Digital Democracy

The Speakers commission has made a call for evidence on:

• The role of technology in helping Parliament and other agencies to scrutinise the work of government
• The role of technology in helping citizens to scrutinise the Government and the work of Parliament
• The nature and format of information and data about Parliament and government that is published online

The nifty thing is they’ll except blog posts, so here goes. Apologies for the showing-off bio at the start but I want to state my credentials.

TLDR (excutive summary)

Data on parliament should be published in a model that could apply to any democratic organisation, not a custom vocabulary. Make all information available via a website as well as machine-readable data. Government should design its processes so that data available to citizens is the heart of internal processes of parliament and the civil service, not an afterthought. All entities involved in democracy (people, places, events etc.) should have an official URI.

About me

About me and our team; my name is Christopher Gutteridge, I work for the Innovation & Development team at the University of Southampton, we are the techies who try to implement new technologies. I work closely with the Web & Internet Science research group headed by Professor Sir Nigel Shadbolt. I launched the University open data service for which we recieved the Times Higher Education award for outstanding ICT inniative of the year 2012. More recently I have launched data.ac.uk which aims to provide a hub for open data in UK academia and to provide services to provide more value from open data. Our most notable success has been with equipment data, where we combine open data on research equipment and facilities to provide a combined dataset and search with the aim of improving collaboration and reuse and avoiding unnecisary purchase of capital items. At the time of writing we are collecting open data from 25 universities and research institutes.

I’m focussing on how we would use digital systems to describe the structure and process of government, and not on the publishing of the many sets of statistics and other documents that government uses to make decisions.

A standard non-UK-specific vocabulary

First of all, I think it’s important to recognise that while a few aspects of UK government are unique, it has lots of structure in common with any other government. When describing elements of our government as open data, we should strive to do so in a way that allows one tool to understand data from different sources. For example, a committee in the UK Government can be described in the same way as a committee in a local council, a university student union, or a charity commitee in Kenya. By using generic, reusuable terms to describe each entity, we encourage the development of software that understands this data, both as a consumer and producer.

This will create an ecosystem of technology that can work with the data and software developed to work over the data about the committees in one organisation can be repurposed for another.

The counter case is if every government in the world developed its own bespoke way to describe it’s organisation and operations then every one will require custom software development and many of the benefits of open data will be lost. It would be like every country creating a different gauge of railway.

If the government is serious about “digitising democracy” then the language of this should be, where possible using existing standards, and if they don’t exist supporting the creation of international standards for communicating the process of governance. I know that The Open Data Institute (ODI) has some of the brightest and best in this area.

Once established standards exist then it is likely that low cost tools can and will be created to generate data in compatible formats to that used by parliament. It makes total sense to move towards mandating that all government processes are recorded as semanticly structured data. For example instead of recording the minutes of a council meeting as Word Documents they should be recorded in a database which stores each item and action and logs attendees and apologies. It’s important that such a system should be very easy to learn to use.

Open Data users and the other 99.99% of citizens

The problem with publishing things as data is that inaccessible to the majority of citizens without the mediation of a software tool. It is likely that a number of tools will be created by third parties, but it should be a principle that all information should be available to anybody with a web browser and basic (web & English*) literacy.

[*possibly “appropriate language literacy”, I don’t know if some records may be in Welsh]

This doesn’t have to be a complex tool, with fancy javascript graphing and so forth, but anybody should be able to set out to find a fact through the parliament website and if it exists, they should find it.

Eat your own dog food!

Wikipedia explaination of “Eating your own dog food”.

This is an important principle. The information used by the citizens should, as much as possible be the same information used by the civil service, councils and parliament. Rather than have a watered down view of the “real data”, the citizens should have access to the same basic information as that used to run the government. Obviously there will be parts that are privilaged or confidential, but the heart of the system should be the information available to all. Actually, the core description of our government and its processes should not be something we’re granted, it should be something we, the citizens, collectively own.

The easy solution would be to make openness and transparancy a “bolt on feature” onto the normal business of government, but that’s likely to be cut when money is tight or staff are overworked. I want to see something more radical, which is to say that the core of government should be assumed to be open and available to all, with some confidential processes on top of that. Obviously there will be some parts of government structure and meetings which are confidential for reasons of security, but these should be the exception, not the rule.

A principle here is that the majority of information communicated between departments should be open. Obviously excpetions exist for privacy and other good ethical or legal reasons, but generally I would like to see civil servants using the same data sources as the public. A good example here is how useful http://data.police.uk/ has been to the police’s own staff.

I’m aware that this would take many years to achieve even in part, but the principle is worth striving for as it will get better information for both citizens and government.

Authoritative Identifiers make everything better

One of the most important things the government should be doing is ensuring that the elements of government have authoratative and persistant identifiers. Schemes for identifying things should be independent of the department issuing the identifier as departments of government change, but that should not impact the ID schemes. Many things in the UK already have a clear ID such as charities, companies, car license plates, and that’s good. In addition I would like to see a formal idenfitication code for all the key parts of the democratic process:

Places: Regions, Council Wards, Parliamentary Constituencies

Events: Elections, Committee meetings, committee membership changes, debates, decisions, change of legislation

Legislation: Acts, sections, bills. http://www.legislation.gov.uk/ is doing pretty well on this front.

People: MPs and other elected officials should all be assigned a unique governent ID which should persist in identifying them in all public roles. However civil servants have more of a right to privacy and under some circumstances it may be better to identify them by their role than by the individual person.

A role: MP for Southampton Test, Chair of the Badger Goalpost committee, Minister for Cheese and Wine. These represent the position, not the person(s) filling the role. A role may be empty at some times. A role may have the dates it was created and disolved.

Person filling a Role: A relationship to an organisation or a responsibility, such as a membership (or chair) of a committee. It’s useful to assign a unique ID to the actual relationship between a person and role as it may have metadata such as a start date and end date, and whom they succeeded and by whom they were succeeded.

The reason it’s so valuable to create all these IDs is that the process of transparancy doesn’t stop at the boundries of government. It gives us, the citizens, the ability to unambiguously review and annote the business of government. It means that diverse groups can provide tools that augment the process. By using the authoratative IDs for each element of the democratic process these systems will be naturally compatible with each other and those of the civil service, once again the analogy of using the same gauge of railway is apt.

Oh, and those unique IDs should be URIs, of course. Although there’s the question of what happens to data.gov.uk URIs if Scotland goes its own way.

Thanks for the opportunity to comment on this. I’m excited to see what happens next.

I’d encourage friends and colleagues to chip in with comments if I’ve missed something important.