|5. System and design issues (Statistics Netherlands)||Statistics Netherlands||7. Lessons learned (Statistics Netherlands)|
6.1 Overview of roles and responsibilities
Apart from the general default roles that Documentum provides (such as Author and Owner) DSC distinguishes
- Metadata Administrator (responsible for, e.g., the maintenance of office-wide metadata standards, such as the classification of statistical topics)
- Metadata Designer (responsible for e.g., the correct design of data sets and for supplying correct variable definitions)
- User (allowed browsing the metadata base and extracting statistical data according to his need)
6.3 Training and knowledge management
All metadata will be stored by the owners of the metadata with assistance of well- trained metadata experts.
6.4 Partnerships and cooperation
6.5 Other issues
- Maintaining good quality of metadata is considered a serious issue. ISO 11179 presents some guidelines and rules that are adopted by SN. It is a challenge though to formulate quality guidelines in such a way that metadata is interpreted the same way now as it will be at a later moment. Also, it is non-trivial to formulate such guidelines for metadata that is intended for (various categories) of non-experts.
- Acceptance of the Data Service Centre by the statistical divisions. The Data Service Centre introduces a new way of working with an initialadditional workload for statisticians. This benefits the potential users of the data primarily and not so much their producers. This means that producers must be convinced that they are users (of other data) as well, so that they will see the overall benefit of spending time and money to describe their data in a user-friendly way.
- Additional Workload. The normal mode of operation is that metadata is produced primarily during the design phase of a statistic, of which the activities involved are usually part of a design project. For various reasons, storing and correctly describing existing data sets (SN's statistical history) with the use of the DSC is usually not organized in a project context however, which puts an additional workload on those departments that are engaged mainly in statistical production.