RE: Reading level and metadata

Daniel and others,

CAST has done a bit of work in this area and remains unhappy with the
readability measures that currently use fixed lists, special vocabulary
matches, sentence length, and other wildly differing methods to determine
readability.  The best methods are expensive and time consuming and are not
based on simple algorithms.  We will settle on one (or more) shortly because
we feel that those who use our text-to-speech eReaders should be provided with
clues about how easy content might be to read prior to using tts support.  It
will provide an estimate at best.

It is a very imperfect science but educational publishers have to do it for
textbooks using controlled vocabulary, sentence length, paragraph length,
concept difficulty, etc.

All of this is made more complicated by the mix of readability and intended
audience.  We don't really want adults reading material which was originally
written for 4th graders simply because that is the appropriate reading level
for their skills.  This might be ok if they are still struggling with learning
to read but it is not ok if they are trying to figure out why there was a War
Between the States.

By the way, I am convinced that metadata wrappers should be used for this
purpose.  Our commercial affiliate, Universal Learning Technology (ULT),
created a tool that creates IMS compliant metadata for text and media objects
stored in databases so that they can be used to create web pages on the fly or
re-purposed by professors and/or publishers for course creation.  It is now
part of a product called Bravo but will soon be joined with ULT's newly
acquired WebCT to provide this capability to web course authors.

Chuck

************
Chuck Hitchcock, Director
Universal Design Lab (UDL)and
Product Development,
CAST, Inc.,
39 Cross Street, Peabody, MA 01960
Voice 978 531-8555
TTY 978 531-3110
Fax 978 531-0192
<http://cast.org/>
<http://cast.org/bobby/>

Received on Tuesday, 22 June 1999 16:24:08 UTC