Chapter 2: Smoother pebbles and the shoulders of giants: the developing foundations of information science – David Bawden

In which the author traces the development of five main themes in information science research: the information discipline, foundations of the discipline, nature of information, relations between discipline and profession, and education for information science.

(David Bawden is Professor of Information Science at City University in London and director of the Information Science Centre within the University’s Department of Information Science. He is also editor of the Journal of Documentation. He began his professional career in pharmaceutical research information services.)

The emergence of an information field was clear, but it was deemed more as an ‘applied multidisciplinary art’ rather than ‘an academic and applied study’. It lacked a conceptual base to build quantitative analysis of the objective observables.

But is it realistic to approach information science using the physical sciences model? One information scientist in the 1980s suggested that information science might just be an integrating science, pulling together contributions from other sciences – linguistics, mathematics, computer science, etc – and ‘may evolve into a body of knowledge and methodology distinguishable from other sciences’.

Nonetheless, without an academically coherent core, it is difficult for the results from information science to filter into other disciplines.

The author argues for a ‘more fundamental adoption of concepts and perspectives’ rather than ‘a more superficial adoption of techniques and algorithms’.

Of notable papers that established a foundation for the discipline, the author listed:

(a) Bertie Brooks.

The foundations of information sciences, Part 1: philosophical arguments, Journal of Information Science, 2(3/4), pp125-133, 1980. (Note: Brooks published a series of 4 papers in the first three issues of JIS.) This paper uses Popper’s ‘three worlds’ ontology as a foundation for information science, and presents a fundamental equation. Brooks then dealt with the mathematics required to describe and analyse information spaces. (For more on Bertie Brooks, see this special issue of Journal of Information Science, 16(1), 1990.)

Popper’s ontology comprises three worlds. World 1 was physical items, world 2 was the subjective mental state and personal knowledge, and world 3 was communicable information (e.g. contents of books, database). This ontology is no longer taken seriously in philosophy, though it still influences information scientists.

Brooks proposed a fundamental equation of information systems:

K[S] + dI = K[S + dI],

where K[S] is the knowledge structure and dI is increment information. However, Brooks acknowledged that this is still a conceptual representation, with terms ill-defined and interdependent, such that they cannot be used for calculation.

(b) B. Hjorland

Hjorland put forward a socio-cognitive perspective using domain analysis. For more:

- Theory and meta-theory of information science: a new interpretation, Journal of Documentation 54(5), pp 606-621, 1998.

- Epistemology and the socio-cognitive perspective in information science, Journal of the American Society for Information Science and Technology, 53(4), pp 257-270, 2002.

- Domain analysis in information science: eleven approaches – traditional as well as innovative, Journal of Documentation, 58(4), pp 422-462, 2002.

Of notable papers on the nature of information, the author listed:

Jason Farradane

- The nature of information, Journal of Information Science, 1(1), pp 13-17, 1979.

Farradane considered information as ‘any physical form of representation, or surrogate, of knowledge, or of a particular thought, used in communication’. This viewpoint is echoed today amongst those who view knowledge as personal and information as what is communicated. Such a definition of information would lead to information science developing as an experimental science. For him, philosophical conceptualisations and mathematical modelling were secondary to experimental analysis and predication.

This contrasted with Brookes’ view that knowledge and information are the same in essence, differing only in coherence. This echoes another common representation today of information in the form of an information hierarchy: data-information-knowledge-wisdom.

On the relation between the discipline and profession, the author finds one mainly of disconnect. The researchers find their research disregarded by practitioners, while practitioners find research irrelevant.

On the future of information science, the author believes that it needs a technology invariant core. Only then can it not be incessantly threatened by new technologies and be able to exploit (rather than react to) these technologies. He sees this core as the study of users and the underlying social and organizational structure of information.

On the relations between information science and other disciplines, the author noted how professionally, IS in UK has merged with the library profession, while IS in the USA has merged with information technology. The author feels it useful to distinguish between information science (which deals with ‘individual information behaviour seeking and use’) and the collection sciences of librarianship and records which has a role as custodians.

Chapter 3: The last 50 years of knowledge organization: a journey through my personal archives – Stella G. Dextre Clarke

In which the author uses her personal experience to highlight the developments of different knowledge management concepts and tools, such as classification, thesauri, and taxonomies.

(Stella G. Dextre Clarke has worked in the public sector on information retrieval. She is best known as the principal architect of the UK Integrated Public Sector Vocabulary.)

The run-up to the 1950s: the rise and rise of classification. Classification (e.g. schemes such as DDC, UBC, Bliss and the concept of faceted classification) had seen its heyday. There were new developments, but none that made fundamental advances. Thesaurus were still new. The 1959 thesauri by B.E. Holm was the first fully operational thesaurus.

1960s and 1970s: era of thesaurus. Many scientific and engineering thesauri were published (e.g. by the armed services, scientific institutes). Since the alphabetical system in thesauri limited information retrieval, they were soon integrated with faceted classification schemes. Early computers searches had incorporated thesauri for better retrievals. With computers still limited in memory, research generally aimed to keep things small.

Card systems. The high cost of computers meant that elaborate card systems were more popularly used. In one such system, keywords could be combined to refine searching. For example, to search “pruning of roses”, one would pick out the cards for “pruning” and “roses” and looked for documents common to both. Huge cards were designed for such purposes, with up to 10,000 punched holes (each representing one document). By putting the cards on a lightbox, one could see the common document-holes illuminated. There were also mechanized card systems, but these were far less elegant and used brute force to run through all the possible combinations. They also tend to have limited index terms. Index card catalogues were also common, but they require duplicates of each document-card for each subject heading. To manage such an unwieldy system, some places designed excellent thesauri with inter-term relationships to facilitate the post-coordinate searching.

1990s: PCs and Taxonomies. While PCs and the internet could in theory bring all the resources to people, the reality was that finding relevant resources was still difficult in practice. Taxonomies became the rage, even though they were just adaptations of classification schemes and controlled vocabularies. With the internet came inter-operability, hence there was a rise in meta-data research, e.g. Dublin Core Metadata Initiative and ISO standards.

Other recent efforts and technologies: ontologies, folksonomies, organising images and multimedia resources, cross-language retrieval; clustering, collaborative filtering, Google-type search algorithms.

Overall… the author felt that there had been no fundamental developments of principles in recent times, but mostly adaptations of new technologies.

Chapter 4: On the history of evaluation in IR – Stephen Robertson

In which the author describes and discusses evaluative experiments in information retrieval, paying particular attention to TREC.

(Stephen Robertson is a researcher at the Microsoft Research Laboratory in Cambridge, UK. He has a part-time professorship in the Department of Information Science at the City University, and is a Fellow of Girton College, Cambridge.)

The author recommends two key sources:

Pre 1980s: Information retrieval experiment, Karen Sparck Jones (ed), 1980.

Post 1980s: TREC: Experiments and evaluation in information retrieval, E.M. Voorhees and P.K. Harman (Eds), 2005.

Classification and worldviews. Each classification system carries with it a worldview of the underlying structure of human knowledge, thus it was difficult to adopt a more functional and experimental approach to deciding their relative merits.

Cranfield experiments. Initiated by Cyril Cleverdon, Librarian at then Cranfield College of Aeronautics. This set of experiments met the philosophical objectives by pitting classification schemes against each other. Each scheme was represented by a team of its experts, and together, the four schemes in the competition covered the spectrum of information organization.

These schemes were: Universal Decimal Classification (hierarchical library classification), Alphabetical Subject Catalogue (subject headings expressed as phrases), Faceted Classification Scheme (combing elements from different facets to form a more complex category, and Uniterm System of Coordinate Indexing (with terms freely assigned and combined).

Two experimental results the author highlighted were: (a) there was not much difference in performance, (b) that faceted classification moved from worst to best when a different form of card index was used, suggesting that relative merits depended a lot on the practicalities of implementation.

From Cranfield 1 to Cranfield 2. Cranfield 1 used a document retrieval test: the author of the document phrased a query that allowed the document to be retrieved. This was clearly too different from real life searches when users do not know which documents are relevant. In Cranfield 2, the issue of relevance was probed and judged using a number of methods, from manual searching to citation-based analysis.

Notable research groups: Bruce Crott at the University of Massachusetts Centre for Intelligent Information Retrieval, Nick Belkin at Rutgers University, SMART group at Cornell University, the Okapi group in the UK.

TREC. The Text REtrieval Conference was led by the US National Institute of Standards and Technology. It is a competition of topical searches. However, its collection size is still very small, so it is questionable the extent which successful methods are scalable at web-scale. Developments it spurred include ranking algorithms, relevance feedback (e.g. using feedback to deduce relevance of a new stream of documents), and bringing in machine learning.

Tackling the web. The web differs from TREC collection: much more diverse, of extremely varied quality, of a much larger scale. However, the web has useful features such as hyperlinks and anchor texts that describe the new page. (Wens: I would add that user behaviour would be different too; given the quantity of information, the next best article would still be relevant, whereas the next best article in a small collection could be quite far off.)

Problems with TREC. TREC had stimulated research, improved experimental methodologies, and advanced information retrieval technologies, but (a) there has been some narrowness in outcomes due to the competitive nature with a lot of focus on effectiveness and less in-depth failure analysis, (b) the lab is not real life, and labs reduce many real-life variables into simplistic abstractions, (c) goals defined are restrictive and artificial compared to issues faced in the real world.

“Information Science in Transition“ (edited by Alan Gilchrist) is a commemorative collection of essays that discusses the development of information science in UK in the past 50 years. This collection first appeared in a special issue of the Journal of Information Science in 2008 to commemorate the founding of the UK Institute of Information Scientists, and was re-published in a book form by CILIP in 2009.

The essays are wide-ranging: from knowledge management to highly specialised branches such as chemoinformatics, on the practical applications of information science in information policies and health information systems, new issues arising from electronic publishing and social software.

The table of contents is as follows:

Guest editorial: Meeting the challenge – Brian Vickery
In which the author discusses how the fast changing capabilities of computers and the internet presents challenges.

1. Fifty years of UK research in information science – Jack Meadows
In which the author surveys the Journal of Information Science and Journal of Documentation for major research themes, and finds repeated interest in information retrieval, information seeking, bibliometrics and communications. The author briefly looks into impetus for these research focus and some of their major projects.

2. Smoother pebbles and the shoulders of giants: the developing foundations of information science – David Bawden
3. The last 50 years of knowledge organization: a journey through my personal archives – Stella G. Dextre Clarke
4. On the history of evaluation in IR – Stephen Robertson
5. The information user: past, present and future – Tom Wilson
6. The sociological turn in information science – Blaise Cronin
7. From chemical documentation to chemoinformatics: 50 years of chemical information sciences – Peter Willett
8. Health informatics: current issues and challenges – Peter A. Bath
9. Social informatics and sociotechnical research – a view from the UK – Elisabeth Davenport
10. The evolution of visual information retrieval – Peter Enser
11. Information policies: yesterday, today, tomorrow – Elizabeth Orna
12. The disparity in professional qualifications and progress in information handling: a European perspective – Barry Mahon
In which the author gives a quick outline of the developments in Europe for information professionals, and concludes that these developments have been mostly disappointing and late.

13. Electronic scholarly publishing and Open Access – Charles Oppenheim
In which the author reviews of developments in electronic publishing. A comparison of the two main types of open access -  ‘gold’ open access journal route and the ‘green’ repository route – in terms of quality, cost, copyright issues, business models and publishers’ reactions.

14. Social software: fun and games, or business tools? – Wendy A. Warr
In which the author introduces a number of social networking tools (e.g. blogs, social networking sites, social bookmarks), and describes adoption of these tools by information professionals (e.g. UK National Archives wiki, CILIP’s meeting on Second Life) and publishers (e.g. Nature’s Second Nature on Second Life, Elsevier 2collab and Scirus Topic Pages).

15. Bibliometrics to webometrics – Mike Thelwall
16. How I learned to love the Brits – Eugene Garfield

Guest editorial: Meeting the Challenge by Brian Vickery
(Brian Vickery was a chemist turned information scientist. He helped to form the Classification Research Group in 1952. Before he retired, he was Director of the School of Library, Archive and Information Studies at University College London.)

What is the work of information scientists? To this, Vickery suggests that just as medical practitioners diagnose and treats, information scientists diagnose information needs in different user situations, and address the need by developing a range of methods to identify and provide that information.

It will not come as a surprise that Vickery highlights the computer and the internet as the two important challenges for information scientists. Part of the problem lies in how fast the capabilities of these two technologies are changing. Computers used to be very limited in data storage and manipulation; there was a time when filenames cannot be more than 8 characters to save memory space, but today’s MP3 player has memory space that could rival a mainframe computer then. The internet began as a form of person-to-person communications for file sharing and writing memos, and today it has not only grown exponentially in scale, it has also thrown in business services, applications, social networking, and other functions that are beyond the experience of library and information professionals.

Vickery believes that there is still a need for information professionals to organize ‘knowledge organization systems’ for users to navigate the ocean of knowledge. Such systems have to trade comprehensiveness for simplicity and ease of use. The structural tools that were developed over the decades – hierarchical classification, facets, thesaural relations and so on – can still be put to use (e.g. corporate information systems have been spending effort on taxonomies). The challenge is to integrate these effectively into computer-based information systems: making use of the computer’s capabilities for complex manipulations of texts, graphics and data, in a way that address information users’ needs and behaviours.

Vickery also acknowledges the need for re-thinking information science at a more fundamental level. He quotes fellow contributors to this collection of essays:

Barry Mahon: “It remains to be seen if the current professional bodies recognize the benefits to be gained by embracing the ideas created by the newcomers to the information sector.”

David Bawden: “We have not fully worked through the insights of the founders of the discipline, let alone replaced them with entirely new insights.”

Stella Dextre Clarke: “The greatest part of the credit [...] goes to enhanced information technology rather than fundamental new thinking in knowledge organization.”

His conclusion: the coming future can be a period of experimentation that leads to new ways of handling information.

Chapter 1. Fifty years of UK research in information science – Jack Meadows

(Jack Meadows was a physicist turned historian of science turned information scientist. He headed the Department of Information Science at Loughborough University.)

For this essay, Meadows surveyed two main library and information science journals: Journal of Information Science and Journal of Documentation.

One consistent theme throughout the past fifty years was information retrieval. In the past twenty years, new clusters of research were formed around information seeking, bibliometrics and communications.

Information Retrieval

Two landmark research projects were the Cranfield experiments in 1950s and the TREC (Text REtrieval Conference) which began in the 1990s and is still a major focus of research today.

The main drawback in early research was that it assumed that all retrieved documents had equal relevance. Later research gave rise to new ways of determining relevance, e.g. methods such as frequency of words in different parts of the documents.

Later research made use of user feedback, e.g. Okapi which had an interactive approach that incorporated user’s feedback of relevance into new searches. The focus on user interaction led to cross-fertilization between information seeking and information retrieval.

Chemoinformatics and Bioinformatics began as specialised branches of information retrieval of chemical structures. They have become a sub-field in themselves today, playing important roles in drug discovery. (The development is elaborated in Chapter 7.)

Information seeking

Information seeking saw a shift from quantitative studies to qualitative ones during this period, with increasing interest in social science information seeking. Methodologies have been adopted from the social sciences, e.g. soft systems methodology (P.B. Checkland and J. Scholes, Soft systems methodology in action, 1990).

There was also a shift from looking at usage of information to the methods which users employ to seek information. One model developed looks at the links between information needs and information-seeking behaviour (T.D. Wilson, On user studies and information needs, Journal of Documentation 37(1), pp 3-15, 1981).

There was also interest in new interfaces (e.g. B.A. Vickery & A. Vickery, Online search interface design, Journal of Documentation, 49(2), pp 103-187, 1993).

Communications

Research in this area focused on the interactive exchange of information rather than documented information.

Information Chain

Scholars have been interested in R&D communication, especially through formal publications such as journals. This chain had four players: author -> publisher -> library/information centre -> reader. Traditionally, the information scientist would focus on the last two players in the chain. But this chain has been broken by digital publishing and communications, forcing a re-examination of the role of the publisher and information centre. Information scientists thus have to study the whole chain, beginning with the author.

Open access

There has been a lot of debate about open access, such as issues of copyright. In Britain, this is further complicated by EU regulations, which differ from its traditional approach.

Meadows provide some references: D. Nicholas et al, The impact of open access publishing (and other access initiatives) on use and users of digital scholarly journals, Learned Publishing, 20(1) pp 11-15, 2007; E. Gadd et al, RoMEO studies 1: the impact of copyright ownership on academic author self-archiving, Journal of Documentation, 59(3) pp 243-277, 2003; C. Oppenheim, Does copyright have any future on the internet? Journal of Documentation 56(3) pp 279-296, 2003.

A quick glimpse at their current cloud tag reveals a few dominant clusters: social issues (e.g. healthcare), US foreign policy (esp with the Muslim world) and terrorism, economic policy & energy & environment.

While these video lectures are already freely available on the university’s websites (e.g. MIT’s OpenCourseware, Stanford’s Engineering Everywhere and Yale’s Open Courses), it’s absolutely amazing when they are all aggregated into one user-friendly website.

See how stats should be used as Hans Rosling debunks myths about the “developing world” at TED using visualisations of time-series country statistics.

The trendanalyzer he used is now freely available at gapminder.org for anyone to play with and appreciate country trend data as diverse as income per person, mathematics achievements, and consumption of sugar and sweeteners.

Behind this somewhat enigmatic domain name is a rich and wonderfully accessible set of country statistics gathered from reliable sources such as OECD and CIA World Factbook. This website has been around since 2003, and has received a number of excellent reviews, from Harvard Business Review, amongst others.

The data is grouped logically into the usual categorical suspects (e.g. economy, education), which helps you discover some strange and fascinating statistics (e.g. number of roller-coasters per capita). The search function works pretty well too.

The beauty of nationmaster is that it presents the data of that year from all participating countries in a bar chart, sorted in descending order, which allows you to decide quite quickly whether to use the data. You can also pre-select the countries in the “statistics” tab, or zoom into all the data available for a single country in “country” tab.

The flipside of bar charts is that you have to manually extract the data should you want to perform trend analysis across years or manipulate the data further, which is very likely if you are going to use the data for staff papers or projects, instead of just entertainment.

Afterall, bar charts are still quite primitive in the world of data visualisation. By this, I don’t mean fancy 3D effects that add nothing to the meaning of the data. I’m referring to other forms such as treemaps, mapping data onto geographical locations, animating trends… essentially, methods that highlight the meaning the data is conveying.

I’ll be putting up more links on other statistics and data visualisation websites. If you know of some, please share them with me.

Singapore’s National Library has begun digitizing books about Singapore, the highlights of which are rare books dating back to the 1800s.

Typical items in this digitized collection are memoirs (e.g. “Memoir of the Life and Public Services of Sir Thomas Stamford Raffles” by his wife in 1830) and guides (e.g. “A Handbook of Colloquial Malay as Spoken in Singapore” published in 1878).

What you get are pdf versions of the original book in the National Library’s holdings, so you see the yellowed pages and original typefaces, and also barcodes and stamps of ownership. The pdf version is excellent for searching within the document, but you wouldn’t be able to search the document text through the website’s search engine.

If you are a teacher thinking of “creating different experiences to spark student’s imagination about Singapore” (Recommendation 8, Report by Committee on National Education, MOE, 2007), you could…

• Get your students to read through chapters in the memoirs, and act / draw out the scenes
• Get your students to read about events, and re-write the same incident through the eyes of a journalist
• Use the more historical sentence examples in the language guides to form the first lines of composition passages

• Use excerpts from the memoirs and contrast them with more official accounts

I’ll like to hear about how these books can be used in lessons, so if you have dones so, please drop me a note.