Perils of Popularization
I read Feynman's Rainbow, and liked it, especially for it's evocation of the ambience of Caltech in the Feynman and Gell-Mann era, but I haven't read Euclid's Window. Robert Langlands has though and you might say his opening sentence telegraphs his opinion:
This is a shallow book on deep matters, about which the author knows next to nothing.
The review in the AMS Notices is long, erudite, passionate and boundlessly hostile. He clearly thinks Mlodinow gets almost everything wrong, but what really angers him is that (he claims):
...[the book] is certainly thoroughly dishonest, but not to any purpose, rather simply because the author shrinks from nothing in his desperation to be “readable and entertaining”.
He makes a gesture at even-handedness:
The book is wretched; there is no group of readers, young or old, lay or professional, to whom I would care to recommend it. Nonetheless, there are several encomiums on the dust-jacket: from Edward Witten, the dean of string theorists, and from a number of authors of what appear to be popularizations of mathematics. They are all of the contrary opinion; they find that it is “written with grace and charm”, “readable and entertaining”, and so on. Perhaps the book is a hoax, written to expose the vanity of physicists, the fatuity of vulgarizers, the illiteracy of publishers, and the pedantry of at least one priggish mathematician.
The review itself is a fascinating and fact-filled read for someone interested in the history of mathematics. It is almost the outline of the more serious and scholarly book on the "deep subject" that he wishes someone better had written - but says he is unqualified to write.
His viewpoint is the Olympian one that I suppose one should expect of a long-time inhabitant of what Einstein called a "quaint ceremonious village of tiny demigods on stilts." (The Institute for Advanced Study). I doubt that popularizations of mathematics or physics can exist in that thin air.
The link is via Peter Woit at NEW. Peter is mainly concerned with the deep matters of Geometric Langlands, another one of the remarkable mathematical connections of String Theory.