I thoguht in
the end he managed to get the FET circuit to be better than
 the light bulb, but it took a lot of work (and probably wasn't worth it). 
    No.  His initial attempt with a FET didn't perform all that well (if
 memory serves), but in the end his very complex circuit with (I think)
 three high-end opamps and an LED-driven photocell did end up
 outperforming the incandescent lamp design. 
 
Isn't that basically what I said?
I didn't; realise he used an LED + LDR in the cirucit. It's a long time
sinc eI read the book, so I must have forgotten that. I'd sort-of assumed
it used only purely electronic (as opposed to optoelectornic) parts
 
That's a fascinating read; I'll try to find a pointer if you're
 interested. (I have the saga in a paper book, "Analog Circuit Design:
 Art, Science, and Personalites", one of my favorite books) 
 YEs, a great book. Along with Bob Pease's book ('Analog circuit
 troubleshooting' or soemthign like that). Well worth reading. 
 
    I shall have to obtain a copy of this book. 
 
I will try to dgi out my copy and post the exact title, ISBN, etc if you
like. I maye have misrememebered the title above, but _anything_ by Bob
Pease is IMHO worth reading (ditto Jim Williams) [1], and if you find a
book with those sorts of words in the title, it's probably the one.
[1] I don't necessarily agree with everythign they write, but it's all
worth reading. To be abvle to disagree with it you have to understand it,
and you'll learn a lot by so doing...
-tony