Path: utzoo!utgpu!water!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!rutgers!uwvax!oddjob!uxc!uxc.cso.uiuc.edu!a.cs.uiuc.edu!m.cs.uiuc.edu!liberte
From: liberte@m.cs.uiuc.edu
Newsgroups: comp.software-eng
Subject: Re: Language-based editors
Message-ID: <39400001@m.cs.uiuc.edu>
Date: 29 Jul 88 22:52:00 GMT
References: <2135@mhres.mh.nl>
Lines: 54
Nf-ID: #R:mhres.mh.nl:2135:m.cs.uiuc.edu:39400001:000:3059
Nf-From: m.cs.uiuc.edu!liberte    Jul 29 17:52:00 1988



Here is a repost of our earlier announcement.

------

We have released Leif, a language oriented editor that is based on 
a recognition paradigm.  In a recognition paradigm the programmer enters 
text and the editor incrementally analyzes this text to determine structure.  
Many other language oriented editors are based on templates.  In these 
editors, the program is constructed by expanding templates.  Template 
editors prevent the creation of syntax errors by restricting the possible 
edits.  In Leif, the programmer may use all of the text editing commands 
of a host editor (GNU Emacs) to create a program.  The programmer may also 
use language oriented commands.  At any point, the syntax analysis may be 
consulted to locate syntax errors or display the structure of the program.

By using a standard text editor as the front end, Leif offers several benefits. 
First, there is very little overhead in learning to use Leif.  By learning 
only one new command, the "move to syntax error" command, Leif can be useful.  
Other commands allow a programmer to see the structure of the program or 
examine a syntax error.  A second benefit is that the user can edit the 
program as text or as syntax simultaneously.

Leif has respectable performance.  It uses an incremental parsing algorithm
to analyze changes to the text and to maintain a parse tree.  If a change is 
made that affects only a local part of the parse tree, only that part is
Analysis does not occur on every key stroke.  Instead, Leif allows any 
number of consecutive changes to be processed as a unit. The programmer may 
use Leif transparently and request information from the parser when it is 
needed.  Once started, Leif does not require any actions by the programmer 
for it to follow an editing session.  Until a Leif command is activated, 
the user does not notice any change in the editor.  This translates into an 
efficiency benefit and further enhances the ease of use.

A language for Leif is specified as a lexical analyzer and LALR(1) grammar.
Standard tools, Lex and Bison (similar to Yacc), are used to process the
specification.  Once the tables produced by these tools have been converted
into an appropriate parse table format, Leif use the tables and edit text 
in the language.  Prototyping the syntax of a language is easy and a 
designer of a language may specify a language and experiment with it quickly.  
The language need not be a programming language, but could instead be a 
specification, design or text processing language as long as a language 
fits in the lexical analyzer and parser paradigm.   

Leif is available via anonymous ftp from a.cs.uiuc.edu.  The files are 
located in the subdirectory pub/Leif.  Consult the file README.DIST that
is in that directory to decide which files to retrieve.  Documentation
is available in the file leif-doc.tar.Z.  (If you are unable to retrieve
the files via ftp, please contact us.)

Please send Leif related queries to ihnp4!uiucdcs!leif or leif@a.cs.uiuc.edu

Bill Smith,	Dan LaLiberte,	Roy Campbell