The Integrated Public Service Vocabulary: A confusion of poly-hierarchies.

By Jan Wyllie.

There’s that old Irish joke about asking directions with the punch line: “You wouldn’t want to be starting from here.”

It is a perfect description of the events that led to the launch of the Integrated Public Service Vocabulary (IPSV) in the UK in April 2005. The IPSV is a product of what one participant called an “impossible” task: the merging of three conflicting vocabularies into a single taxonomy covering the whole world of national and local government interest and action, as well as ‘community information’.

The IPSV aims to include all concepts in the compact but all embracing Government Category List (GCL), the newer, less-developed Local Government Category List (LGCL) and the much bigger thesaurus called SeamlessUK – all created by different public-sector bodies and their consultants. The headings at the top levels of the IPSV broadly follow the GCL, but have been adapted to accommodate LGCL structures.

The job would have been hard enough if the three vocabularies had been created for similar purposes. However, the GCL was designed as an umbrella structure to facilitate broad categorisation of resources across the public sector. The LGCL was focussed on local authorities, in the context of organising websites, while the seamlessUK taxonomy was for indexing resources at a much greater level of specificity where many of the resources were in databases where full text was not available to the search engine.

To make matters worse, the process of taxonomy integration is reported to have been driven more by political expediency and turf protection than by the principles and processes of good taxonomy design and constructive collaboration (see Taxonomies: Frameworks for Corporate Knowledge, Ark Group). Participants also say that directives from the Office of the Deputy Prime Minister (ODPM) – whose Best Value Performance Indicator (BVPI 157) for e-government has driven the whole undertaking – were less than helpful to local government about exactly where and when to use the new taxonomy.

Unsurprisingly, the resulting IPSV is a compromise, if not a fudge. Now, compromising and fudging may be useful attributes in the realm of politics, but they are certainly not virtues in the design of taxonomies where improved clarity and common understanding is the goal.

The latest twist in the tale is that the IPSV is not the last word in vocabularies. The IPSV was designed as a common classification and indexing tool, not a website and portal navigation tool. So a group of local councils in the UK (Brent and the West London Alliance) collaborated to produce the Local Government Navigation List (LGNL), originally know as the Poly-Hierarchical (navigation) System (PHS), a simpler web-navigation taxonomy. LGNL is also designed to be used with local-council workflow systems. At the moment, using the IPSV is mandatory, while groups of local councils can decide for themselves whether to use the LGNL for web and intranet navigation.

The top levels of LGNL and IPSV are similar, but also significantly different. Although the purposes of the IPSV and the LGNL are different, these nearly-the-same-but-not-quite classifications are unlikely to contribute to clearer thinking or presentation by government. For instance, in the IPSV, searching for “woodland management” brings up the preferred term “Forestry” under Environment – but not under Business – and is associated with non-preferred terms, “Tree felling”, “Tree planting” (plantations) and “Timber production”. Searching for “woodland management” in the LGNL yields a different preferred term, “Forest and woodland management”, under Environment & Planning/Conservation, Environment & Planning/Countryside and under Environment & Planning/Parks and open spaces, none of which appears at all in the IPSV. While IPSV has related terms, the LGNL does not.

Both systems suffer from the problem that they often fail to produce logically meaningfully related lists of terms, especially at the lowest levels. For example, “Forest and woodland management” in the LGNL is in a list that includes “Grass cutting”, “Events” and “Boats and boat rides”. Although the LGNL sticks to the three-level hierarchy, despite its tendency to mix up apples and oranges, the IPSV is inconsistent in the number of levels of terms employed. Neither produces lists of items of equivalent importance at each level in the hierarchy.

Already to get around this inconsistency, Lincolnshire Council has to make up its own terms at the lower levels of the IPSV (about 40 so far) just to import data into the fields of its CRM system. Robert Sprigge of Advance Telematics who works on the system, said: “It varies a lot more than two or three levels. Some areas have five levels not just two or three. An unfortunate but common situation is where there are a few child nodes but not enough to cover real-life scenarios.”

The IPSV and the LGNL represent two different perspectives on the world of government. How these two perspectives are reconciled will be a major challenge when it comes to delivering outcomes and meeting targets. However, a far greater challenge will be showing that the elimination of locally defined classification and filing systems improves performance in the three broad purposes of taxonomies – information retrieval and sharing, collaboration, and intelligence gathering and analysis.

There is a danger that taking away people’s power to manage their own information in their own way could significantly reduce government performance at all levels. At least individuals and small groups have their own ways of remembering their own private taxonomies and storage schemes for which they alone are responsible. If the centralised system does not work, the power of individual memory – “I made this place for it” – will be crippled, lost in the system.

The decision to merge the government vocabularies into the IPSV while at the same time making it mandatory across the public sector appears to have been taken in haste. It is questionable whether the issues of implementation, including the huge cost of training in both time and money, have been sufficiently considered. The public sector employees responsible will have to understand how to work with a thesaurus, according to the ISO2788 standard. If the indexing is not completed to standard across government, then retrieval will become unreliable.

It is also possible that the use of any kind of common indexing will improve performance in the areas of knowledge sharing, collaboration and intelligence. To get around the shortcomings, people could negotiate and agree conventions; sub-group classification schemas (eg, for CRM applications) will form; and life will go on gradually improving.

Either way, the exclusive adoption of the poly-hierarchical form of taxonomy misses out on the potential benefits of a faceted approach. A difficulty with the poly-hierarchy is its assumption that the goal is to enable people to see information from their own point of view, rather than giving it the best description possible. So “Woodland management” is classified under “Parks and open spaces” for the park wardens, under “Conservation” for environmentalists and under “Countryside” for the farmers? Do they all want to know the same things about woodland management? What happens to woodland management businesses? How many other points of view should be catered for? Where does woodland food production (berries, venison, squirrel meat etc) go?

The goal of a multi-faceted taxonomy is to give people a common understanding of how information is organised, a superior table of contents, rather than a superior alphabetical index. In a multi-faceted taxonomy, “Woodland management” would combine the term “woodland” from a subject hierarchy (physical things in the world under, say, Land) and the term “management” from, say, an activities hierarchy. If the item was about “conservation”, the word would come from another facet or hierarchy, say, Issues. And so on. The power of multifaceted taxonomies to describe sources is so much greater than poly-hierarchical taxonomies.

Jonathan Engel of InfoArk, who worked on the original LGCL, points out that government list users are already taking the first step towards a multifaceted approach, if, in addition to the IPSV, they also tag items with terms from other official lists, such as the Local Government Audience List (LGAL), the Business Category List (LGBCL) and the Local Government Interaction List (LGIL), thereby giving different perspectives on (or facets of) the same item.

A multi-faceted taxonomy could incorporate all government activity and interests in a coherent and intuitive framework – a language, not a list. It is now time for the government to commission the design of this meta-language in order to build on the hoped for outcomes of this first phase of taxonomy application. It could then be called the IPSL (Integrated Public Service Language), which public servants, politicians and even the public could learn. The hope must be that the IPSV is a step in this direction.

Jan Wyllie is the author of Taxonomies: Frameworks for Corporate Knowledge, which is published by Ark Group.

<BOXOUT>

Overview of UK government taxonomy family members

e-Government Interoperability Framework (e-GIF) and Technical Standards Catalogue (TSC) http://www.govtalk.gov.uk/schemasstandards/egif.asp
e-Government Metadata Standard (e-GMS) V3.0 http://www.govtalk.gov.uk/schemasstandards/metadata_document.asp?docnum=872
e-Government Metadata Standard (e-GMS) for intranets V1.0, 8 Nov 2004. http://www.govtalk.gov.uk/schemasstandards/metadata_document.asp?docnum=909
e-Government Metadata Standard (e-GMS) for websites http://www.govtalk.gov.uk/schemasstandards/metadata_document.asp?docnum=806
Integrated Public Sector Vocabulary (IPSV) http://www.esd.org.uk/standards/ipsv/
Government Category List (GCL) http://www.govtalk.gov.uk/schemasstandards/gcl.asp
Priority outcomes: explanatory notes for practitioners Version 1.0 http://www.idea.gov.uk/transformation/downloads/2community.pdf
Integrated Public Service Vocabulary (IPSV) http://www.esd.org.uk/standards/ipsv/viewer/
Local Government Category List (LGCL) http://www.esd.org.uk/standards/lgcl/viewer/
Local Government Navigation List (LGNL) http://www.esd.org.uk/standards/lgnl/viewer/
Local Government Services List (LGSL) http://www.esd.org.uk/standards/lgsl/viewer/
List of Government Lists http://www.esd.org.uk/standards/