Eli Bendersky's website One of the classics of software engineering landed on my desktop (in an electronic-Russian-translated-version) the other day. I decided to read it not because I thought I can learn a lot from it (it's quite outdated, and I've already read a few more modern books on the subject), but to enjoy the book everybody talk about (most of those other books rely on this one for references this way or another).
Well, what can I say... it fulfilled my expectations exactly. I didn't learn a lot from it, but got impressed by the ideas it presents. Most impressive is that these ideas survived about 20 years of a boom in the software world, practically unchanged.
The book isn't too technical (if you need technical, read Code Complete), it's more managerial. Maybe it's one of the reasons I didn't *really* like it. I'm not a management type, after all. Sure, these things are important - measuring productivity, dividing PYs and teams for best performance, but I'm clearly more for the technical stuff.
I think that this book should be read for the historic perspective only... if you really want to learn these things, far better and newer books exist - Code Complete, Peopleware, Rapid Development, etc...
Addendum (2018-03-03): just re-read this book, 15 years later and with much more managerial experience by now. It's a good book, and I liked how it presents fundamental ideas about productivity in 1975, that are still true today more than 40 years later. Also, it's pretty awesome to read so many classical articles and even quotes that became part of the SW engineering jargon by today.