@fluidlogic In their defence, though, old and supposedly reputable companies sometimes also miss. I recall reading a pretty weird book, full of dubious claims on parsing techniques, from Wiles, around the late Bush crisis.

No Starch Press seems to be one of the few tech publishers that consistently puts out high-quality books. O'Reilly and Addison-Wesley get pretty close, too; I'd even accept an argument that they're essentially on the same tier, they just do more books, so their occasional questionable editorial choices are statistically indistinguishible from No Starch Press getting nearly everything right.

One needs to be vigilant with Apress; they seem to believe in a process more than expertise proper, and probably just can't really tell a good tech book from a bad one. And Packt, well, is the blind pig that might sometimes find a tasty acorn, but I'd never bet on that. Perhaps not exactly a vanity press, but awfully close.