WRITTEN ON November 7th, 2005 BY Luke Razzell AND STORED IN Transformational Government
The British Government is making all the right noises about their new IT strategy: the strategy paper emphasises the importance of both integrating diverse information resources on one hand, and yet providing flexible, adaptive access to those resources to each individual user and community on the other. Also to be applauded is the decision of the responsible cabinet minister, Jim Murphy, to participate as an author here at Ideal Government. Welcome to the Blogosphere, Jim (it’s writing, Jim, but not as we know it ; ).
However, there remain at least two flies in the Government’s technological ointment:
1) XML’s brittleness
XML, the data model around which e-Government’s technological strategy is centred, is inherently inflexible. This effectively necessitates a centralised, top-down approach to information design across the e-Gov network, as XML-structured data can only be exchanged between parties who agree in advance how the data will be described.
Government has stated twin aims of building IT solutions that are “designed around the citizen or the business”—with their diverse needs and viewpoints—and to “move to a shared services culture [...] in information and infrastructure”. To achieve both these aims it may well be necessary to move beyond an dependence on standards-based data models, such as XML, for the exchange of structured data, in favour of new approaches which allow for more flexible integration of diverse structured data sources.
In the physical world, where we (mostly!) manage to sustain interaction and communication within relationships, communities and societies despite the diversity of characters and view points within those social groupings, by bridging intuitively across our overlapping yet different perspectives (you say “turquoise”, I say “light-blue”, we both understand the general concept).
A virtual environment that allowed us each, as individuals or communitites, to describe information in our own way, yet at the same time to tap into others’ differently-described information, would enable us to interact online with one another in ways that are familiarly human, rather than according to the requirements of the machines that facilitate our interactions. This is a vision that is surely worth pursuing!
Disclaimer: My company, i-together is developing a data model and technology for data exchange between diversely-described structured data sources.
2) The identity-centric data exchange conundrum
A second, and related challenge for Transormational Government is identity-centric data management. “Identity-centric data management” refers to the management of data exchange between individuals on one hand and communities, business and government on the other in a way that offers multi-party self-determination and security across distributed networks (such as the e-Gov network). This is a problem that is still very much in the process of being solved by industry. The SAML/SOAP-based Liberty Alliance framework, for example, has been shown to have a conspicuous security, and hence privacy weakness from the individual user’s point of view.
The Microsoft-led “identity metasystem” initiative, founded on the WS-* (Web Services) protocol stack, aims to provide a comprehensive framework for identity-centric data management that comprehensively solves the problems of multi-party privacy and self-determination. However, discussion over how to actually implement such a metasystem is ongoing: Web Services are based upon XML, and I feel that the standards-dependent rigidity of the XML data model (discussed above) is a very significant inhibiting factor to the progress of the discussion. Also, SOAP-based Web Services are highly complex to deploy, so any network solution that depends upon them runs the risk of excluding the majority of small-scale web developers from that network.
Other data exchange technologies that offer promise with regard to abstraction from a specific data model, and multi-party self-determination, privacy and security include the secure data exchange technologies of Credentica and the open YADIS data-exchange protocol.
Credentica are working with Microsoft to ensure that the former’s technology can be deployed in conjunction with the latter’s InfoCard personal data management solution, within the metasystem framework.
YADIS eschews the complexities of SOAP-based Web Services in favour of the simpler, widely-adopted and more transparent REST-ful approach (RSS is an example of a REST-ful web service), and looks set to achieve strong take up across social web applications like blogs, wikis, photo-sharing services and so forth. However, as a very young protocol, YADIS is currently insufficiently-specified to facilitate the industrial-strength data-exchange required by e-Gov.
Transforming aspirations into understanding—through considered dialogue
It’s clear, then, that we still have a long way to go with the development of technologies that enable the realisation of the Government’s vision for public services. A catalyst to progress will quite possibly prove to be the solution of the problem of exchanging data across diversely-described data stores: until we do solve this problem, we will be unable to embrace in our solutions, at a technological level, each user’s expression of their unique, human identity. But whether or not that turns out to be the case, I leave you with the thought that, in sustaining an inclusive, meritocratic and distributed conversation about the subject across the blogosphere, we are already taking big strides in the right direction. : )
[Cross-posted at weaverluke.]
7 Responses to “Transformational Government—in need of a transformational data model?”
i wanted to comment.
your choice of words about XML seems unfortunate to be honest. especially given that you recommend RSS, which is an XML-based format! and also call RSS restful – what exactly do you mean by that?
i am all for critique, but i just couldnt parse this screed effectively.
Thanks for the comments.
John:
I agree that common standards for data description are beneficial with regards to the situations you describe. However, e-Gov is not just about inter-governmental or market-regulated data exchange: it is also about data exchange between diverse *people*, as individuals and communities of practice, on one hand, and business and government on the other. And people function most effectively and happily when they are able to relate to others according to their own point of view, or ontology.
For this reason, in my view, we need a data model that integrates *both* exact and approximate, highly-structured, semi-structured and unstructured data descriptions.
James:
I’m not sure if you mean something more general than your specific criticisms when you say my “choice of words about XML seems unfortunate”, but let me answer your individual points anyway.
RSS is indeed an XML format, but its widespread uptake is regarded by most industry commentators as stemming from its extreme simplicity—it is an XML “microformat”. RSS is highly effective at purveying *semi-structured*, public information (such as that found on this blog!).
Unfortunately, the very key advantage of RSS—its simplicity—also limits it radically as a potential vehicle for the kind of rich-semantic structured data that e-Gov deals with. Hence my perception of a need for non-XML approaches to exchanging *structured* data.
RSS is REST-ful in the sense that each feed can be identified by a standard URL. I do not need to use SOAP to access an RSS feed, just type the URL into my feed reader. That is REST.
You are right that YADIS is young but it is an open project in which anybody can participate (see yadis.org) If anybody has specific experience in the area of e-gov requirements, by all means, please feel free to get involved and make sure we are doing the right thing.
Ok now we’re cooking with gas.
I understand the microformat argument, and it makes sense to make it explicit. if you’re going to crititize particular XML schema, or stacks based on those schema, then I can understand you better.
RSS is far from a universal panacea for workflow, and there are standards issues to deal with their too.
its important to define terms clearly in order to support the distributed dialogue you talk to in your blog. A statement that XML is inherently inflexible is problematic, in that regard.
Well, I guess I do see with XML an inevitable trade-off between semantic depth (i.e. richness of structure) and re-useability.
If you’re interested, I wrote an essay a while back whose third section discusses this structuredness/flexibility trade-off issue with current tech in the context of RSS, tagging, APIs (i.e. XML) and (keyword) web search. The abstract:
http://www.i-together.net/weaverluke/2005/03/phase-behaviour-in-human-communities.html
Regarding the suitability or otherwise of XML, I think it is about horses for courses. There are some basic elements of e-government info exchange that lend themselves to pre-determined standards based on XML-like structures for interoperability. But Luke’s point is really that most meaningful e-government interactions are with or between *people*, and these require a greater tolerance of ambiguity, less pre-determined structure and more flexibility in general. Perhaps we need a new approach for these.
People have been banging on about the (theoretical) importance of data standards for years and years, but there are not that many compelling examples of these standards in action. For example, in the e-learning field during the 90′s everybody was obsessed with the need for common standards to support interop, but few actually pursued interop for purely commercial reasons. In retrospect, althoguh theoetically logical, a lot fo this was a waste of time. It may be that in some cases a simpler approach is also better, learning from the way that microformats such as RSS have spread like wildfire.
So, I think there’s something in what Luke says and if we are ever to rescue e-gov from the deadening clutches of Microsoft, IBM, EDS and Capita contracts then we might have a chance of achieving it ;-P
On XML brittleness …
I agree that XML is “strict” and requires that government and business parties exchanging data do so on the basis of agreed meta-data. But, such inter-governmental or market-regulated data exchange should follow strict protocol anyway either (a) in a free state with limited government or (b) in markets cleared by efficient prices.
In government, a data exchange not strictly limited to lawful or conventional data-sets is called interrogation. In commerce an exchange of goods based on a combination of some commodity or money price and “further consideration” is discriminatory and monopolistic.
So, I see “requires XML” or the “inhibiting effect” of the “the standards-dependent rigidity of the XML data model” as a virtue of the MS approach.