The past year has seen a lot of innovation in the design of legal research. In the blink of an eye (in this market anyway), we’ve seen the launch of WestlawNext, Lexis Advance, data visualization and Forecite from Fastcase, and new mobile apps for all three systems, to name just a few. It’s been a long time since there’s been that much innovation in legal research, and all at the same time. Sometimes it feels like we’re living in a fast-moving future, with very advanced and modern technology to address our work.
But it just isn’t true. One day not far from now, we’ll look back and laugh at how cool we thought these systems were. This is the Mesozoic Era of research, especially in user interface and design, which has really changed very little from its original search-results-document heterodoxy from 40 years ago. Our research systems (even the best ones) are dinosaur-primitive, and we’re really just beginning to see what the future might look like.
A Conversation About the Future
On Friday, Dec. 10 at 3 p.m. EDT, I’ll be participating in a discussion about that future with Rich Leiter and Roger Skalbeck, who are hosting a Law Librarians Conversation podcast about the future of user experience in legal research, with Jason Wilson, Vice President of Jones McClure Publishing; and Tom Boone, Reference Librarian at Loyola Law School. [For more information or to register, click here.]
In an effort to hit the ground running in our conversation, we’ve each agreed to write some initial thoughts in blog form. You can find the group’s thoughts here (and if you care about information systems or software design, these should be required reading — some terrific thinking in these posts):
Rich Leiter: Reflections on the End of the World Wide Web and the Future of the Internet as an Information/Service Resource
Roger Skalbeck: Mobile Legal Research: Do we NEED an app for that?
Jason Wilson: Law Librarian Conversations: The Future of Interface Design
Tom Boone: From interface to extinction: law school librarians beyond Thunderdome
During the live discussion, co-hosts Richard Leiter and Roger Skalbeck will consider three broad topics relating to interface for legal research tools:
1. Native App vs. Browser-based tools
2. What interface innovations can we borrow from other applications?
3. Modularity & Interoperability: Will we ever have modular legal research tools? (I’m thinking APIs, a plug-in architecture or interoperable products)
For some background on this conversation, I’d like to discuss some assumptions that make legal research needlessly simplistic, in the hopes that we can challenge them and make research and discovery more interesting.
Discovery Beyond Search
Inherent in the traditional notion of legal research is “search” — the idea that we will type keywords or phrases into a box and get back a list of results. And this works today in roughly the same way that it always has — keyword or natural language search, bound by Boolean operators (a syntax so counterintuitive that we have to train law students to forget everything they know about more intuitive search), producing a long list of text results with highlighted keywords that we sift through, one by one.
The process has been busted for years, and we’re just not doing enough to fix it.
Traditional legal research really only does two things: filter once (or twice, if you “focus” and search within results) and sort once. They filter the entire universe of legal precedent with Boolean keyword parameters, then they sort the result (for caselaw, traditionally the highest court first, most recent case to the oldest). The poor researcher then sweeps the Augean stables by reading documents one by one. This is true not just for legal research, but for Web search as well.
Starting in 2003, we did some things at Fastcase to change up the order a bit — allowing users to customize how results are sorted to bring the best ones to the top, to reduce the search costs for finding the most germane documents — but at core this is really little more than a filter and a sort. (We’re flattered that you can now find some of these sorting innovations in the new versions of Westlaw and Lexis – for a great discussion of design principles of all three products, check out Amy Eaton’s great summary (CRIV Sheet Nov. 2010 at p.6) of our panel last summer at AALL in Denver.)
To really innovate in legal research, we’ve got to move beyond simply filtering and sorting, and we’re just starting to see some really inspiring innovations that might show us the way.
The law isn’t a series of flat documents — law cites to law, and these citations create a beautiful information architecture that is woefully underutilized. We’ve done a little work here, integrating citation analysis into results on Fastcase (akin to a Shepards column in search results that allows users to sort most cited documents to the top).
But we’ve got to move beyond hyperlinked citations. There is so much more to do here — including topical (semantic, if you must) links, geographic links, linked entities (judges, parties, corporations), or even procedural connections between documents (linking, for example, denials of motions for summary judgment, or final orders).
Text search results can be very powerful, but they’re far from the only way to find germane documents from a filtered subset of a database. Charts, graphs, clusters, histograms — there are hundreds of ways to display results that tell the story better than search results. We’ve done a little work here, creating interactive timeline maps of search results. Here, too, there is so much work to be done.
The guys at Computational Legal Studies, IBM’s Many Eyes Team, and Wolfram Research have done some gorgeous work, and there are well-established leaders like Hans Rosling and Edward Tufte in the field. Data visualization tells us so much truth that text obscures, and as Jeff Jonas has pointed out, when you enable linked data and data visualization, you don’t need the old search paradigm. Data finds data.
Finally, there is contextual discovery — finding things by reference to nearby navigation elements. This has lots of promise, in part because we already do this some today. When browsing through hierarchical texts such as an outline of statutes, we’re really just relying on context (the outline provides the context). Google reports that it’s working on contextual discovery — search without searching — that proactively will suggest things based on context (such as a restaurant’s menu on your mobile browser when you walk in).
Forecite on Fastcase is a similar, non-search contextual discovery feature. When a Fastcase user is looking at search results, we treat the set of search results as a context, then review all of the cases cited by docs in the search result. If that turns up a frequently-cited opinion that isn’t in the search result, Forecite proactively suggests it to the user.
The network knows more than the researcher or any individual document — we can use novel user interface tools to expose knowledge just waiting to surface. David McCandless, author of Information is Beautiful, has an elegant way to describe this new frontier of discovery beyond search:
“Data is the new soil. For me it feels like a fertile creative medium, and data visualizations and infographics are flowers blooming from this medium. If you look [at data sets] directly, they’re just a bunch of numbers and disconnected facts. But if you start working with them and looking at them in a certain way, interesting things can be revealed.”
If data is the new soil, my team is working in the garden. We’re working on about three dozen UI innovations right now to improve the transparency, customization, visualization, consistency, and elegance of Fastcase. These are still early times in search, but there is innovation everywhere around us — it’s a really exciting time to be working on these issues. Should be a fun podcast conversation, too — Rich, Roger, Jason, and Tom are some of the best thinkers about design, and it’s always fun to see the future through their eyes.