Sometimes 'ahead of its time' is not such a good thing. Specifically, when the available hardware can't keep up with what the system design demands[1], e.g. a responsive smartphone touch interface created in an interpreter (as far as I can see without even having a JIT compiler) on a 500Mhz embedded CPU, without GPU acceleration[2]. It was ahead of its time alright.
Sure, running too taxing software on unaccelerated hardware is a bad thing. But webOS wasn't slow. It had sporadic slowdown and then it really was slow, but in normal state it sure could compete speedwise.
It was buggy, the scrolling felt slow because the settings were wrong (as proven by the buttah-patch), same for the animations, the default-browser lacks html5-support, the connectivity of the phones is bad, pages often don't load without a reason even when on wlan (and when not, it often fails to fetch a slowly loading page bit for bit, which was made even worse because the browser would reload that page the moment it was loaded because he was inactive too long(!)) - it wasn't performance that was the issue here.
So maybe LG really can take the system and make something out of it. The tile-system sure could work on TVs designed to be used with a remote-touchpad-controller.
Mercer is a bit... biased. He was pushing a debased Java system that took five lines of code to center text, and couldn't implement Duarte's ideas properly. So webOS becoming, well, webOS was in reaction to that and a rushed effort to come up with something else in time for CES 2009.
The Palm folks then spent too many months de-Mercerizing the code for webOS 2.0...
[1]http://hardware.slashdot.org/story/12/01/02/0213204/insiders...
[2]http://forums.webosnation.com/webos-discussion-lounge/295123...