Exploring maps for the mind

What do you type into the search bar (or chat box) when it's time to:

• Prepare for school or job applications?
• Scope out a trip somewhere nearby this weekend?
• Dive into ways to explore knowledge beyond the query-in-a-text-box?

Hey – these are not just simple queries! They represent looong journeys of discovery, with many searches and sources to stumble through. A Google search or chat with GPT can take you to starting point as long as you know where to go, but what if there isn't a specific destination in the first place?

---

Recently, I foraged through research literature (information retrieval, HCI) with the hope of finding crumbs to guide my own exploration of interfaces that support wandering, serendipity, freeform curiosity. I'm sharing this here as a collection of the best pickings!

---

Into the bushes: berrypicking and foraging

There's the classical model of information retrieval that influenced major search engines to focus on matching specific queries with specific documents (remember Google's "10 blue links"?). However, researchers have critiqued the inflexibility of this lookup-based model and proposed alternatives.

In 1989, Marcia Bates proposed a berrypicking model based on how people's needs evolve throughout the search process as they encounter new information. There is no one single set of best results; people collect fragments bit-by-bit through each step.

Evolving search needs (source: The design of browsing and berrypicking techniques for the online search interface)

In the late 1990s at XEROX PARC, Peter Pirolli and Stuart Card draw an analogy between how people look for information with animals looking for food. In their theory of information foraging, people try to move efficiently through patches of information by trying to lower the costs of moving between patches (or maximizing their gain from the current patch).

So searchers (and people, generally) are non-commital, munch-loving beings who tend to flit around until finding something that smells good enough to investigate – that is, until coming across another patch worth the leap.

In turns out there's a new-ish kid on the research block (est. 2005) building upon those models of exploratory information-seeking. Enter exploratory search!

How exploratory search is positioned and sized relative to related research desciplines

---

"Exploratory search makes us all pioneers and adventurers in a new world of information riches awaiting discovery."

---

Defining exploratory search

Exploratory search tasks are the realm of seeking nutritious knowledge to fuel action. They are en-route in the climb to higher levels of learning (like in Bloom's taxonomy). Unlike lookup tasks associated with careful queries and single answers, there is a sense of a journey that is open-ended and without a clearly defined goal.

These activities often happen in parallel, with rapid shifts between them. Lookup tasks can be embedded in learning or investigation. Source: Macc

Building upon Marchionini's widely cited definition, Ryen White and Resa Roth highlight (in Exploratory Search: Beyond the Query-Response Paradigm) the association with "wayfinding" behavior. Searchers are like travellers in an unfamiliar landscape with little clue about where to go, if even a destination exists. Just like in most wayfinding tasks, the searcher needs a way to conceptualize the entire space.

So...would it be alright to just call all this "browsing"? Not necessarily, although browsing behavior is involved in exploratory searches – and so are lookup tasks and queries aiming at something specific. In the entire journey of exploratory search, people use browsing and focused strategies in tandem.

White and Roth's overview of behaviors involved in exploratory search. Picture someone zipping up, down, and around between browsing and deeper-dives.

---

"Effective exploratory search systems will maintain a balance between analytical and browsing activities and support a symbiotic search relationship between searcher and system."

---

Features of exploratory search systems

While reviewing books and articles about exploratory search, I found examples of tools from the past. Retro UIs galore! But the ideas behind them are still relevant. Needs for browsing overviews, diving into details, and interactive filtering are also reflected in Ben Ben Schneiderman's influential articles on Direct Manipulation and Visual Information.

Examples of key features to support exploratory search:

Dynamic queries - Move incrementally through the information space by directly manipulating parameters that instantly update the results.

Film Finder (1994) plots films from a large database (positioned by production year and popularity) and allows users to use sliders as filter controls

Meaningful groups - Explore and filter results that are organized by categories (AKA facets) or clustered, possibly in hiearchies.

Relationship Browser (2007) is a research tool for exploring data sets. It shows facets/tags based on a query and a "tag cloud".

Grokker (an actual company from 2001 - 2009) was a search engine organizing results into topics and subtopics that users can zoom into.

Information visualization - Leveraging how fast and effortless it is to understand graphical information (compared to text). However, historic studies of visualizing result clusters have had mixed results about their appeal.

Hoeber et al. generated a visual map of the query space that helps users refine their Google searches.

---

"The pleasure in using these systems stems from the capacity to manipulate the object of interest directly and to generate multiple alternatives rapidly."

---

🚧 🚧 🚧 Writing in progress 🚧 🚧 🚧

In the meantime, sharing this list of Are.na directives that resonate very much :]