As a lawyer’s work is expressed in spoken and written words, these are the primary output options for any legal software, with printed output as the simplest one to begin with. This does not pose much of a technical issue and basically consists of automated filling of court forms and generating court bundles, the former being PDF forms and the latter any text format being capable of indexing, paginating and inserting evidential documents. PDF forms can be processed using the PDF Toolkit and as regards bundles, they can be e. g. Latex or LyX documents exported into PDF. Suitable legal citation systems are available on an open source basis, such as Bluebook in the U.S. and Oscola in the U.K., to name just the two largest legal markets. All the generating steps are taken in one process scripted in Python or any other language of one’s preference.
Generating the output documents being the essentials, the next step is automated advanced processing of the very content of the text so produced. And that is when legal technologists’ programming steps in: if the software can compose the chronology of events leading to litigation, make relevant legal conclusions itself and suggest the content of statements of case, then a lawyer user’s focus can shift from repetitive text editing to legal matters at the highest possible level.
Approaches to Codifying the Law
There have been various approaches how to enable computerized processing of the law. Some of them focus on converting human readable laws into machine readable form, others concentrate on semantic analysis of legal texts. So, projects like the Estrella Project, OWL2, CEN MetaLex, Oracle Policy Modeling or the Hammurabi Project have developed special programming languages that could serve as protocols for exchange of legal information. Projects using artificial intelligence such a IBM Watson, IP Soft Amelia, Ravn, on the other hand, use natural language processing methods to extract legal logic directly from sources of human laws. Systems already in practice are mostly based on legal rules hard coded in an existing programming language of their choice and update the embedded legal knowledge either by direct coding, for instance Stanford, A2J Author, and Modria or via a graphical user interface so that no coding is required to amend existing legal rules or add new ones, which is the method used by Neota, or by filling in web page forms as with Donotpay. These are, however, tied to their own expert system. Overall, none of the implementations has managed to create a large library of parsable law, only specialised collections or incomplete libraries and so a general standard is still missing.
Garret Wilson pointed out both procedure oriented and object oriented modeling concepts as a common denominator between lawmaking and software design. While software development is a rather recent discipline, in law the analogy has its roots in legal theories of ancient Antic philosophers. And while the law has developed these concepts into an elaborate system contained in myriads of legal texts written in what is familiarly entitled as “legalese”, the need for formal expression in computer science led to the development of the Unified Modeling Language as a means of systematizing the design of ever more complex computer programs.
The ORIG & OBITER ICO is now live. Join the cryptoparty at http://www.LegalMachinery.com/ or follow updates at https://twitter.com/psmlondon and subscribe to the ORIG Telegram channel at https://t.me/origico and the OBITER channel at https://t.me/obiterico
Note: For fully referenced text see the white paper at http://www.LegalMachinery.com/