Scheme implementation techniques?

[mcvax!kth!draken!liuida!mikpe@uunet.uu.net (Mikael Pettersson)]

What are the approaches that have been used to implement interactive
Scheme systems? I am more-or-less familiar with the following ones:
	1) interpreting unprocessed S-expressions (parse trees)
	   (using "meta-circular" or flat coded interpreters)
	2) interpreting processed S-expressions (syntax trees)
	   (this is what my current interpreter does)
	3) bytecode compilation and interpretation
	4) compiling to native machine code
	   (is it worth while to always use compilation rather than
	   having an interpreter with a compile option?)

Are there any other approaches worth considering?
-- 
Mikael Pettersson, Dept of Comp & Info Sci, University of Linkoping, Sweden
email: mpe@ida.liu.se  or  ..!{mcvax,munnari,uunet}!enea!liuida!mpe