Upgrading the GDDD towards the ideal pan-European GI marketing platform


Marcus Wandinger, MEGRIN Internet Services Manager

MEGRIN, 6 - 8, avenue Blaise Pascal, Cité Descartes,
Champs-sur-Marne, F-77455 Marne-la-Vallée Cedex 2, France
Telephone: +33 (0)1 64 15 32 40 Facsimile: +33 (0)1 64 15 32 19
Email: Marcus.Wandinger@megrin.org Web: http://www.megrin.org

Keywords
          metadata, pan-European datasets, mapping, geographic information.
1 Abstract
          MEGRIN is the "implementation and commercial branch" of CERCO, the forum of the Heads of the European National Mapping Agencies (NMA) counting 35 members. CERCO and MEGRIN believe that the European and global dimension of Geographic Information (GI) is increasingly becoming a key issue for the future development of our sector. Consequently any "national" GI initiative should be encouraged to consider the cross-border and international potential use of their products, and take into account the more global patterns that emerge from the regional, i.e. European initiatives.

          GDDD (Geographical Data Description Directory) is MEGRIN's metadata service describing the main GI products from NMAs of 20 European countries, a pioneer initiative in international implementation of the emerging CEN-TC287 standards on metadata. The service has been praised for its quality and consistency by all experts in the sector. However, standards practice and web technologies are quickly progressing, and the GDDD has to adjust to the new conditions.

          Within a new project named "La Clef" (sub-titled "unlocking the public sector GI"), the "old" GDDD will be greatly upgraded:

          The main purpose behind the project is the opening of the GI market across borders and across European cultural diversity. MEGRIN will develop and maintain the necessary infrastructure for the marketing and information exchange platform. Its success will ultimately depend on national organisation's willingness to collaborate in linking their national specifics and initiatives to that broader European platform.
2 Introduction
          As a commercial organisation MEGRIN (Multi-purpose European Ground Related Information Network) represents, and is funded by, a grouping of 19 European National Mapping Agencies (NMAs). It aims to bring a European dimension into NMAs national activities. MEGRIN aims to meet the increasing demand for pan-European data by improving international users access to national datasets. It does this by providing information about the digital data available now in 22 countries, and is creating harmonised pan-European datasets. Although MEGRIN comprises only 19 full members, its everyday partners are all European NMAs, the 30 plus CERCO members.

          The European geographical information market is still very nationally focused, and this means that users requiring pan-European datasets face a number of obstacles in acquiring the information they require, including:

          Recognising the growing market demand, MEGRIN has been established to focus on two areas of activity: providing metadata and creating new harmonised digital datasets. This paper focuses on MEGRIN's metadata activities.
3 Geographical Data Description Directory (GDDD)
3.1 Description of the GDDD
          MEGRIN's GDDD metadata service provides information about 250 digital datasets available from the NMAs of 22 countries of Europe. The GDDD was also the first pilot implementation of such a scale of the provisional European metadata standard CEN prENV 12657 of the CEN/TC287. Its current structure has existed since its introduction in November 1994. In 1996 it became more accessible as the widespread use of World Wide Web (www) browsers became commonplace.

          A central database is maintained in Paris, with NMAs providing updates as necessary. NMAs are provided with a software update tool which includes the current version of the GDDD database. They can update information about existing datasets or add new datasets before returning the media to MEGRIN, when the new data are then transferred to the main central database.
          Web pages are derived automatically from the database, and the information is made available free of charge to all Internet users. In a way GDDD can be described as an Internet gateway dedicated to European Geographic Information.

Information in the GDDD falls into the following categories:

The GDDD enables international users to locate datasets of interest, and contact the appropriate NMA. It has provided a good first step in the provision of geographic metadata.

Fig. 2: System overview over present GDDD: Central database at MEGRIN, which is updated with data collected from NMAs and from which static web pages are derived.

3.2 Access statistics
The Website log files are analysed and show that the GDDD attracts large international users, including:

          In the table above the columns indicate the areas of GDDD accessed i.e. the overview details of data, information relating to the commercial terms and technical issues, and details of the individual NMA organisations.

Although these statistics provide only a partial picture of usage, they do at least provide an indication that:

3.3 GDDD Strengths & weaknesses
Strengths:

          Although it provides users with these benefits, it is clear that the system needs to be upgraded to overcome some of its shortcomings, and to keep up with changes in technology and increasingly sophisticated user demands for metadata.

Weaknesses:

          The updating of the descriptions are felt to be time consuming by the NMAs because of the need to translate nationally owned metadata and the real or apparent duplication of efforts. A short term solution could be a simplification of the GDDD content, consistent with a shift of the service's focus from technical to marketing. But more efficient procedures based on automatic links and translation must be investigated - some of these issues will be developed further later in this paper.
4 La Clef: developing GDDD to a future one-stop geographic information shop
          Users require data which is ever easier to access, and would ideally like to use a 'one stop shop' to view, purchase, and be supplied with a wide range geographical data, including topographic data, remotely sensed imagery, geological and demographic data.

It is likely that such shops will be created incrementally. For the development of GDDD the next steps should be:
1. distributed metadata systems will initially link the various (existing) national databases of metadata,
2. the databases of metadata will then develop to enable on-line data sales.

Fig. 3: Planned structure of new MEGRIN metadata service as it will be realised within the La Clef project

4.1 Stage 1: Distributed systems
          Although metadata systems such as the GDDD have large numbers of users, they tend to be nationally-based and are often specialist services restricted to a specific discipline (and, maybe a specific provider, such as topographic data from the government). In their current form, they do not allow users to access the geographic information directly, but are directories which enable users to discover what exists. These need to be developed if geographic information is to reach as wide an audience as possible.

          In the European context, a distributed metadata system will face various challenges in integrating the various services, and issues such as different languages, application fields and standards will need to be overcome. This section aims to provide an overview of the issues which will be important in developing distributed systems.
          All metadata systems need to provide information which is easily accessible, reliable, and up-to-date.

The following will have significant impacts on the success of distributed systems:

          Within La Clef, these topics will be investigated by connecting at least GEIXS, the metadata service of EuroGeoSurvey, to the system. GEIXS is right now under development, a prototype is already accessible in the www.

          These topics will also be addressed in the Info2000 Project ESMI, co-funded by the European Commission DGXIII, where MEGRIN, a partner in the Consortium, will bring its practical experience in running a "pre-distributed-metadata system".

4.2 Stage 2: On-line data sales
Although the immediate aim is to connect metadata systems, there is growing demand for on-line access to data. To be truly useful, metadata services will need to be more closely integrated with the data purchasing process. Ultimately, users are interested in accessing actual geographic data. Metadata is the means to achieve the end, not an end in itself.

          On-line supply is currently hindered by various factors, including the wide range of geographical data transfer formats. If on-line supply is to be successful, the interoperability between different systems needs to be improved. Various investigations are underway including the OpenGIS(tm) and OGDI consortia, standardisation bodies, while software vendors are developing products which can change formats on the fly without the user needing to know. There are also the issues of data security, secure financial transactions and a range of marketing topics. For example, the existence of a network of agents and distributors acting for an NMA may make it only possible to introduce direct on-line sales slowly due to existing long term commercial contracts which must be honoured.

5 Issues to be dealt with when developing GDDD within La Clef project

5.1 Data model standardisation
          It is likely that implementing databases of metadata according to standards will not only provide more reliable information to users, but should also help the data providers by:

          Unfortunately, there are many metadata standards to choose from! These include the draft European standard from CEN/TC287. Internationally, ISO's TC211 is developing a standard which will ultimately replace any European standard. Finally, activities such as the Dublin Core or the CEO recommendations may also provide useful input, as they propose a very simple and limited set of metadata, easy to comply with, and easily open to several sectors of data, while CEN and ISO are GI focused.

5.2 Semantic standardisation
          The use of standard data structures will not necessarily ensure that the metadata is of consistent quality, completeness and accuracy. Indeed, the quality of the metadata may be more important than the data structure, since this is what users will see and use. Without reliable information, users are unlikely to use the service.
          All data providers and system users must have the same understanding of terms. Without this, the search results which are presented to users will be meaningless and/or incorrect. This means that all data providers must either attach standard keywords to their descriptions, or a thesaurus mechanism needs to be created to provide a common view of individual implementations. The task of harmonisation is likely to be complex, as there is considerable scope for differences to occur (for instance, between disciplines, within the same discipline in different countries, and even between different organisations in the same country).

          The need for semantic harmonisation is widespread: for example, there is not yet a standard for location references, with which to tie together the geocoded data and geographic information. Where standards do exist, they are not always implemented !

          MEGRIN's experience indicates that a reliable metadata service depends on the quality of the data descriptions. This is particularly significant in a multi-national environment. The initial data collection exercise raised problems because different organisations had used different keywords to describe datasets based on the same data model. This disagreement in choice of keywords was significant for complex themes, e.g. 'topography' and 'land cover', while straightforward keywords, e.g. 'road network' and 'railway network', were generally interpreted consistently. This clearly illustrated the importance of having the same understanding of the terminology.

5.3 Languages
To propose a real solution for semantics consistency, MEGRIN, the services of which aim at covering all Europe, cannot be satisfied with only one language. There are even more critical reasons to point out the inadequacy of a uniquely English language service:

5.4 Search queries
          Ideally, the user who is looking for data would like to locate the data he is interested in by pointing at the area covered by the project he is working on. Solving the main semantics and language issues should allow this possibility.

A text based query such as "Find objects = 'roads' in area = 'Benelux'" would be close to requests expressed in natural language. There are a number of ways to develop such a relatively user friendly search query. Knowledge about regions or geographical locations can be stored by the way of geographical keywords. Different relatively common solutions for look-up tables have to be considered, when on the pan-European level, each having its strengths and weaknesses according to specific approaches:

6 Partners in GDDD future
          MEGRIN is already involved in the ESMI project on distributed systems. Its main role will be focused on providing the project with multilingual "core-metadata" or "discovery-metadata", consistent with the most pertinent standard(s), and relevant to the GI community profile. Efforts will also be made to open the scope of the project to trans-sector applications, in order to be able to develop services across sector boundaries, including such areas as environment, transport, remote-sensing, geology, statistics, etc.

          Although MEGRIN is already currently working with a number of government organisations and representatives of the private sector, we are looking for formal and informal partnerships which are necessary to help facilitating the access to GI data for all range of users.

You may be able to help us for the development of the GDDD, where we foresee needs of partners in: