|Compiler Compiler Compiler email@example.com (Daniel C. Wang) (2001-03-22)|
|Re: Compiler Compiler Compiler firstname.lastname@example.org (J.H.Jongejan) (2001-03-26)|
|Re: Compiler Compiler Compiler email@example.com (Mike Dimmick) (2001-03-26)|
|Re: Compiler Compiler Compiler firstname.lastname@example.org (2001-03-26)|
|Re: compiler compiler compiler Dr_Feriozi@prodigy.net (2001-03-26)|
|Re: Compiler Compiler Compiler email@example.com (Kevin Szabo) (2001-03-27)|
|Re: Compiler Compiler Compiler Trevor.Jenkins@suneidesis.com (Trevor Jenkins) (2001-03-27)|
|Re: Compiler Compiler Compiler firstname.lastname@example.org (Chris F Clark) (2001-03-27)|
|Re: Compiler Compiler Compiler email@example.com (Ingo Dittmer) (2001-03-27)|
|Re: Compiler Compiler Compiler firstname.lastname@example.org (2001-03-27)|
|Re: Compiler Compiler Compiler email@example.com (2001-03-31)|
|[7 later articles]|
|Date:||26 Mar 2001 13:51:13 -0500|
|Posted-Date:||26 Mar 2001 13:51:13 EST|
>Just curious, but has there been any research done on
>parser-generators generators? i.e. some magic tool that will take a
>specification of my favorite programming language and produce a yacc
>like parser generator.
>The first non-trivial thing anyone does when they invent a new
>programming language seems to be rewriting yacc so they can bootstrap
Or maybe a compiler compiler compiler compiler that first invents the
new programming language?
But seriously, C++ is the only programming language that I can think of
that seemed to need a non-traditional parsing method. The relative
proliferation (actually there are very few) of new tools falls mainly
into two categories.
1. Improvements to, or ports of the venerable YACC. (bottom up)
2. LL(k) parser generators. (top down)
The main motivation for number two is summarized on page 273 of Aho and
Ullman's "The Theory of Parsing, Translation, and Compiling."
"When we look at translation, however, the left parse appears more
The gist is that top down syntax directed translation would be
preferable if only it were not so hard to write the grammar for the
language. This mainly applied, I think, to LL(1) grammars. Left
factoring tends to mutilate many LL(1) grammars beyond recognition.
LL(k) greatly reduces the need for left factoring, so it is much easier
to produce a usable grammar.
The SLK parser generator home is http://pages.prodigy.net/dr_feriozi
Return to the
Search the comp.compilers archives again.