This book is dry and uninspirational. Its exercises do their best to be 'writing' exercises, whereas they have to be exercises for 'hearing' .
Positive Points: Its expansions on Piston's first edition do some logical revisions, altering the ordering of chapters, and making the book more -systematic-; some of the exercises look a lot like the exercises in the initial chapters of Piston's 'Counterpoint', so that's also an indication that the latest DeVoto edition tries to be careful about the contrapuntal dimension of music that you simply cannot ignore if you're asking the student to write exercises other than block-chord progressions.
I'm pretty much a novice in 'Harmony' and 'Counterpoint', but I have to say that among all those books that I have examined, the best one by far was Sessions' book. ***You have to evaluate a harmony text by the types of exercises it has, because a harmony text is nothing other than a simple guide to have you discover certain things for yourself, preferably with the assistance and critical evaluation of an experienced tutor.*** I've tried to benefit from the first and fifth editions of Piston's "Harmony", Gauldin's "Harmonic Practice in Tonal Music", Rimsky-Korsakov's little known "Harmony", Schoenberg's Harmonielehre (A work of genius, which really should by evaluated on a different plane than what we're discussing at the moment.), Hindemith's "Traditional Harmony", and Sessions' "Harmonic Practice". Gauldin's book is a confused appempt to reconcile Schenkerian pedagogy with the general frame of mind that you see in Piston. The result is only un-inspirational, and 'dry'. What are we aiming to do in 'harmony' study? Are we not trying to acquaint ourselves with ears as unprejudiced as possible, with the materials and technicalities of a certain system, which has not just descended from the skies, but has evolved, -as a very efficient and easily comprehensible way of organizing the thought-in-tones? How can we do that? Not by memorizing formulae, or individual 'functions' for sure: You can't learn a language by memorizing sentence-forms. 'Detailed' as the Piston-DeVoto-Gauldin approach may be, it lacks the fundamental element that the student must feel himself free with his exercises, from beginning to end, so that he later has the ability to 'talk', and perhaps, much later, to express himself with/in that language.
Sessions incorporates the figured bass-melody harmonization-figured soprano scheme of Hindemith with the Schoenbergian idea of having the student listen and discover for himself the meaning of harmonic usages by concocting his own structures from the beginning. That's a very interesting synthesis (remember how Schoenberg hated melody-harmonization!), and I must add that he also takes the relevant aspects of Schenkerian theory. From the beginning, he makes you think in contrapuntal terms: You can't do that if your 'book' forces you to think in terms of roman numerals, and in terms of roman-numeral-restrictions that seem to belong not to a liberated understanding of tonality, but to a very curious 'style' within functional-harmony. That's why people have accused the Piston-DeVoto book of having nothing but note-drawing exercises: I concur with them!
BTW, I taught myself harmony from this book when I was in high school (with the guidance of a teacher reviewing my exercises) and tested out of all theory at a prestigious university. Basically I got the equivalent of an undergrad theory background from studying this text.
Also recommended: PENTATONIC SCALES FOR THE JAZZ-ROCK KEYBOARDIST by Jeff Burns