[FoRK] Arc, arc, arc.

[Jeff Bone]

This proves the point I was making about the relatively unimpressive  
amount of work that goes into implementing e.g. an Arc.  Less than a  
week out and somebody's gone and done an alternate implementation of  
most of it --- enough to bootstrap, i.e. run arc.arc --- in Javascript.

   http://halogen.note.amherst.edu/~jdtang/arclite/

Heck, what this guy has done is far *more* impressive in many ways,  
as the delta between the substrate and target languages is far  
greater than the distance between e.g. Arc and it's MzScheme host.   
In fact, this Javascript implementation strains the bounds of what  
you might consider metacircularity. He's also added some  
innovations:  first-class macros, a small but significant change to  
the binding behavior of set, and Unicode support. :-P ;-)

PG will correctly say that the heavy lifting inherent in Arc is  
figuring out what the language should look like and how it should  
behave, not the implementation.  And I agree with that to a large  
extent.  I suspect we just disagree on how much heavy lifting even  
that actually involves...

jb