Thu, 21 Aug 2008 07:29:44 -0700 (PDT)

Related articles |
---|

Recursive Descent Parsers only with LL(1)-grammars? formatzeh@gmx.de (Gilbert Mirenque) (2008-08-20) |

Re: Recursive Descent Parsers only with LL(1)-grammars? parsersinc@earthlink.net (SLK Mail) (2008-08-20) |

Re: Recursive Descent Parsers only with LL(1)-grammars? torbenm@pc-003.diku.dk (2008-08-21) |

Re: Recursive Descent Parsers only with LL(1)-grammars? jaluber@gmail.com (Johannes) (2008-08-21) |

From: | Johannes <jaluber@gmail.com> |

Newsgroups: | comp.compilers |

Date: | Thu, 21 Aug 2008 07:29:44 -0700 (PDT) |

Organization: | Compilers Central |

References: | 08-08-041 |

Keywords: | parse, LL(1) |

Posted-Date: | 23 Aug 2008 14:43:57 EDT |

On Aug 20, 7:33 pm, Gilbert Mirenque <format...@gmx.de> wrote:

*> Hello NG,*

*>*

*> I'm not so familiar with compiler construction just interested in it.*

*> What appears curious is that I have heard that it is only possible to*

*> implement recursive descent parsers for LL(1)-grammars. But I imagine*

*> that it isn't a problem to look ahead more than just one symbol. So my*

*> question is why it is only possible for LL(1)-grammars?*

LL(1) is far too limiting the possibilities. Maybe the people confused

it with a fixed look-ahead, which LL(k) is traditionally known for.

But with ANTLR (http://www.antlr.org) now LL(*) is supported, which

allows infinite look-ahead. Basically LR and LL are pretty much equal

in recognition power (although people are going to correct me on

that ;).

Johannes

Post a followup to this message

Return to the
comp.compilers page.

Search the
comp.compilers archives again.