Quantum Mechanics by Dirac I took quantum mechanics from our school's electrical engineering department. It was a grad level class designed for students working in device physics, thus it covered a lot of materials: from the basics (Schrodinger's equation, tunneling, the harmonic oscillator), to statistical physics (variational methods, Fermi-Dirac, Bose-Einstein, and Boltzmann distribution functions), as well as some solid state physics basics (simple models for metals, semiconductors).
I then went on to take solid state physics, which used Ashcroft&Mermin, and Lundstrom.
Now I no longer plan to work in device physics for my phD, but I still want to have a good understanding of QM and Solid state physics.
I was working through the Griffith text, hoping to graduate toward the Shankar text when I came across Dirac's book. It seemed really elegant and focuses on intuition first. I was wondering if anyone would recommend going through Dirac's text before going to Griffith's? It makes more sense to me but most curriculums never even touches Dirac's book.
Thanks,
Al
 A: Read Dirac's book. It is a complete introduction to quantum mechanics, done very elegantly and very physically, with the type of insight that only a founder can give. There is no substitute. Dirac was never used as a textbook, because it is too good, people don't assign good books in elementary classes.
The only things that you need to read in other places are the path integral, which is covered well in Yourgrau and Mandelstam and Polchinsky's string theory books, and ironically the Dirac equation, which you need to work out for yourself, because I don't know a great source.
A: Dirac is definitely worth reading, but you don't necessarily need to read it before you work through either/both of the others. You may even find it useful to read it (again) after working through the others. I would actually say that the three texts are linearly independent; they each have their own strengths and weaknesses. Dirac's explanation of the formalism is fantastic, and his solutions of the simple problems actually makes them comprehensible. The class I used it for stooped using it after the chapter on perturbation theory, so I don't know how good the later chapters are. Griffiths is typically used for physics undergrads who have some physics, but haven't necessarily gotten to all the math that they'll need. It's probably the best of the three for that. As mentioned already, Shankar is typically used for grad classes, but my undergrad class used it. You don't necessarily need to do Griffiths before you do Shankar. Shankar spends a lot of time going through the mathematical formalism, so if you're uncomfortable with any of that, you'll either get comfortable with it or decide to drop back to Griffiths. A really good check is to try to work your way through the problems in Chapter 1.
One big downside to Dirac is that it doesn't contain any problems. So if you want problems to work, you'll have to either pull them out of the other two, or find some on the internet. Dirac's book does discuss the Dirac equation (though see @Ron Maimon's comment below). He does not discuss Feynman path integrals. Griffiths discusses neither of those two. Shankar discusses both. I also just noticed that Griffiths has a section on solids, which includes Bloch's theorem and band structure. So you might find that a useful connection to the solid state stuff. I remember I was not comfortable with that my first time through solid state after Shankar; it wasn't until I had the same instructor explain perturbation theory and band structure to me that I got it.
