Re: The melting ice technology (2): levels

lgm@polaris.ih.att.com (Lawrence G. Mayka)
Sat, 14 May 1994 19:17:45 GMT

          From comp.compilers

Related articles
The melting ice technology (1): compilers & interpreters bertrand@eiffel.com (1994-05-09)
The melting ice technology (2): levels bertrand@eiffel.com (1994-05-09)
Re: The melting ice technology (2): levels peter@objy.com (1994-05-12)
Re: The melting ice technology (2): levels mhcoffin@tolstoy.uwaterloo.ca (1994-05-13)
Re: The melting ice technology (2): levels papresco@undergrad.math.uwaterloo.ca (1994-05-13)
Re: The melting ice technology (2): levels lgm@polaris.ih.att.com (1994-05-14)
Re: The melting ice technology (2): levels pdlogan@ccm.jf.intel.com (1994-05-16)
| List of all articles for this month |

Newsgroups: comp.lang.eiffel,comp.compilers
From: lgm@polaris.ih.att.com (Lawrence G. Mayka)
Keywords: interpreter, Eiffel, Lisp
Organization: AT&T Bell Laboratories, Naperville, Illinois, USA
References: 94-05-019 94-05-046
Date: Sat, 14 May 1994 19:17:45 GMT

mhcoffin@tolstoy.uwaterloo.ca (Michael Coffin) writes:


      [re Bertrand Meyer's articles on compilation and interpretation]


      I don't think there's any reason to be defensive about this, by the way.
      Many very useful languages---Scheme, Lisp, Elisp, Perl, Icon, and Awk, for
      example---are usually implemented using the same basic approach: an
      interpreter with compiled versions of the time-critical pieces to avoid
      most of the performance penalty. Although some of them allow compilation
      of code as well as interpretation (Icon and the Hobbit comiler for Scheme,
      for example), none of them I'm familiar with seem to integrate
      interpretation and compilation as nicely as Melting Ice.


Actually, most commercial and free-of-charge Common Lisp implementations
nowadays compile down to machine instructions, plus calls to the run-time
library as necessary. Most also include an interpreter, for the sake of
top-level interaction more than anything else; and of course compiled and
interpreted code can freely reference each other.


But this all only demonstrates that the principle of offering more
than one space/run-speed/compilation-speed tradeoff is a sound one.
--
                Lawrence G. Mayka
                AT&T Bell Laboratories
                lgm@ieain.att.com
--


Post a followup to this message

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