| Manifesto - Object-Oriented Legal: | = | |
|
|
| All legally-relevant documents - all contracts, statements of work, primary sources such as cases, regulations and statutes, and all electronic transactions under them - can be converted into text “objects”. | = | |
|
| Objects are nearly universally used in computer programming for the organization of complex information and authoring. Objects improve quality, focus work, and provide targets for continuous improvement (“many eyeballs”). Objects can perform a similar function in law. Software sourcing and deployment has rigorous methods and mature tools. The methods and tools of software development, collaboration and maintenance can be directly adopted for legal deployment when we express legal documents as objects. (Objects; versioning; Git; GitHub. Week 1) | = | |
|
| OTF | = |  |
| (From OpenTrustFabric.org - Modelling the EU Economy as an Ecosystem of Contracts.) | = | |
|
| Documents can be converted to text objects by breaking out the text components into a list of names and components, finding and using structural similarities among them, grouping these lists into higher-level lists and linking the lists. This creates a “graph” of linked “maps”, a simple, source form of semantic web. (Prose Objects; graph databases; semantic web.) | = | |
|
| These materials can be worked by lawyers using an "IDE" such as VS Code. The fit is very good. | = | |
|
| Expressing documents as objects makes them substantially “computable”, even if we defer any attempt to restate meanings embedded in documents as algorithms (Rules-as-Code, Legalese, Lexon,etc.). With a bit of usage of the system - particularly in expressing the chain of document events of real transactions - humans can see the patterns and pattern-recognition bots (AI, etc.) can make informed guesses about the likely consequences of various paths and propose paths to users. Waze for Law. Puppy deliveries by Amazon. Use of pattern-recognition rather than algorithms can reduce the reductionism of algorithmic specification and will find optimizations beyond those specified by human programmers. (Computable contracts. Contract Incompleteness.) | = | |
|
| An object-oriented approach does not exclude Rules-as-Code. Prose Objects are a complement to and facilitator of Rules-as-Code. The patterns found in the Prose Object paths can provide guidance and correction for fully algorithmic decision-making (Rules-as-Code). Conversely, when there is an adequate available algorithmic expression, the Prose Object can defer to the algorithmic expression, leaving performance and next steps to the algorithm. (See the ACTUS thought-piece.) The algorithmic bot can write consequent steps back into the collection of Prose Objects, creating a log and maintaining the “truth” and completeness of the collection. This relationship of Prose to Code is functionally identical to, though conceptually the mirror-image of, blockchain-based smart contracts that include text. (Accord Project, OpenLaw.io). It is also a complement to tightly-coupled systems of Code and Prose, such as Legalese and Lexon. Those systems can act as bots or as aids to humans, writing new Prose Objects into collections. Rules-as-Code and text are not and cannot be exclusive in a viable system.. (Rules-as-Code, smart-contracts, blockchain.) | = | |
|
| At the boundary of Prose Objects and Rules-as-Code is the systemic handling of text components. There is a soft boundary between enumerating a particular list of elements - for instance a list of information that is to be treated as confidential - and programming a machine to handle the list. This list-handling arises throughout the authoring of legal documents and can become a form of programming language for lawyers. We have a design for this, but it needs scrutiny, re-imagination and implementation. (Smart Prose Objects.) | = | |
|
| An object-oriented approach to text can be inclusive of and a complement to AI. AI can find patterns, identify new knowledge, avoid local maxima and errors associated with algorithms - hard-wired logic. The text of documents can improve on AI by setting forth principles and rules that are understandable to humans, providing theories and principles that go beyond mere recognition of patterns. In most transacting, the hand-off between machine and human can be easily accommodated - most transacting does not have the same debilitating time-urgency as self-driving. (Supervision by AI, AI ethics, HAL, HumanCompatible.ai.) | = | |
|
| Even modest use of “Prose Objects” can leverage law’s greatest technique, “codification” in the sense of the Napoleonic Code, the Uniform Commercial Code and the codes for data management that nations and enterprises are frantically attempting to negotiate now. Objects obviate the need for repetition, reinvention, rehashing. They provide targets for continuous improvement. They are tools, not rules, obviating the need for committee work and universal agreement, enabling small groups to move ahead as fast as they want, greatly accelerating convergence even as committee work is reduced. Coders and the open source community show how to use iteration to “codify” in the legal sense, bottom-up. (LegalCodification.md, UCC, Napoleonic Code.) | = | |
|
| Codification will allow us (lawyers) to systematize legal drafting in ways that were previously impossible or impractical. We will work in structure, starting from the humble NDA. It has and requires a system of “Persons” (identities), of Places (jurisdictions, delivery points, spatial reasoning), of functions (payments, notices, dispute resolution). We will be able to make these concrete and find the boundaries (interfaces, hand-off points) with technical systems such as databases, algorithms, identity management, fulfillment systems. | = | |
|
| Advanced techniques of linguistic analysis can identify patterns in legal language and legal logic. Most of this effort is done in the direction of converting unstructured text to structured form. This can be applied to text in Prose Objects to find patterns, and should be aided by the human-declared structuring of lists. More innovatively, it can also be applied in the opposite direction - linguistic patterns can provide authoring patterns - expressed in Prose Objects. This can be done at very simple, unsophisticated levels such as the common device of a list of conditions triggering an operational impact - guiding the author into an unambiguous pattern for drafting. (Grammatical components. Green Eggs.) | = | |
|
|
| Ted Wang challenge: | = | |
|
| Ironclad funding. | = | |
| https://www.theinformation.com/articles/mary-meekers-vc-firm-bond-invests-in-ironclad-at-nearly-1-billion-valuation | = | |
|
| IDE for Law: | = | |
| https://law.mit.edu/pub/whatwouldanintegrateddevelopmentenvironmentforlawlooklike/release/2 | = | |
|
|
| OpenTrustFabric.org | = | |
|
| Common-Legal-Platform: | = | |
| https://www.common-legal-platform.org/ | = | |
/Docs/G/NW-NDA/99/WiP/Manifesto.md