Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!cs.utexas.edu!usc!apple!motcsd!hpda!hpwala!hpavla!rowland
From: rowland@hpavla.HP.COM (Fred Rowland)
Newsgroups: comp.unix.questions
Subject: Re: style
Message-ID: <920004@hpavla.HP.COM>
Date: 2 Jan 90 12:53:07 GMT
References: <1524@utkcs2.cs.utk.edu>
Organization: Hewlett-Packard Avondale Division
Lines: 48

> I recently discovered a command on my workstation which I had previously
> been unaware of.  It it on a DECStation 3100 (Ultrix) and it is called
> "style".  What I would like to know is what are the supposed meanings of
> the various "readability grades"?  Is higher better?  What is the maximum
> possible grade?  What criteria are used?  Here is some sample output.
> Can someone explain what this means?

> readability grades:
        (Kincaid) 13.9  (auto) 15.0  (Coleman-Liau) 13.2  (Flesch) 14.7 (38.9)

> . . .

> David L. Battle


There's a complete writeup of 'style' and its companion programs 'diction'
and 'explain' in an article

     Writing Tools - The STYLE and DICTION Programs

     L. L. Cherry and W. Vesterman

dated November 22, 1980.  Unfortunately, my copy doesn't include the name
of the journal, but I'd say there's a good chance that it was Communications
of the ACM.  Good luck!

Briefly, the different readability scores are based on different formulae
derived by different people using different types of test reading material.

Kincaid is based on Navy training manuals ranging from 5.5 to 16.3 in reading
grade level.

ARI is based on text from grades 0 to 7.

Coleman-Liau is based on texts ranging from .4 to 16.3.

Flesch is based on grade school text, years 3 to 12.

This quote may sum up what you need "Coke (private communication) found
that the Kincaid Formula is probably the best predictor for technical
documents; both ARI and Flesch tend to overestimate the difficulty; 
Coleman-Liau tends to underestimate."

But try to find the full article; there's some good reading there.


Fred Rowland
Avondale Division/Hewlett-Packard