at the end of which I suggested a plan for Firefox.  

Briefly: Fork a stable release, if necessary, farm this out to a tech partner who is good at managing large developer communities and enterprise users. This version only changes in two areas: improvements to the rendering engine and security updates. It's a lean platform for plug-ins. 

You want rendering to evolve quickly so as not hold back developers of websites, to avoid disasters like IE6. But if stability is important, you shouldn't have to accept insecurity. The pace of evolution and safety should not be related. 

The other fork is a the "rock and roll" thread, where changes are frequent, and innovations many.