025.431: The Dewey blog

Didn't all this get sorted out in, like, 1876?

Okay, we know his behavior wasn't always exactly saintly, but to ask "Was Dewey the anti-Christ?" seems a little harsh. This is the question posed by Beth Dempsey in her article "Power users: Designing buildings and services from the end user's viewpoint transforms access for everyone" (Library Journal 130, no. 20, December 2005, pp. 72-75). Apparently the DDC "violates the principles of good information architecture on multiple levels;" and, instead of using Dewey numbers on signs and shelves, libraries should follow the chain bookstores' lead and "put cookbooks on a shelf marked 'Cookbooks'." Here at Dewey Manor we're in full agreement with the main thrust of the article, which is that it's often difficult for library users to find their way around in library buildings, but that there are a few simple steps that librarians can take to help users more easily find the stuff or the services that they want. Hey, better signage -- what's not to like? But we'd like to submit that it's not really the case that libraries have to choose between Dewey numbers and verbal labels for subjects. Many libraries have signs that use a combination of the two -- which can only be a good thing because the Dewey system provides a number of benefits that you don't find in the average bookstore: (a) a way of describing subjects that is independent of any natural language, (b) a standard order of main classes, and (c) hierarchical arrangement of topics within main classes, and that's just for starters. For the user whose first language may not be English, and who would appreciate the opportunity to browse shelves of works on different but related topics, arranged in a systematic, standardized order that only needs to be learned once, and only to the level of detail that meets the needs of the individual, surely a combined DDC/verbal system of signage can only be helpful? Or ... did Dewey really get it so wrong? Dear readers: let us know. (And thanks to Alice for pointing us to the LJ article from It's all good.)

ASIS&T 2005: Day 4

So the ASIS&T schedulers knew what they were doing all along, of course, by saving the best till last. In a session on the “Use of classification in information seeking” jointly sponsored by SIG/CR and SIG/USE, Barbara Kwaśnik (Syracuse University) described the variety of ways in which the effectiveness of information retrieval systems can be improved by the partitioning of collections of resources into meaningful clusters. Barbara enumerated the challenges faced by the user of retrieval systems, such as the formulation of queries in the absence of domain knowledge and the interpretation of large, undifferentiated sets of search results, and showed how the classified organization of resources can help the searcher meet each of those challenges. Barbara allowed that no single classification scheme is appropriate for all purposes, but her characterization of a “good” scheme as one that is elegant, parsimonious, expressive, complete, memorable, usable, flexible, and hospitable reminded Team Dewey of one scheme in particular. (Clue: Not Zastrau.)

Joseph Busch (Taxonomy Strategies) then treated us to an excellent overview of the role of classification in search systems that support organizational goals such as e-commerce and e-government. Joseph argued that search engine users typically only think they prefer “type-and-go”/“feeling-lucky” interfaces over those that allow browsing of faceted hierarchies of classes, mainly because designers are not being as creative as we might expect them to be in the development of browsing-based systems. Joseph referred to research by Hearst, Dumais, and others that quantifies not only the extent of user preference for faceted interfaces but also the magnitude of the benefits (in terms of cost and time) to be obtained by organizations implementing such systems. Joseph demonstrated a number of existing, web-accessible systems that offer browsing-based access to information about products, services, and documents (see, for instance, a British supermarket’s faceted interface to its database of wines), before describing a case study of the opinions of users of a system in the domain of agricultural economics which showed that users found it easier to search by non-topical facets (i.e., geographical area and commodity type) than by topic. Joseph’s conclusion that the best search interface is the one that offers faceted navigation in combination with keyword searching serves as a ringing endorsement of the approach taken by DeweyBrowser.

ASIS&T 2005: Day 3

The parallel session – indubitably the most helpful of conference traditions. What could be more sensible than to schedule at precisely the same time the only two sessions (out of more than 60) that include the keyword “ontologies” in their title? What motivation could anyone possibly have for attending both “Science ontologies and the semantic web” and “Metadata and ontologies”? None, obviously, so you can imagine Team Dewey’s jubilation at having to choose between the two. “Science ontologies …” (sponsored by ASIS&T’s Special Interest Group on Scientific and Technical Information Systems, SIG/STI) won the coin-toss, and the highlight was Ralph Hodgson’s measured (and refreshingly jargon-free) tour of meaning quadrants, OWL triples, and knowledge technology trees. Ralph (TopQuadrant, Inc.) cunningly compressed a day-long seminar into 20 minutes, teasing us with the promise of a forthcoming beer ontology (“Beer … is_a … beverage,” anyone?) that will surely revolutionize our understanding of the semantics of the brewing industry. One minor disappointment for Team Dewey was the acronym-per-slide ratio of this session: an average ApS score of 30 really doesn’t cut the mustard in the modern age, and as far we could determine very few of the acronyms used seemed to have been made up specifically for this session – a sad indictment of the current state of R&D in acronym science.

ASIS&T 2005: Day 2

With the pre-conference seminars and workshops now but fond memories, the ASIS&T conference proper got underway on Sunday, and your intrepid representative from Team Dewey decided to indulge an interest in subject access to visual resources by checking out a pair of sessions organized by SIG-VIS -- the Special Interest Group on Visualization, Images, and Sound. In the first, Paul Roy from Arius 3D demonstrated hardware and software for scanning and visualizing digital images of three-dimensional objects, and other panel members described studies of the various ways in which such technology is being used to enhance the level and quality of access to information about cultural, educational, and scientific resources. Speakers in the second session explored the roles that collections of digital representations of museum artifacts play in the lives of museum visitors. Paul Marty (Florida State University) presented the results of a survey of visitors' opinions which demonstrated that the relationship between a physical museum and its digital counterpart is typically complex, and (for instance) that visitors very strongly prefer to satisfy certain of their interests by interacting with one or the other. Ken Hamma (J. Paul Getty Trust) showed how museums can engage the attention of younger visitors by providing online access to collections through virtual communities such as Whyville; and Layna White (San Francisco Museum of Modern Art) and Anne Gilliland (University of California, Los Angeles) summarized what researchers working on the Museums and Online Archives Collaboration (MOAC) project have learned about educators', students', and museum professionals' use of online information about museum objects.

Two of the themes that emerged from the discussion are, first, the awareness of a convergence in the descriptive practices of librarians, archivists, and curators, such that professional staff of all three kinds of institutions are increasingly recognizing the benefits of close collaboration on projects in the field of cultural informatics -- to the extent that LAM (libraries, archives, museums) is now one of the more common acronyms in use at this conference; and second, the promise of collaborative indexing systems that allow users of digital collections to assign their own labels to representations of cultural resources. In the final talk of the second session, Hsin-liang Chen (University of Texas at Austin) asked whether the positive reception of initiatives such as the Cleveland Museum of Art's Personal Collections project was indicative of a paradigm shift in our understanding of how objects are best described. Ken Hamma followed up this point with a tantalizing reference to a new project in which the Getty Museum will be exposing its collections of digital images for harvesting by other services, with the intention that those third-party services may then choose either to provide users with access to Getty-supplied metadata or to create and offer access to new, community-specific descriptions and indexing.

ASIS&T 2005: Day 1

This year's meeting of the American Society for Information Science and Technology (ASIS&T) kicked off on Friday, and a delegation from Dewey Manor was there on Saturday for the 16th Workshop of ASIS&T's Special Interest Group on Classification Research (SIG/CR). Workshop chair Jack Andersen (Royal School of Library and Information Science, Denmark) and his committee had put together a fascinating program of talks, and participants greatly appreciated the amount of time that was built in for general discussion.

Unfortunately, scheduled keynote speaker Joacim Hansson (University College of Borås and Göteborg University, Sweden) was unable to attend, so Fernando Elichirigoity (University of Illinois at Urbana-Champaign) got things underway with a critique (co-authored by Cheryl Knott Malone of the University of Arizona) of the North American Industry Classification System's coverage of the "new economy" that is based on open-source software production. Fernando and Cheryl argued that data on the productivity of this sector remains unrecorded due to the absence of appropriate classes in NAICS, and that this absence is partly a result of NAICS' use of copyright-holding status as a criterion when identifying productive units.

Melanie Feinberg (University of Washington) reported on her comparative study of the ways in which library collections arranged using different classification schemes express the relationship between feminism as a concept and women as a social group. Melanie's paper was followed by discussion of how researchers might distinguish in such studies between the role of the classification scheme itself and the role of individual classifiers' interpretations of the scheme; some attendees identified an urgent need for further user studies in this area.

Hans Jørn Nielsen (Royal School of Library and Information Science, Denmark) presented the first of the day's two papers on the indexing of imaginative literature (i.e., fiction), asking whether it is desirable or even possible to index the subject matter of such literature given that it is characterized by ambiguity and plurality of meaning. Some attendees suggested that all texts, literary or otherwise, have different meanings for different readers; does that mean that all indexing is worthless? The group briefly considered the apparently timeless question of whether manual classification remains necessary given the capabilities of automatic indexing systems, before discussing the value of a "third way": collaborative indexing, by members of particular communities of practice, using a wiki-based system to assign multiple class-labels that represent multiple points of view.

After lunch, Frank Exner, Little Bear (North Carolina Central University) gave a detailed account of the issues involved in establishing effective authority control over North American Indian names. Three challenges in particular stand out: the diversity of patterns formed by the constituent elements of American Indian names; the fact that many people retain more than one name at any given time; and the fact that many people are known by different names at different points of their life. Little Bear concluded by making specific recommendations for changes in AACR2.

In a paper co-authored with G. Craig Murray, Kara Reuter (University of Maryland, College Park) summarized four cognitive-psychological accounts of the variation observed in children's and adults' performance of tasks that involve categorization. Kara argued that such performance depends primarily on context, and that the development of categorization ability should most properly be understood as the development of cultural expertise. Kara identified the need for further study of the implications of these findings for our understanding of children's and adults' ability to use bibliographic classification systems.

In the final paper of the day, Rune Eriksson (Royal School of Library and Information Science) revisited the theme introduced earlier by Hans, focusing on the ways in which the classification of imaginative literature has historically been treated by the DDC. Rune spoke persuasively of the benefits of a faceted approach that allows for the expression of the themes, styles, narrative structures, and sub-genres of literary works (as well as language, form, period, and place), and the group identified the potential value of browsing-based systems that take advantage of the DDC's faceted structure. And then, suddenly, it was 5 o'clock. Everyone agreed that we'd done a good job, and that we should celebrate by finding the nearest bar. Cheers!

Eccles takes the cake

This one falls squarely in the category marked "You couldn't make it up." Bear in mind, if at all possible, that what you are about to read (which, with apologies to all fans of cheap sarcasm, we present without further comment) is not a lost Monty Python script, but a transcript of what really was said in the UK Parliament's House of Lords on June 9, 2005.

Lord Harris of Haringey asked Her Majesty's Government: Whether the time spent preparing the e-Government Unit's document Tomatoes are not the only fruit: A rough guide to taxonomies, thesauri, ontologies and the like, represents value for money.

Lord Bassam of Brighton: My Lords, yes, the document was published in 2002 by the Office of the e-Envoy, at the request of technical users in government who were new to the subject. It was produced in-house at an estimated cost of less than £100.

Lord Harris of Haringey: My Lords, I am grateful to my noble friend for that information. I ask him to congratulate the civil servants concerned on the diligence and speed with which they must have produced 12,000 words and four charts on the subject of Tomatoes are not the only fruit, containing such gems of information as: "How long has it been for many of us since the primary meaning of the word 'mouse' has been 'a small furry animal that frightens elephants?'," or the information that carrots can be either salad or root vegetables. That will no doubt come in very helpful in promoting e-government. Can we also congratulate the authors of the Guide to meta-tagging with the Integrated Public Sector Vocabulary, which gives another eight pages of valuable advice and information? It includes the information that the phrase "Common Agricultural Policy" may appear under the phrase "European Union" or under "Farming" but will mean the same under both. Given the diligence of the civil servants in the unit, can the Minister assure the House that the same energy and effectiveness is being applied to delivering information security throughout the public sector? ...

Lord Bassam of Brighton: My Lords, I shall of course pass on my noble friend's congratulations. However, I have a sneaking suspicion that civil servants are as we speak listening carefully to his kind congratulations and warm words. ... The document, although it has attracted a certain levity, is I am sure most useful to those who work in government IT services.

Earl Ferrers: My Lords, can the Minister not pass on congratulations to the civil servants on producing a document that is completely incomprehensible to a normal person and really does not make any sense at all? Why cannot they learn to write English?

Lord Bassam of Brighton: My Lords, having looked at the document, which has a modest number of words, I disagree with the noble Earl. I believe that IT users in the business will probably find it very helpful and useful. Certainly, when I read it, I began to understand notions relating to taxonomies, thesauri and ontologies.

Lord Maclennan of Rogart: My Lords, although the use of what the Civil Service calls "controlled vocabulary" may operate as a disincentive to get online to the Government, can the Government say whether there has been a significant improvement in the access to e-government over the past two years? ...

Lord Bassam of Brighton: My Lords, there have been improvements, and I am sure that the e-Government Unit is aware of its role and responsibility in ensuring that these improvements continue. ...

Viscount Eccles: My Lords, is the Minister aware that if he were unfortunate enough to have cancer of the bladder, medical advice would be that he should eat no more tomatoes? Is that information on the website?

Lord Bassam of Brighton: My Lords, not that I am aware, no.

Lord Peyton of Yeovil: My Lords, I really wonder whether the noble Lord is aware of the extent to which he has attracted to himself this morning the wholehearted sympathy of the House at the appalling ordeal that he has had to go through in not giving a single answer to a question and really fluffing what he has said beyond the limits of comprehension.

Lord Bassam of Brighton: My Lords, the noble Lord always makes generous remarks across the House, and I suppose that I have to be the beneficiary of those remarks on some occasions.

A voice of reason

Commenting on folksonomies ("personal Dewey Decimal System[s]"), Ben Vershbow at the Institute for the Future of the Book reckons that, when it comes to making sense of tags, "librarians may in fact become the most important people of all ... They are after all the original masters of metadata ... Librarians could become the sages of the web." We like it, Ben. But next time, less of the "may," "could," and "become," okay?