The Steele quote was to a Lisp audience. They were a bit unhappy with the things. Lisp's performance was okay, but it did need a lot of memory in many cases (w

The Steele quote was to a Lisp audience. They were a bit unhappy with the things. Lisp's performance was okay, but it did need a lot of memory in many cases (with a few exceptions). Whenever you developed serious with Lisp in the 80s and early 90s, you had the machines with the most memory around. Using lots of memory then was a delivery problem, because not only the applications were expensive, but also the hardware which was needed to run them. Memory was extremely expensive in the 80s. Java had some influence from Lisp implementations, but not so much from the virtual machine architecture. Virtual machines in software were not the focus of the Lisp world, which had lots of native code compilers. Smalltalk usually used a VM approach. Generally Java did not get much from Lisp anyway. Half-way is not true, maybe 10%. Steele worked for SUN because he was a language spec guy (he wrote books and specs). There were widely used VMs without GC and there were lots of systems with GC, but no VM.