T. G. McFadden Writes:
\"Questia has recently made changes to its search-and-retrieval interface (in response to suggestions from users, according to the marketing side) that represent a pretty fair misunderstanding of how the typical undergraduate will want (or need) to use the database.
Prior to this change, the initial search screen (“Quick Search”) presented the standard author, title-word, and subject options. More advanced variations on these basic themes were available in the “Power Search” mode. Now, however, the initial search screen (still “Quick Search”) combines by default all of these search types into a single search statement. This has the following result, when the search concept is the rationalist philosopher Descartes.
Much MoreAn original “quick” author search on Descartes yielded 5 titles. This (on 5 April 2001) remains true of a “power search” on the same term. However, now the initial “quick search” result is 35 items. The result is an undifferentiated mix of author, title-word, and subject works. It is still true that one of the titles, An Introduction to Philosophy Through Literature, contains a very brief selection from Descartes, amounting in all to about 3 pages. Another title, Dreaming by Norman Malcolm, is not by Descartes at all; the book is in fact about Descartes’ dream argument. I pointed this out to Questia after my original experiment, but these anomalies have not been corrected (nor have any of the others I identified at that time, sofar as I can tell). However, by changing the initial search statement into an inclusive strategy, the problem that some titles actually by an author might, for instance, turn up incorrectly in a subject search, or vice versa, has been neatly solved. Since there is no distinction anymore among bibliographic types within the “quick search”, no titles turn up in the wrong list; there is only one list at this point.
However, the utility of the results of the “quick search” so formulated is highly problematic. If the database grows at a roughly proportional rate for all of its parts (which it won’t, but the example will stand for now), then the “quick search” for Descartes will eventually yield some 250 undifferentiated titles. Moreover, the whole concept of sorting this undifferentiated list by degree of relevance (still an option) is without meaning. The idea of relevance is strictly relational; to try to sort a combined list of this kind by degrees of relevance immediately presents an apples and oranges paradox.
Incidentally, none of the 5 titles yielded by an author search on Descartes are important or even selected editions of Descartes’ primary works. This is also a serious problem.
Tom McFadden, Director
Schenectady, NY 12308