Physical architecture and information architecture NOT

					University libraries and WWW image search
Paul.Nieuwenhuysen @ vub.ac.be               B-1050 Brussel          BELGIUM


 Introduction:                              Problem statement:                        Methods:
 Academic information can be found          How useful and efficient is WWW           The test subject domain is one in
    and accessed increasingly                  image searching nowadays to               which images are important:
    through the WWW, while                     discover information sources,             classical, ethnic African art.
    information in physical                    besides other discovery services       Queries were submitted to public
    documents and library services             offered by academic libraries?            access WWW image search
    remain important.                          More concretely:                          engines and for each query the
 One of the functions of academic           1. How high is the precision of              20 individual results that were
    librarians has become                      results from search systems with          ranked highest have been
    evaluating, selecting, offering            reasonable queries?                       evaluated quantitatively.
    and recommending information            2. Is the danger of misinformation        For each search, the scoring method
    discovery services on the WWW              an issue (beyond the presence of          leads to a positive score related
    to their clients.                          neutral irrelevant search results         to precision between 0 and 20
 Searching for images on the WWW               that lower the precision)?                and to a negative score related
    has become an attractive                3. Does WWW image searching                  to misinformation between 0 and
    starting point to find relevant            reveal specific html documents            minus 20.
    information sources, in                    with images, which have been           Data were collected, analyzed and
    particular in fields where visual          placed on an academic library             plotted, using a spreadsheet
    information is relatively                  WWW site by the                           program on computer.
    important besides textual                  investigator/author?
    information.


 Results:                                                                             Conclusions &
 1. Most positive scores related to retrieval precision were much lower than            recommendations:
    the ideal maximum, even though the queries contained very specific words          1. Information retrieval from the
    and names as formulated by a user familiar with the subject domain, and              WWW through image searching is
    the query syntax was simple and correct.                                             attractive, simple and fast, but
                                                                                         far from perfect.
                                                                                         Image searching deserves a
                                                                                         place in the offerings of
                                                                                         academic libraries and in
                                                                                         particular in those domains

                            J                              J                             where visual information is
                                                                                         important.
                                                                                      2. Misinformation in WWW
                                                                                         searching cannot be neglected.
                                                                                         Therefore libraries should
                                                                                         include this problem in their

     L                                                     L
                                                                                         efforts to increase the level of
                                                                                         information literacy of their
                                                                                         users.
                                                                                      3. Libraries can create digital
                                                                                         information sources and the
                                                                                         chance that these can be
                                                                                         retrieved with current general
 2. Beyond irrelevant, neutral results that lower the precision of a search, a           WWW search engines is high.
    significant number of search results may even mislead users who are                  Of course they should pay
    unfamiliar with the selected subject domain.                                         attention to those aspects of
 3. WWW image searching functioned well for specific HTML documents with                 their procedures that influence
    images, which have been placed on an academic library WWW site by the                the chance that the information
    investigator / author.                                                               source is found through WWW
                                                                                         searching and more particularly
                                                                                         through image searching.

Partly published in            P. Nieuwenhuysen
WWW image searching delivers high precision and no misinformation: reality or ideal?
Issues in Informing Science and Information Technology, Vol. 7, pp. 109-131. ISSN: On-line 1547-5867; CD 1547-5859; Print
1547-5840, peer reviewed journal,
available online free of charge from http://iisit.org/Vol7/IISITv7p109-131Nieuwenhuysen770.pdf


    5th UNICA Scholarly Communication Seminar: « Find it, Get it, Use it, Store it » Lisboa/Lisbon, Portugal      2010-11