Logic Language Law Computing

Today I attended a workshop organised by the Greater London Authority (GLA), which is the citywide government for London. The workshop was held at City Hall on the top floor where we had a splendid view over the Thames, of Tower Bridge, and the Tower of London.
The GLA is in the process of scoping a datastore for information London. The objective is to begin to encourage development of “government 2.0” using open government data along the lines of what has been done in San Franscisco in the US (see an article on DataSF and a post by San Francisco Mayor Gavin Newsom). The principle idea is that by putting data of public interest into the public domain, the government can provide the basis for development of applications and services for the government, business community, and public. For example, using police data, one can generate crime maps.
At the GLA meeting, the objective was to meet with the developer community to get ideas and feedback on what and how the data should be released as well as how best to encourage applications in the near future.
Clearly, the GLA meeting is along the lines of what is happening elsewhere in the UK government (see Digital Engagement at the Cabinet Office, the Office of Public Sector Information, and The Stationery Office).
There were some 70 participants at the meeting, and we can look forward to further information coming from the organisers at the GLA. Some very useful suggestions where made about where to get further information such as the Technology Strategy Board which supports technology development in the UK.
Among the topics of discussion where:

• What sort of data should be released and in what form? There were those who wanted it raw and those who wanted it structured. Likely releasing it both forms will occur.
• How to get licensing for the data? There are a host of difficult issues here, as most of the data is owned or copyrighted by a range of organisations, each of whom wants to control the flow of information, profit from it, or has concerns about security/liability. Moreover, the government contracts information service providers, which process the data, may have some legal claim. Such providers may be required to make their data open.
• How would the data be used? There were many suggestions about data reuse and mash up, mostly along the lines of existing applications such as mapping data to physical maps in order to get ideas about what is happening where in neighborhoods, transportation assistance, information access in a local area, and so on.
• Who would develop the applications and how would development be funded? Clearly there is an issue about funding, but some of the ways around it are to leverage funding between academic, government, and business communities.
• Who to consult about applications? A range of parties might be consulted about what they would find useful, from the person on the street, to members of service organisations (police, licensing, etc), to higher level government organisations. Alternative, the GLA or similar organisations might develop applications which they thought would be useful, then provide them to the public. Here, the focus would be on small, manageable, pilot projects to show proof of concept.

The discussion was very nicely organised and led — several large tables around a circle, several large monitors, several boards for writing, an MC who kept things moving along, and a good overall atmosphere. However, missing were a list of participants, contact information, and a short (three sentence) statement of interest; hopefully, all this will appear soon.
This is all very interesting and exciting, for things are just beginning to happen. However, I have some concerns and realise I have a somewhat different focus.

• There were too few participants from government services and academia. This is also reflected in the gap between the technology and the data, since it is highly unclear who is developing what for whom and what purpose? It is hard to get a handle on how data should be served without some sense of goals. Nonetheless, likely there will be another meet-up at which this more substantive discussion will happen.
• There was only passing mention of ontologies, annotation, information extraction, and the semantic web. The absence of semantic web concepts suggests that “reasoning” and complex information management is not high on the agenda. This is consistent with the family of application ideas (graphs, maps, local information). While I was told that there were ontologies via OPSI, this is not what I understood from John Sheridan in my recent discussion, so I will be eager to see exactly what this is.
• Similarly, there was some discussion about whether there should be or could be standards and schemas for the data. These are always compatible (make them both available), but I see standards are essential for any communication across agencies and localities. There are drawbacks to standards development, but the issue will arise sooner or later in any case.
• There is, as yet, no the pitch. In other words, what is the incentive for anyone to make their data available or for organisations to otherwise cooperate with this endeavour? The only mentions of incentive were government obligation, but this is perhaps the most heavy handed way to make headway. Rather, there should be positive incentives. In addition, the eGovernment agenda should be pushed (e.g. transparency, support for government, participation, efficiency, cost reduction, consistency….).
• There was little discussion of exactly which technologies would be used though RDF/XML and REST were mentioned. These are generic and widespread; is the real hangup right now data access, or is there some technological issue? If I wanted to know what I should need to know to program and provide a simple service, what would I have to know and do?
• Despite the widespread interest in government 2.0, there is little vertical/horizontal integration or communication among the interested parties. There is not, apparently, a coherent website or ‘state of the art’ article with links to the relevant data/functionalities/support organisations.
• There was no over-arching conception of design or context for applications. Likely some sort of ‘apps’ or plugins framework will emerge so that, for example, a local council would build none of its own applications or services, but these would be provided as plugins by independent providers, yet given a consistent style and structure.
• Though there are claims that there have been consultations about government 2.0 with the various interested parties, there is no clear presentation of the results of those consultations. A ‘brain-storming’ site would be very useful.
• It is unclear to me the extent to which the participants have the political/social context in mind. While we were hosted by the GLA and discussed GLA data, the opportunities, limitations, requirements, and objectives of government seem to have entirely overlooked. For example, government is successful (not always, but often) with making and monitoring standards for the public good; as elsewhere, why not here? The requirements of government information provision are different than for commercial provision, especially since the government provides goods and services that would not otherwise be profitable. The government does consult with the public and interested parties in making policy, but in some cases it is crucial that government lead and direct developments; the government is not simply another commerical provider of goods and services, driven by consumer interests; the government has a legislative role. Keeping this in mind may change the sorts of proposals that come out from open GLA data
• There were several discussions about why and how government data should be published. The main points ought to be developed, discussed further, and summarised. Yet, it ought also be pointed out that there is, in the UK, an abundance of information that the government holds about individuals; it is unclear how a ‘firewall’ to protect and promote civil liberties will be set up and maintained; privacy and rights are in fact rather weak in the UK. For example, the NHS is state funded and one might argue certain matters are in the public interest, so open information issues may arise here: will we have ‘disease’ mashups such as there are for broken lamp-posts, but in this case for drug addicts, HIV carriers, swine flu, etc?
• I am particularly interested in legal reasoning, but this is not something on the agenda with respect to this data.

In any case, there is much of interest here and much to look forward to.
Cheers,
Adam Wyner
Copyright © 2009 Adam Wyner