Re: meta circular clojure

2014-02-13 Thread Di Xu
2014-02-13 1:05 GMT+08:00 Herwig Hochleitner : > 2014-02-12 5:36 GMT+01:00 Di Xu : > >> >> all lisp dialect provide `read` function, so if you want to build an >> evaluator, you could just use this function and don't need to do lexical >> and syntax analysis. >> > > Maybe your understanding of th

Re: meta circular clojure

2014-02-12 Thread Herwig Hochleitner
2014-02-12 5:36 GMT+01:00 Di Xu : > > all lisp dialect provide `read` function, so if you want to build an > evaluator, you could just use this function and don't need to do lexical > and syntax analysis. > Maybe your understanding of these terms is different from mine, in my view: A lisp evalua

Re: meta circular clojure

2014-02-12 Thread Herwig Hochleitner
2014-02-12 5:18 GMT+01:00 t x : > > If no such evaluator exists, where is the complexity of a > clojure-in-clojure evaluator that I failed to mention above? > Clojure is a compiled language. This means that even if you leave out any platform issues like bytecode generation, there is still an an

Re: meta circular clojure

2014-02-11 Thread Di Xu
Hi, lisp in lisp evaluator is the simplest evaluator you can build, because lisp uses list to represent its program, and all lisp dialect provide `read` function, so if you want to build an evaluator, you could just use this function and don't need to do lexical and syntax analysis. This is the po

meta circular clojure

2014-02-11 Thread t x
Hi, With apologies for a soft question: * in SICP, the meta circular evaluator of scheme in scheme (chapter 4) is about a page * in Ansi Common Lisp, there's a lisp in lisp in about half a page * if we took Clojure, and ripped out all the JVM support (i.e. gen-class, proxy, (.javaFunc ..