Dear AMICO University Testbed Participants,

I wanted to share the following correspondence with you, as it highlighted for me some issues around indexing the AMICO Library. We'd be very grateful for other observations regarding search results and terms, as we're now thinking about ways to improve access to the AMICO Library.

All comments and suggestions are most welcome and appreciated!

jennifer

EXTRACT OF QUESTIONS:

From: Miranda Marvin <mmarvin@WELLESLEY.EDU>
 To: pwalsh@WELLESLEY.EDU
Subject: AMICO test session
Date: Tuesday, February 02, 1999 5:21 PM

I spent some time testing AMICO and got some rather odd results. Here's my report:

Results of AMICO Session:

1. Simple Search for Red-figure vases yielded 5 vases: the first Black-figure, the other 4 red-figure.

2. Simple search for Black-figure vases yielded 7 vases: 4 of them Black-figure.

3. Simple search for Greek Vases yielded 11 items. 10 of them vases, one a bronze figurine.

Only some overlap with the above vases.

AND AN EXTRACT FROM MY REPLY:

The search results, are artefacts of the keyword indexes that support RLG simple searches..

On "red-figure vases", for example:

* the Etruscan figure comes up because it is "rather similar to the appearance of dancing figures painted on Greek vases,"

* The black-figure vase is painted by artists who "were active in the last decades of the 6th century BC when the new red-figure style came to dominate vase production in Athens,""

A search on "red-figure" in title produced 25 works, all of which are "red figure". BUT NOT all are called vases -- some called kylix or hydria... "Vase" isn't a word that we often find in museum records. Black figure in title produces similar specific results.

Hope this helps.

jt