[port-peer-review] Review La Barre-Dent article
Hi, (01)
Here's the review Mary requested me to do of Kathryn La Barre and Chris
Dent's article (some issues have already been addressed to a certain
extent): (02)
http://lab.bootstrap.org/port/papers/2002/labarredent/index.html (03)
I agree with Philippe that the description currently is too general, and
that terms like cluster, facet, access structure etc. need to be defined
and illustrated with clear examples. (04)
I very much like the focus of the authors on their tool aiming to augment,
not to replace human evaluators. However, my main problem with the paper
is that this key human evaluation process is left implicit and that the
demarcations between the formal and human processes are not too clear. My
main suggestion would therefore be to put more stress on the augmentation
(meta)model of which the knowledge organization process is part, maybe
by means of a figure. (05)
Some more detailed comments: (06)
- (011) "By enabling side-by-side comparison of various methods of
semantic analysis, the means by which each method clusters and the
potential utility of each method will emerge" -> How to envisage this?
What criteria do you use to compare utilities? (07)
- (012) "We built a testbed for evaluating *tools*" -> I am confused: I
thought your method was about organizing *knowledge*, and that you
developed a *tool* to support the human classification process? (08)
- (015) "extraction of recommended resources" -> What kind of resources
and extraction? From where and to what purpose? (09)
- (016) "We contemplate that this traversal function will allow
comparisons that will assist human evaluators attribute facets to the
messages" -> How is this traversal function actually used/integrated in
the attribution process? (010)
- (031) Do you have any experience with coding practices? What are the
incentives for people to code? What types of coding are there, apart from
coding for importance? What are the sources of coding terms? Controlled
lists? How do you resolve coding conflicts, if at all? (011)
- There is a difference between concepts and terms. The focus of the
automated part of your analysis seem to be terms. Have you thought of
extending this work by, for example, using conceptual ontologies and
linking these with term thesauri? What happens after humans have manually
labelled terms/concepts. Do the discovered relations then become available
for automated processing in the next round of analysis? (012)
- Your figures print badly, the node numbers are not readable. (013)
Cheers, (014)
Aldo (015)
==========================================================================
---/// e-mail: ademoor@kub.nl
IN|F/OLAB phone +31-13-4662914/3020, fax +31-13-4663069
|/// home page: http://infolab.kub.nl/people/ademoor (016)
Dr. Aldo de Moor
Infolab, Dept. of Information Management - Tilburg University
PO Box 90153, 5000 LE Tilburg, The Netherlands
========================================================================== (017)