We asked Sue Feldman to write a guest editorial for this issue, drawing from her 2012 book The Answer Machine, published by Morgan & Claypool. The genesis of the book was her article, “The Answer Machine: Are We There Yet?” published in the January/February 2011 issue of Searcher (Vol. 19, No. 1, pp. 18–27). This editorial is abstracted from section 3.2 of The Answer Machine.
Web search is 20 years old this year. The human drive to seek information, coupled with the pervasiveness of personal computers and growing access to the internet, created a perfect environment for the explosion of the web and web search engines. The web unleashed a torrent of information, and invited anyone with a computer and an internet connection to become a searcher.
End users search differently than intermediaries do. Unlike intermediaries, end users know what their ultimate information needs are—and they recognize the right answer when they see it. However, they may have difficulty crafting a suitable query. End users are also less likely to create long, well-formed queries. In contrast, intermediaries are proficient at query formulation and in helping end users pinpoint what they are seeking. But intermediaries can only craft queries based on what they are told (or suspect). End users are less hung up on precision and more willing to browse (but less willing to plow through pages of results).
To accommodate end users, online systems needed to move from expecting a finely crafted query to supporting information exploration. Online system designers had to incorporate what they knew about good human search practices into a new kind of retrieval software to index and retrieve this vast, quickly changing volume of information to accommodate the unpredictable searching skills of millions of new users. Neither careful curation of the collection nor professional intermediaries could handle the expectations of end users for quick access to current information. Technologies that could bake professional searching expertise into the search system were required. In short, although search engines could analyze text and create an index of terms to search against, they also needed to do what professional searchers had been doing instinctively.
In the past decade, the worlds of personal and workplace information seeking have drawn together. We have seen a metamorphosis in the way consumers and workers find and use information. People expect to do their own searching either at home or in the workplace. The lines between work and home have blurred as people tote around tablets and smartphones so that they are constantly in touch with their families, their friends, and their work. Information seeking now happens in all kinds of contexts and on all types of access devices and formats.
Between 1995 and 2005, we saw a shift from information professionals to end users as the primary users of online information systems. Now, valuable information is often freely available on the web, and end user searching has won.
Lost in the shuffle, temporarily, are two key services that professional intermediaries provided: helping end users define what they are looking for, and understanding the provenance of the located information so that trustworthy information is provided to the end user. Both of these functions are difficult to automate effectively, but newer information applications are certainly taking a stab at both.
The consequences of the shift from professional information seeking to everyone as a searcher have been mixed. Certainly, access to a wealth of information has changed how we live, enabling us to accomplish simple tasks such as finding directions, making dinner reservations, or looking up stray facts no matter where we are.
On the other hand, distinguishing between “good” and questionable information is not easy, so scams abound, and decisions are made on incomplete, old, or wrong information, creating information disasters. It’s hard for most people to judge what’s authoritative and what’s not.
Today, we are on the cusp of yet another revolution in information access, this one moving us to pervasive information that is integrated within the flow of whatever else we are doing. We can decide to see a movie, find reviews, view a trailer, buy tickets, and get directions—all while everyone involved is walking to a coffee shop, sitting around the breakfast table, or on the phone. Pervasive information moves us away from information as a separate destination and back to what is more natural—exchanging information when we need it to complete a task.
We now have 40 years of experience in designing online information systems. Search technology is no longer experimental. Analytics for text is well-developed. Many of the challenges are clear:
Searchers have difficulty formulating the right question.
Voluminous collections of information are hard to navigate and analyze.
Language is ambiguous and hard to interpret for literal-minded computers.
Interfaces are still experimental—their effectiveness varies with the user, the user’s task or context, and the user’s background and expectations.
Images and speech are hard to retrieve because they are even more subjective, variable, and difficult to represent than text.
Speech (voice) and gestures are natural carriers of information and need to be incorporated into interaction design.
These challenges will shape the direction for search technology research today and for the next decade. Search is only one starting place among many in the quest for information. The emphasis now is on designs that allow users to browse, explore, and discover within a specific work environment based on a process or task.
These newer, search-based applications, which I call InfoApps, address information-centered processes. Certainly they include search, but they may also add content management and publishing, collaboration tools, marketing tools, sales tools, business intelligence tools, or predictive modeling. InfoApps are designed around a specific process such as e-discovery, marketing, or web publishing. They create comfortable information work environments that integrate multiple technologies. InfoApps have tuned the search and analytics technologies for a particular process: Find an answer, find the best answer, find all possible answers, or even, “What’s here?” or “What should I know about that I didn’t think to ask?”
InfoApps and other steps toward the future will make people’s lives simpler, hiding complex technologies so that users can get their work done without having to be technical wizards.