In online catalogues, subject searching has proven particularly popular: it accounted for about 59 per cent of all catalogue enquiries in a 1983 study (Taylor 2000, p.266; Roe 1999, pp.70-71), a level that appears to have been maintained in the past two decades (Yu & Young 2004), though of course the exact proportion varies from library to library. Consequently, providing effective subject access to information is an essential part of offering a good catalogue. Likewise, users often want to search by subject on other kinds of bibliographic databases, such as those that cover journal and newspaper articles; indeed, subject access is even more important where documents are briefer and thus less sought after specifically. Yet, despite its significance, subject access is the part of information organization that is the most difficult to get right. Although we have made significant progress in providing access to information resources in the digital age, ‘intellectual access to the contents…has improved very little, if at all’ (Lancaster 1999, p.49). Although search engines such as Google have vastly improved subject access to the web over the past decade, even here result sets include a lot of irrelevant materials, and exclude a lot of much more relevant ones.