Interactions with Search Systems – Ryen White

Interactions with Search Systems – Ryen White

by | Aug 30, 2016 | Intranets, Reviews, Search

One of the common challenges I face in working with enterprise search teams is to persuade them that search is a dialogue between the user and the computer. It is not a ‘first strike’ application that is evaluated by the user on the basis of how many thousands of results are returned but on the way it enables them to work though search results interactively. For almost a decade Ryen White has been developing novel ways to support search interaction at Microsoft Research. The scale of his research can be gauged from his list of publications. Fortunately for the research community Ryen has now written Interactions with Search Systems, a book that synthesizes the work that has been carried out on search interaction by the teams at Microsoft and by many other research groups around the world.

Part 1 of this 500pp book provides an introduction, covering collecting and representing search interaction, modelling interests and intentions and summarising the many different approached to modelling information seeking behaviours. In Part 2 the theme is system support, including user interface design, learning and use and collaborative searching among many other topics. Search evaluation is becoming an increasingly important topic as user expectations continue to increase, and evaluation measures and methodologies are assessed in Part 3. The book concludes with a relatively short section on opportunities and challenges. The overall focus of the book is on web search, though in principle search interaction is of course core to enterprise search and library search applications..

There is no doubt that Ryen White has written a book that will become a master text in the way that Marti Hearst’s book on Search User Interfaces was in 2009. Interestingly both books are from Cambridge University Press. There is little that I missed in the book, though I would have liked to have seen more about interaction management in mobile search and I could not find anything about stopping rules and strategies. The bibliography runs to probably around 1600+ references spread over 85 pages. This bibliography is both a strength and a weakness. The strength is the depth of analysis that the author provides on this vast corpus of research. The weakness is that only academics and doctoral students with a well-funded institutional library are going to be able to read the cited papers and make use of author’s insights. It would be a challenging read at Master’s level to students with no prior knowledge of the literature.  Enterprise/web search managers are not going to benefit a great deal from the book. Of course that is not the market that the book was written for, but it is these managers who are looking for insights into search interaction for practical application, especially as open source solutions now provide increased scope for innovation in search application development. That is why Designing the Search Experience with its balance between research and best practice is such an exceptional book.

Once again I find myself disappointed about the quality of an index to a book. Adjacent entries for ‘cursor movements’ and ‘cursor tracking’, each with around sixteen page numbers is just one example. Entries with around 40 results, such as ‘engagement’, ‘histories’, ‘query formulation’ and ‘user studies’ are of no assistance to the reader at all. Why don’t publishers realise that few people will read a book like this from cover to cover, and even if they have done so will wish to dip into it time and time again for advice and inspiration on specific topics? I also found the use by CUP of centered sub-headings at the level unhelpful and if you are looking for screen shots there are virtually none.

As an authoritative book on search interaction research for information retrieval researchers then I have to commend and recommend this book. However its impact on educating search managers and helping them improve user interaction in search applications will, unfortunately, be much more limited.


Martin White