EDITOR’S SUMMARY

At the ASIS&T Annual Meeting in Copenhagen, Peter Ingwersen expressed surprise and gratitude for being honored with the Association’s 2016 Award of Merit. The professor emeritus from Denmark’s Royal School of Library and Information Science thanked peers for recognizing him as a mentor and research originator. Ingwersen observed that his own areas of research in interactive information retrieval (IR) and scientometrics/webometrics have come together as converging lines of interest, though research and analysis are often restricted by system and feature limitations. Experimentation, Ingwersen stated, is too little focused on the effective workings of IR models and contexts, repeatability and negative research results, and is instead constrained by practicalities and pragmatics. Ingwersen expressed pleasure at the acceptance of web redundancy, citing his theory of polyrepresentation and noting the value of weighting term and search results underlying ranking of web search results. He credited the ASIS&T meetings and publications for developments stemming from original IR experimental studies.

KEYWORDS

Association for Information Science and Technology
honors
research and development
information retrieval
webometrics
term weighting
redundancy
reproducibility
ranking


Reflections on Receiving the ASIS&T 2016 Award of Merit

by Peter Ingwersen

ingwersen-photoDear colleagues and friends; ladies and gentlemen!

It was a nice and highly appreciated surprise to me to have the honor to be presented with the prestigious ASIS&T Award of Merit, given in recognition of my contributions to our field of information science. And then here in Copenhagen, my home town, at the first ASIS&T Annual Meeting outside the United States. We are very proud and thankful to Diane Sonnenwald and ASIS&T for this choice. Also, I wish to thank the jury that selected me and the nominators Peiling Wang, University of Tennessee, Knoxville and Virginia Ortiz-Repiso Jiménez, president of the European Chapter, who created the two nomination packages.

I have now had the opportunity to look into the two nomination texts and the many supporting letters from outstanding IS researchers worldwide. I am very grateful for their acknowledgements of my contributions to the field over four decades of research. As with citations received to your work, you obtain a range of new perspectives on your research and activities. It seems that I have had much stronger roles as mentor as well as initiator of research than I perceived. I really appreciate the recognition of that achievement.

Looking back, it is interesting to note two phenomena in association with my lines of research:

1)     How scientometric and information retrieval (IR) research are technology driven and dependent on available features in IR and citation systems.

2)     How redundancy and the ranking principle have become mainstream in research and commercial retrieval environments.

Speaking to the first point, my two lines of research, interactive IR and scientometrics/webometrics, have recently started to be seen in conjunction, including at this conference. In that endeavor I have followed in the steps of B.C. Brookes, Jean Tague-Sutcliffe and Don Swanson. In particular, in scientometric and webometric research it is interesting to observe how much we are dependent of the system structures and features available (or rather NOT available) in the citation databases and search engines on the web. With link-searching features having disappeared from Yahoo (originally in Altavista), webometric analyses have been seriously inhibited. We must rely on independent crawlers and repeatability becomes more difficult to perform. As long as Web of Science and Scopus keep their online analysis features available, research evaluation is doable and repeatable online and independent from the Leiden system.

It is also my opinion, that the IR experimentation, by and large, has been and is a technical venture. Experiments and tests are very often done because they can be done, based on technical progress and constrained by the availability of test collections of special kinds. There is less focus on why and how IR models actually do (not) work or work better than other models, in particular in IR interaction – and in which topical and media environments they are most effective. Repeated experiments as well as negative result papers are rarely submitted or published. But fortunately we see an increasing amount of user-driven IR experiments.

Turning to my second point, I want to comment on how quickly redundancy became accepted by the information profession. Up to the start of the web, redundancy was to be avoided; it was unacceptable except in faceted classification. My theory of polyrepresentation from 1994 and onwards makes deliberate use of redundancy. Polyrepresentation is a consequence of the cognitive perspective on IR, promoted by Nick Belkin, Pia Borlund, Birger Larsen and many others also present today in this room. Today redundancy is all around us, often in the form of chaos on the web. But without redundancy no weighting of terms, records and so forth can effectively take place. And from that follows logically the ranking principle in IR.

With respect to the ranking principle and other algorithmic solutions provided by IR research, we all remember how difficult it was to convince and transfer that ranking principle in experimental IR to major online vendors and applied search environments 1976-1996. They did not believe in it! With the commercial search engines (Altavista, Google, Yahoo) ranking first became a natural way of searching, and it caught on very fast to become the only available and accepted search mode on the web – and elsewhere. Remember that most of those features are grounded in experimental IR and information-seeking studies done since the 60s, in particular driven by research channels like TREC, ACM-SIGIR and, not least, JASIST, ARIST and the ASIS&T meetings.

Once again, allow me to thank you all for this distinguished honor.


Peter Ingwersen, ASIS&T 2016 Award of Merit recipient, is professor emeritus from the Royal School of Library and Information Science (RSLIS), Copenhagen, Denmark. He can be reached at peter.ingwersen<at>hum.ku.dk