|[4 earlier articles]|
|Re: Recursive Descent Parser firstname.lastname@example.org (Randall Hyde) (2000-02-12)|
|Re: Recursive Descent Parser email@example.com (Michael Larson) (2000-02-22)|
|Re: Recursive Descent Parser firstname.lastname@example.org (2000-02-22)|
|Re: Recursive Descent Parser email@example.com (Chris F Clark) (2000-02-22)|
|Re: Recursive Descent Parser firstname.lastname@example.org (Fred J. Scipione) (2000-02-23)|
|Re: Recursive Descent Parser email@example.com (2000-02-27)|
|Re: Recursive Descent Parser firstname.lastname@example.org (2000-02-27)|
|From:||email@example.com (Anton Ertl)|
|Date:||27 Feb 2000 02:38:33 -0500|
|Organization:||Institut fuer Computersprachen, Technische Universitaet Wien|
|References:||00-01-027 00-01-032 00-02-034 00-02-058 00-02-111 00-02-115|
firstname.lastname@example.org (Paul Groves) writes:
>And I agree with Stroustrup (which is nice :-) ), if you're writing a
>langauge as an exeriment - hand writing the parser is the only way to
The way I understand what Stroustroup wrote, is: with his constraints
(C compatibility etc.), using yacc was a bad choice.
If you are writing a language as an experiment, and are willing to fit
the syntax to the tools you use, using a parser generator is a good
idea; in this setting it's even more useful than with a given syntax.
Why? With a hand-written parser, every change in the language can
result in changes affecting significant parts of the parser (e.g.,
changes in the first-sets). The parser generator takes care of these
issues for you.
M. Anton Ertl Some things have to be seen to be believed
email@example.com Most things have to be believed to be seen
Return to the
Search the comp.compilers archives again.