Lisp 1 Programmer's Manual (1960) [pdf]
history.siam.org
history.siam.org
12k is pretty impressive for a complete interpreter, including standard library. I don't really have anything useful to add to that, other than that I like it. It'd be nice if the java VM was 12k as well.
Also, I hadn't seen the Flexowriter before, that's a pretty awesome piece of hardware. It looks like it comes straight out of the Fallout universe.
|It'd be nice if the java VM was 12k as well.
Why would that be nice? You surely know how complex the JVM is, and how different it is from a basic, single-threaded LISP interpreter running on bare metal on proprietary IBM hardware.I do know how complex the JVM is, and I don't like it. Yes, clearly 12k won't get you the JVM. I do think there is a smaller, leaner runtime for java possible. Having everything under the sun in the standard library is not a feature to me.
As for the runtime size, the Truffle AOT compiler might help if it ever gets added to the reference JVM. Which is dependent on Graal replacing Hotspot I guess.
Another thing is the modularization, that will hopefully finally come with Java 9.
Having stuff in the standard library sounds like the definition of a feature -- well, actually multiple features.
Plus, it's not like those are loaded into memory when not used.
Lastly, there are several stripped down java vms, some are even used for real time / embedded work.
All those libraries buy me the freedom to do a bunch of different things. That freedom has some value, even if I never use it.
This is often cited as the first known usage of the word "object" in computer science.
"[...] The second phase of this was to finally understand LISP and then using this understanding to make much nicer and smaller and more powerful and more late bound understructures."
http://userpage.fu-berlin.de/~ram/pub/pub_jf47ht81Ht/doc_kay...