Re: RPN as intermediate code?

jaidi@technet.sg (Nor Jaidi)
Sat, 26 Aug 1995 03:18:07 GMT

          From comp.compilers
Related articles
RPN as intermediate code? gorelick@esther.la.asu.edu (1995-08-22)
Re: RPN as intermediate code? lkaplan@BIX.com (1995-08-25)
Re: RPN as intermediate code? jaidi@technet.sg (1995-08-26)
Re: RPN as intermediate code? odersky@ira.uka.de (1995-08-29)
Re: RPN as intermediate code? ukln@rzstud2.rz.uni-karlsruhe.de (1995-08-31)
Re: RPN as intermediate code? hbaker@netcom.com (1995-09-02)
Re: RPN as intermediate code? mejohnsn@netcom.com (1995-09-11)
Re: RPN as intermediate code? bevan@cs.man.ac.uk (1995-09-06)
Re: RPN as intermediate code? Robert.Corbett@Eng.Sun.COM (1995-09-12)
| List of all articles for this month |
Newsgroups: comp.compilers
From: jaidi@technet.sg (Nor Jaidi)
Keywords: interpreter, design
Organization: Compilers Central
References: 95-08-164
Date: Sat, 26 Aug 1995 03:18:07 GMT

>And why can't I find any literature about this sort of thing? All the
>compiler books I can find say 'Here is reverse polish. Its a good
>intermediate representation because its stack-like', but then they go off
>and do something completely different.


Try Threaded Interpreted Language (TIL) published by Byte. It is a very
old book (1970's?) discussing the implementation of a language that
looks like Forth. (May be it was Forth's forgotten predecessor?). The
compiled code (as well as the source code) is in RPN. If I remembered
correctly, almost everything, (+, -, if, while, etc) are compiled into
"procedure calls". But I gather performance is not high in your criteria
list. The good thing about TIL is that if you stripped away its editor and
translator, the run-time part that is left is very small.


Nor Jaidi
jaidi@technet.sg (Now)
jaidi@enterprise (Future)
--

Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.