Harvard Law Library Readies Trove of Decisions for Digital Age by Erik Eckholm.
From the post:
Shelves of law books are an august symbol of legal practice, and no place, save the Library of Congress, can match the collection at Harvard’s Law School Library. Its trove includes nearly every state, federal, territorial and tribal judicial decision since colonial times — a priceless potential resource for everyone from legal scholars to defense lawyers trying to challenge a criminal conviction.

Now, in a digital-age sacrifice intended to serve grand intentions, the Harvard librarians are slicing off the spines of all but the rarest volumes and feeding some 40 million pages through a high-speed scanner. They are taking this once unthinkable step to create a complete, searchable database of American case law that will be offered free on the Internet, allowing instant retrieval of vital records that usually must be paid for.“Improving access to justice is a priority,” said Martha Minow, dean of Harvard Law School, explaining why Harvard has embarked on the project. “We feel an obligation and an opportunity here to open up our resources to the public.”
…While Harvard’s “Free the Law” project cannot put the lone defense lawyer or citizen on an equal footing with a deep-pocketed law firm, legal experts say, it can at least guarantee a floor of essential information. The project will also offer some sophisticated techniques for visualizing relations among cases and searching for themes.
Complete state results will become publicly available this fall for California and New York, and the entire library will be online in 2017, said Daniel Lewis, chief executive and co-founder of Ravel Law, a commercial start-up in California that has teamed up with Harvard Law for the project. The cases will be available at www.ravellaw.com. Ravel is paying millions of dollars to support the scanning. The cases will be accessible in a searchable format and, along with the texts, they will be presented with visual maps developed by the company, which graphically show the evolution through cases of a judicial concept and how each key decision is cited in others.
…
A very challenging dataset for capturing and mapping semantics!
If you think current legal language is confusing, strap on a couple of centuries of decisions plus legislation as the meaning of words and concepts morph.
Some people will search it as flatly as they do Google Ngrams and that will be reflected in the quality of their results.
Yet another dataset where sharing search trails with commentary would enrich the data with every visit. Less experienced searchers could follow the trails of more accomplished searchers.
Whether capturing and annotating search trails and other non-WestLaw/LexisNexis features will make it into user facing interfaces remains to be seen.
There is some truth to the Westlaw claim that “Core primary law is only the beginning…” but the more court data becomes available, the greater the chance for innovative tools.