I was inspired to write this post by the discussion at Coordination Problem regarding the use of Man, Economy, and State (MES) as a principles textbook. MES is Murray Rothbard's magnum opus on economic principles. The concern that one commentator had was the utility of MES as a preparatory text for higher-level economic study. I think the main concern is not the quality of economic insight one will get from Rothbard, but rather the methodology used by Rothbard.
The Austrian economists use verbal logic and a priori reasoning to develop theory. Mainstream economics uses symbolic logic and develops hypotheses that are empirically tested. Austrian economics weaves its theory into narrative history to explain economic events. Mainstream economics uses econometric modeling to see if the hypothesis fits the data. The skill sets for each are somewhat different, so I understand the concern.
As one trained in the mainstream who is transitioning to the Austrian view, I think I have some insight here. First, at the undergraduate level, the main tools and conclusions offered by Austrian economics and mainstream (neoclassical) economics won't be much different. Supply and demand curves are derived in different ways but this follows from different views on the nature of homo economicus. This, though, is an important point. Neoclassical economics assumes the economic agent is rational and greedy. Austrian economics assumes the economic agent is just an ordinary person with his or her own tastes and desires. Austrian economics is based on a much more realistic view of humanity than neoclassical economics. So while Austrians might talk about utility, they don't think that utility curves can be modeled as functions of wealth, for example - it's much more complex than any math function can capture.
So for the undergrad, a class using MES would be better, I think, than the typical mainstream class. Would is hamper the person who wanted to major in economics? After all, the upper level classes may not use Austrian economics texts. As I said, the main knowledge expected from an intro class would exist, and then some. The MES trained student would know more economics but have a different view of some key areas. Chiefly, the Austrian view of production is substantially different from the neoclassical view. But this is a good thing - the neoclassical view is easy to learn, but the Austrian trained student will be able to intellectually engage and criticize this story, rather than just accept it. This may be frustrating for some professors, but will be very fruitful for the student.
Well, that's micro, but what about macro? Well, the Rothbard-trained student would have a hard time with a Keynesian-style macro, but not as much of a hard time with a neoclassical-style macro, simply because the Rothbardian student would announce "bullshit!" so many times. In the Austrian view, all economics is micro, and anything termed "macro" is just a statistical description of some economic data. As Roger Garrison (I think) said, there are macro problems but only micro solutions. But, neoclassical macro is often modeled in micro terms, but here too is a problem since the neoclassical style makes heavy use of the "representative agent" technique and Austrians can't accept that all people are the same, or similar often to use one or two households to represent an entire population.
Now, what about graduate study? Well, I did Ph.D. level economics without any senior economics courses from my undergraduate or Master's degree. The thing about Ph.D. level econ (my book was Mas-Collel et al. if you want to have a look at what we did) is that it doesn't assume you already know economics. It does assume you're pretty good at math, however. I viewed my Ph.D. micro as applied mathematics, and I didn't find I really increased my economic intuition or really learned the 'economic way of thinking.' So MES won't prepare you for that - but I think it's a critical complement to it.
Thus, my conclusion is that studying Austrian economics as an only course in economics would give the students all the grounding they need to really understand the world and especially the effects of policy interventions. For an econ major, it is an invaluable complement to mainstream learning, but may require the student to do a little extra math on the side if the student wants to major in mathematical economics. And finally, for the graduate student in econ, MES won't help with your classes, but it will be an invaluable antidote to them.