Mark Me Up
Structuring the Supreme Court's judgment in
A collaborative and decidely lo-fi experiment to explore effective ways to structure judgments.
On 2 October 2019 the Legal Education Foundation (TLEF) published an extensive report authored by Dr Natalie Byrom setting out a "29-point plan for tackling ‘digital exclusion’, and ensuring the government’s £1bn court reform programme delivers access to justice for all court users."
Dr Byrom's recommendations address a wide range of issues that are bound up in the proliferation of justice-related data, including mechanisms for the collection of data about court users' vulnerabilities; monitoring outcomes for represented and unrepresented litigants and ethical controls governing the use of data collected throughout the justice process.
The report is also significant because it mounts a clear set of recommendations addressing the ways in which judgments should be handled as a unit of data in the justice system.
In her report, at paragraph 4.47, Dr Byrom observes:
There was a high degree of consensus across stakeholder groups regarding the urgent need to understand and reform the system for providing public access to primary legal data i.e. judgments.
The need for effective and meaningful access to the decisions of judges flows from the fact that the UK is a common law jurisdiction governed by the rule of law. Dr Byrom at para 4.48,
In a common law system, providing access to judgments is crucial in order for individuals to understand their legal position and initiate or defend legal claims. It is commonly argued, and uncontroversial to state that ‘the rule of law requires that the public have access to the law—to all the binding norms and authoritative interpretations of them’. The provision of access to judgments is also critical to the delivery of principles of open justice: stakeholders argued that in the context of a reformed system where physical hearings were to be reserved only for those cases that could not otherwise be resolved the vital importance of providing access to judgments and capitalising on opportunities for reformed systems to deliver what Professor Richard Susskind has described as enhanced ‘informational openness’ is underscored.
Turning to address the status quo, Dr Byrom says, at para 4.49:
The current system for disseminating judgments is opaque and complex—stakeholders reported that ‘the bulk of the task of publishing judgments has been delegated to nonpublic institutions, some of which have a profit-driven motive’.
The report then goes on the identify three main problems areas:
- Comprehensiveness of judgment coverage on BAILII
- Comprehensibility of the judgments themselves
- The format in which the judgments are made available
It is the third of these problem areas that Mark Me Up has been devised to explore.
Imposing structure on judgments
At para 4.54, Dr Byrom makes the following important finding:
Currently the sites providing free access to judgments publish them in non-machinereadable, unstructured format. Stakeholders from lawtech and for-profit publishers reported that the failure to publish judgments in a structured, machine-readable format that defines particular elements e.g. party names, decision etc creates barriers to entry for lawtech start-ups, particularly those operating in the not-for-profit sector, as the costs associated with preparing data are prohibitive. Stakeholders argued that judgments should be made available in an open, machine-readable format (such as XML) using a consistent and open XML standard. A common set of meta-data fields should be applied and XML schema used should be capable of distinguishing the applicative/procedural part of the judgment from the mere representation of the document. Stakeholders reported that the development of open data standards for publishing judgments would align England and Wales with developments in Europe and the USA, helping to ensure our ongoing competitiveness.
Structure and why it's important
There is a widespread and justified belief that judgments, when viewed together as a collection of documents, represent an important source of data for a range of social and economically-driven applications.
To do things with data, we need to impose a consistent system of structure upon it. We intuitively apply structure to all sorts of data we produce without thinking about it. If you've ever made a table in Excel, with labelled rows and columns, you're applying structure to data. If you've ever written a document in Word with headings, subheadings and paragraph numbers, you're applying structure to data.
If you have ever been given a spreadsheet produced by someone else and found it littered with random empty cells, columns out of alignment with the rest of the data and inconsistent use of cell formatting, you've seen first-hand the problems caused by poorly structured data.
The structures we apply to data directly affects our intepretation of the data and the degree to which further downstream operations can be applied to it.
Modern judgments take a narrative form with numbered paragraphs. They are written in Word and text heavy. This makes them challenging to work with from the offing even when there is some structure to them. However, more often than not, as a source of data, judgments are made available in two formats that make them even harder to work with: RTF (rich-text format) and PDF.
Para 4.54 of Dr Byrom's report advocates for a new approach based on a consistenly applied and maintained machine-readable format (XML is being the cited example, but there are others). We agree wholeheartledly.
What does structure look like?
When we talk about machine-readable formats we are not talking about formats of information that allow a machine to understand meaning (that's a long way off). Instead, we're talking about formats that allows a machine dice and slice the data in a way that makes it easy for us to extract things from the data.
This webpage has been written in a machine-readable format: HTML. HTML breaks up the content of this page into smaller units. We can write instructions in other languages to do things to those smaller units (e.g. make this bold).
The task at hand
The imposition of structure is a vital prerequisite for better open access to the decisions of judges. It is our assessment that HMCTS (to whom Dr Byrom's recommendations are ultimately directed) are in the best possible position to lead the way on making this a reality.
The community has an obligation to support HMCTS in this regard by providing it with tangible examples of what we would useful. Mark Me Up is about making those tangible examples.
Mark up the Supreme Court's decision in R (Miller) v Prime Minister
We have published an open Google doc consisting of the text of the Supreme Court's recent judgment in the Miller case.
Your objective is to annotate the document to indicate the ways how you think it should be structured in a machine-readable way.
You do not need to worry about adding XML tags that are perfectly formed (although, you can if you want to). But you're encourage to highlight the text and add comments describing your ideas.
The only thing we ask you to do is refrain from deleting anything in the document!
The goal here is not to achieve perfection, but to get the ball rolling and put pixels on the screen.