Re: Opinions about "epsilon" Symbols in Parse Trees

Hans-Peter Diettrich <DrDiettrich@compuserve.de>
10 Jun 2005 22:16:32 -0400

          From comp.compilers

Related articles
Opinions about "epsilon" Symbols in Parse Trees drikosv@otenet.gr (Evangelos Drikos) (2005-06-09)
Re: Opinions about "epsilon" Symbols in Parse Trees schmitz@i3s.unice.fr (Sylvain Schmitz) (2005-06-10)
Re: Opinions about "epsilon" Symbols in Parse Trees DrDiettrich@compuserve.de (Hans-Peter Diettrich) (2005-06-10)
Re: Opinions about "epsilon" Symbols in Parse Trees mefrill@yandex.ru (mefrill) (2005-06-12)
Re: Opinions about "epsilon" Symbols in Parse Trees news4e71@yahoo.com (0x4e71) (2005-06-12)
Re: Opinions about "epsilon" Symbols in Parse Trees DrDiettrich@compuserve.de (Hans-Peter Diettrich) (2005-06-12)
Re: Opinions about "epsilon" Symbols in Parse Trees drikosv@otenet.gr (eDrikos) (2005-06-13)
| List of all articles for this month |

From: Hans-Peter Diettrich <DrDiettrich@compuserve.de>
Newsgroups: comp.compilers
Date: 10 Jun 2005 22:16:32 -0400
Organization: Compilers Central
References: 05-06-052
Keywords: parse

Evangelos Drikos wrote:


> The e-free Syntax Rule:
> "<Identifier> ::= <Latin Letter> [ { <Latin Letter> | <digit> } . ] "
> can be restated as:
>
> <Identifier> ::= <Identifier start> <Identifier part>
> <Identifier start> ::= <Latin Letter>
> <Identifier part> ::= { <Latin Letter> | <digit> | NONE } .
> /* where NONE is the "epsilon" */


Isn't NONE an implied alternative for the whole loop, in case of zero
occurences?


I suspect that you choose a bad grammar syntax for your purpose, pure
BNF might be a more appropriate choice.


DoDi


Post a followup to this message

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