From my holiday viewing habits, Netflix now seems fairly (and justifiably) certain that I will be intrigued by content categories such as “Crime Comedies”, “Absurd Satires”, “Witty British TV Shows” and “Critically-Acclaimed Movies from the 1980s”. A recent article in The Atlantic tells us that Netflix has nearly 77,000 micro-genres and still deeper micro-tagging and rating of the underlying content. This data, combined with monitoring of viewing and clicking activity allows Netflix to, in the words of Netflix’s VP of Product Todd Yellen, put “the right title in front of the right person at the right time.”
The Netflix magic is achieved through a combination of human touch and machine-driven effort rolled into a regularly tweaked algorithm. In addition, while the Netflix genres go beyond the standard Hollywood categorization (Rom-com, buddy flicks, period-piece, family-friendly), the descriptors are kept to 50 characters or less.
Legal professionals may see a partial parallel to the subject matter indices and taxonomies available through commercial services like the Canadian Abridgment, the Halsbury’s collection or the Maritime Law Book key numbers. But for self-represented litigants, even those with access to these or similar services through courthouse libraries, there remains the challenge of knowing what category applies to their situation. Could a Netflix-like overlay to the underlying case law provide value through re-categorizing material into multiple and overlapping micro-categories that is more responsive to the manner in which SRLs understand their own issues?
Netflix demonstrates that with enough micro-tagging, you can create many paths to the same destination. Even if some of the paths seem odd or unlikely, this approach allows for more combinations than even the most talented group of human editors can conceive and manage. Of course, there may be a limit and things could be taken a wee bit too far. But even from within the depths of absurdity inspiration can rise. [For those that prefer their absurdity straight-up, The Atlantic built a highly entertaining Netflix Genre Generator that spits out such must-see gems as: “Dysfunctional-Family Psychological Space-Travel Slashers Set in Biblical Times About Marriage Directed By Herbert Ross”, ” Gangster Disney Thrillers Based on Bestsellers For Ages 0 to 4” and “Tearjerker Thrillers Based on Books From the 1940s About Cats”. ]
Fight-the-System Road Trip Cases Set in the Caribbean About Jurisdiction Rendered By the SCC
Do you recognize Club Resorts Ltd. v. Van Breda, 2012 SCC 17 in that title?
Maybe, maybe not.
However, if a client approached your firm about a dispute concerning activities that took place in another country, a quick reference to “forum non conveniens” or maybe “private international law” in your research tool of choice would lead you quickly to this decision. What would lead a self-represented litigant to this decision?
In her two-part series on Slaw in December describing the potential of providing legal services in a coaching model, Dr. Julie Macfarlane observed that “some SRLs have become fairly competent legal researchers” through extensive use of and hours spent on CanLII. While certainly true, it is likely equally true that many of those hours were spent dealing with the very steep learning curve of sifting through the irrelevant to find the (hopefully) relevant. Very few SRLs faced with taking the first step in addressing a problem that spans jurisdictions would begin their search with “forum non conveniens” to figure out where the action is properly brought.
Effective legal research always involves elements of advance planning, diligence and serendipity. How you frame the question can have a significant impact on where the research takes you and how you know when you are done. Working from the assumption that not all SRLs (or lawyers, for that matter) follow a focused process such as that offered here by Catherine Best of Boughton Law, it stands to reason that non-traditional methods of uncovering information and of finding end points (i.e., when you keep finding the same information and nothing new is coming up) are needed.
Research begins with a story not a topic or target
CanLII ran a one hour experiment on the afternoon of December 5th in which we provided a live-chat help feature. The objective was to improve our understanding of how users interacted with the site tools and to learn where we could make improvements to the overall user experience. Beyond that, the experiment gave us great insight in to research strategies of professional and novice alike.
Over that one hour, the chat window was presented to 1900 visitors and we carried our 176 live chat sessions. A sampling of the questions exposes some of the challenges that could be lessened with a Netflix-style approach to organizing case law content:
- Hi, I am looking for arbitration decisions involving dismissal due to absenteeism, and if possible where the absenteeism involves addiction.
- Any advice on searching for titles of equivalent legislation in various provinces? I’m looking for other provinces’ equivalent to the ‘local authorities election act” in Alberta.
- Can I appeal a default judgement after 3 months due to family death?
- I’m looking for a case about at what point the duty to preserve evidence arises.
- I am looking for a judgement or legislation, to help with the following question: in a divorce judgement, there was an order to pay child support. However, the date where the defendant needs to start paying that child support has not been determined. Is the date retroactive to the date of the judgment?
Except for those well-schooled in the use of Boolean operators, jumping in to these topics through keyword searching could be quite frustrating, or at least challenging and time consuming. But if the words, alone or combination, could be presumed to mean something against a list of micro-tags, then maybe the first step needn’t be as difficult.
The author of the Netflix piece in The Atlantic scraped the site to identify each and every uniquely-numbered genre id and then further parsed that information to decode Netflix’s genre vocabulary and grammar. The result was 242 tags (not including actor, director, creator names) across 10 top level categories including such things as 146 adjectives, 18 top-level genres and 12 temporal or geographic settings. (He put the full list in a google doc spreadsheet.) As most avid Netflix users are coming to realize, the micro-genres created from this are surprisingly good at bringing order to flood of available content. Is it reasonable to consider that applying this approach to court judgments could be useful as well?
Getting from idea to reality will be the challenge. For Netflix, this was an effort several years in the making and one that requires continual investment of human and technological resources. In the legal information sphere, we can see examples of the blending of professional effort and leading technologies in services like Lexis Advance and WestlawNext. As this pdf from Westlaw describing its 22-step case treatment and tagging process demonstrates, this is not a trivial exercise.
As noted above, Netflix wants to put the right movie in front of the right person at the right time. In seeking to promote access to justice through supporting self-represented litigants, it’s certainly worth exploring a variety of ways of putting the right case in front of the right person at the right time, even if we come at it through application of unusual and non-traditional methods.