Re: User-oriented BNF "walker"?

Joachim Durchholz <>
20 Jan 2003 23:59:27 -0500

          From comp.compilers

Related articles
User-oriented BNF "walker"? (Chuck Lutz) (2003-01-17)
Re: User-oriented BNF "walker"? (Joachim Durchholz) (2003-01-20)
Re: User-oriented BNF "walker"? (Andreas Gieriet) (2003-01-21)
Re: User-oriented BNF "walker"? (Andreas Gieriet) (2003-01-21)
Re: User-oriented BNF "walker"? (2003-02-06)
Re: User-oriented BNF "walker"? (Rodney M. Bates) (2003-02-11)
| List of all articles for this month |

From: Joachim Durchholz <>
Newsgroups: comp.compilers
Date: 20 Jan 2003 23:59:27 -0500
Organization: Compilers Central
References: 03-01-089
Keywords: parse, tools
Posted-Date: 20 Jan 2003 23:59:27 EST

If I'm correctly informed, what failed were syntax-directed editors.

From my perspective, the failure was because a lot of programming
activity is restructuring code and moving code snippets across the
boundaries of constructs, and in a syntax-directed editor, you have
navigate into and out of constructs where a text editor lets you move
the code just a handful of lines up or down.

You want a different tool: something that lets you explore the BNF
specification of a language, play with it, see what results if you
substitute the definition of B into rule A -> x B y | x B z, etc. I
know of no such tool; the various compiler-compiler toolsets might
have one. If I were asked to write one, I wouldn't see any challenges
in compiler theory, but I'd see a challenge in getting the tool easy
and intuitive to use.

[I used some of the 1970s syntax directed editors which were indeed
awful. The only good one was Steve Wood's Z, which worked by pattern
matching the plain text in the file rather than by trying to edit a
parse tree directly. -John]

Post a followup to this message

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