Instead of theoretical I should have said "research-grade". People sometimes don't realize how big the gap between a research finding and a product that actuall

Instead of theoretical I should have said "research-grade". People sometimes don't realize how big the gap between a research finding and a product that actually delivers market benefits is. Often the discovery is only 5% of the work, and the rest isn't just grunt programming, but deep applied research with big technical challenges. That Graal has shown that the idea can actually yield competitive benefits in practice, running on real applications, and at acceptable costs (that are being continuously reduced) is a great achievement. Graal/Truffle is, obviously, not magic, and cannot convert any interpreter to a competitive optimizing compiler (it doesn't even beat C2 on some important workloads, which is why it is not replacing it just yet), but as a first approximation for what it does, I think that interpreter->optimizing compiler is a fair description, and even from a full-picture point of view, Truffle languages are competitive at a significantly reduced cost to other approaches we see in industry. Applied research that yields significant bottom-line benefits is precisely what a technological breakthrough is.