The Corporate Document Repository was launched in 1998 and currently contains more than 17000 FAO publications in digital format. The repository has been developed using Web technologies. The metadata is stored in an Oracle relational database together with all the information related to the workflow and the security rules, while the documents are saved in a file server in order to have a standard file system repository and a structured database engine for retrieval and workflow management. This architecture ensures stability, scalability and robustness.
State-of-the-art technology, using XML based import and export procedures, are used to facilitate the exchange of data with external partners and member countries and to improve the search functionality. For instance, FAO shares a common Information Finder with 16 centres of the Consultative Group on International Agricultural Research (CGIAR) [9], where documents from these Centres and from the Corporate Document Repository are presented together [11].
The first issue raised when the development of the Corporate Document Repository started was to find the most appropriate format for document storage, which can be retrievable and re-usable also in the future, regardless of which direction the technology development takes place. It was decided that the format of storage for the Corporate Document Repository should be the XML, because it is a non-proprietary format and allows separation of content from layout and hence can be repackaged and presented in different formats. Based on this decision, and considering that Microsoft Word is the most widely used authoring tool, Word templates have been created in FAO to ensure that styles are applied properly to a document and that the content is kept well separated from the layout. This facilitates the conversion process because it is possible to ‘map’ the styles used in Word to elements in XML or in HTML (see figure1).
Figure 1. Converting styles used in a Word Template into XML and HTML tags.
However, to use XML as the standard format for a digital library requires that all internal electronic publishing workflow is XML-based and not ‘paper-oriented’, as usually is the case. In an organization such as FAO - where there are close to three thousand persons working around the world - it is not easy to quickly change established rules and procedures. As of today, the MS word templates are used only by a small part of the organization, mostly where the production process is centralized, such as that for FAO’s governing body meetings. The Document Repository also accepts PDF files. In the trade off between having the most suitable formats and the necessity to disseminate FAO’s information, this has shown to be a necessary compromise.
Together with the Corporate Document Repository, a workflow system was proposed to collect metadata through the course of the document/publication production process. Some of these processes include content preparation, translation, editing, desktop publishing, printing, conversion for the Web. Customized version of this tool, which incorporates metadata gathering as well as workflow control, were offered to FAO technical Departments. The idea was to have authors or producers of documents to deliver the necessary administrative and descriptive metadata.
This experiment suggests that the complete decentralization of metadata production is not possible within FAO. It will not only bring down the timeliness and completeness of the resource acquired but also the quality of the metadata provided unless adequate incentives are provided to the authors.
