4. Benefits of the AGRIS AP

The AGRIS Application profile as an exchange format addresses the significant aspects of metadata interoperability. It:

- reuses content designators recommended by Dublin Core. It also uses elements that have been declared in other standards such as AGMES and AGLS

- uses XML and RDF syntax for coding. These syntaxes and widely applied for exchange and storage of information.

- is both human and machine readable.

This interoperability allow for various value-added services.

1. Exchange of agricultural information. Many partners, especially those from developed countries, are now contributing to the AGRIS central database using the AGRIS AP. The AGRIS AP enables exchange of data from systems that are using cataloguing and management rules other than those prescribed by the AGRIS guidelines. For example, the Netherlands AGRIS Resource centre uses a local format for management of its resource; however, it has submitted data using the AGRIS AP for the AGRIS central database.

2. Harvesting of metadata and associated content for Open Archives. The generic level of the AGRIS AP compliments the unqualified Dublin Core metadata set. This is the recommended format used for metadata harvesting in the Open Archives Initiative^[16]. The AGRIS AP facilitates exposure of the AGRIS content to the OAI systems, making it harvestable and available to a wider audience.

3. Possibility to access the actual resource from the Web. A recent study concluded that more and more resources are being retrospectively added to the Web. Based on this study, availability of good quality metadata allows for retrieval of the original resource, regardless of its actual location on the Web^[17].

Additionally, the XML format combined with XSL Transformations can allow for information to be used in many unique ways. This is potentially limitless. For example, a user could take a record encoded in XML, place it into a special XSL style sheet, and format a perfectly styled bibliographical reference in a word processing program.

4.1. AGRIS AP exchange layer avoids the need for a single format

Mapping all of the different formats (MARC21, UNIMARC, FINMARC, RUSMARC, JPMARC, AGRIN to name just a few) has proven to be practically impossible, especially since changes to formats occur with some regularity. As a result, there has been tremendous pressure for everyone to accept a single MARC format, but this also involves no less work to convert entire catalogues - except for those lucky few who already happen to use the chosen format.

Another suggestion is to use an “exchange layer” that would serve for exchanging all bibliographic information, whether it is in one of the versions of MARC ISO2709, or any other, perhaps relational database, structure. Therefore, if you could put in one field, you would receive the corresponding field for the other formats. [See Figure 4]

For example: in MARC21, information for the “Publisher” goes into field 260, subfield b; in UNIMARC, it goes into field 210, subfield c; in AGRIN3 (used by FAO), it is placed in field 402.

Figure 4: Mapping different formats to the AGRIS AP format

The advantage of the situation illustrated above is that each format needs only to create conversions to and from the exchange layer and avoid the need to create separate conversions to and from all the other formats. Any changes within one format would not result in reciprocal changes for everyone else, since the changes would affect only their own output/input to the exchange layer. If this worked for all fields and all formats, any record could be shared with any database. Of course, local editing would still be necessary in many cases, but no one would need to change anything within their own local databases. The AGRIS AP attempts to be this “exchange layer”.

4.2. Platform independent exchange to facilitate interoperability and reusability of information

The method to achieve simple metadata exchange is through the use common metadata standards and a common syntax, the XML format. The use of both these aspects in the AGRIS AP enhances the possibility for exchange avoiding many of the earlier problems related to systems. The result now is that the format or structure of any database is irrelevant for metadata exchange.

4.3. Supporting multilinguality

The xml:lang=“ ” attribute is used for elements for which it was considered necessary to know the language of its content. This extensibility enables multiple values of the specified field in any language. It was already mentioned that titles were indicated as being only in English, Spanish, French or Other. However, with the new specifications, title element can be in any language as long as the language is indicated using the xml:lang attribute.

On the structural side, there are different methods of inputting non-ASCII character sets, including Windows character sets, DOS, ISO, along with special bibliographical character sets, as exists in MARC-8^[18]. Unicode is now widely accepted as a standard and has been implemented in many systems; AGRIS-AP has chosen to accept UTF-8^[19] character encodings.

4.4. Maintain a level of quality in the collected information

Certain elements are critical for searching and necessitate the use of metadata schemes, thesauri and controlled lists. The use of these schemes also assures a level of consistency to be achieved in the collected information. The AGRIS AP recognises such elements and provides controlled vocabularies, lists and Schemes for these elements.

The following example shows how the AGRIS AP offers a means by which different controlled vocabularies and recommended schemes in agricultural sciences and technology could be used. The schemes that are used for the subject element are specific to the Agricultural Community. They provide the source information and thus the possibility of providing value-added searches.

The following subjects have all been given to the same item: the first is AGROVOC^[20], second is CABI Thesaurus^[21], and third is Library of Congress Subject Headings^[22].

With minor changes to the DTD, AGRIS could also accept records catalogued with keywords from other thesauri as well. Therefore, a record catalogued with keywords from the Bibliotheque National de France and the Swiss National Library would be validated with the modified AGRIS DTD.

Thus, all the individual terms can still be used for searching, but each one can also be “fine-tuned” to search only with terms from RAMEAU^[23] or AGROVOC, if desired.

4.5. New possibilities

The AGRIS AP’s structure and content plays a major role in enhanced searching and retrieval of agricultural documents. The XML structure, in which the metadata is encoded, significantly improves the usefulness of the data. AGRIS data, exported using the AGRIS AP, has been reused in different scenarios to achieve value added services.

Using the AGRIS AP has the following advantages that support searching and retrieval.

• The ability to search multiple databases with a single search has long been possible through the Z39.50 protocol^[24], but has not been widely implemented because of limitations on searching and record display. Using AGRIS-AP and web services^[25] offers possibilities that are far greater than ever before. Web services allow for superior searching capabilities, while XSL Transformations^[26] can sort records, eliminate duplicates and do further processing of the records received by the user.

• OpenURL^[27] and SICI^[28] are just a couple of the newer standards that have proven to be highly successful in assisting search and retrieval of journal articles. Metadata records must be interoperable with these standards and others that may arise.

• RSS newsfeeds^[29] have become very popular and are an excellent example of reusing XML applications that were originally designed for non-bibliographic uses. Installing an RSS feeder is relatively easy, and can be used to notify users of new items in their own, specific interests.