|Target Users||Publishers, book distributors, booksellers, libraries|
|Nature of the initiative||Private|
|Contact||Antonio María Ávila
Executive Director FGEE
DILVE (Distribuidor de Información del Libro Español en Venta) is a service provided by the Federación de Gremios de Editores de España (FGEE – Federation of Spanish Publishers Guilds).
The ultimate goal of this service is to provide rich, updated and complete metadata about Spanish books to worldwide actors of the book value chain. This information is directly managed in DILVE by the corresponding publishers using a wide set of tools and services that caters to the varying needs and technical skills of the different publishing houses. DILVE’s core is based on ONIX for Books (see www.editeur.org/8/ONIX).
DILVE allows publishers to manage their catalogues in the ‘cloud’ using different mechanisms, ranging from FTP uploads to interactive data entry, in different formats, from ONIX (both versions 2.1 and 3.0 of the standard are handled) down to free-form CSV files. DILVE includes tools for assessing the quality and completeness of the data.
Publishers keep the property of all metadata they upload and they explicitly allow its distribution by DILVE.
A wide range of access mechanisms and format is also available for users when extracting information from DILVE. Files can be obtained in ONIX (2.1 and 3.0), in user-defined CSV formats, as well as other Spanish book industry standards. These files can be sent via mail or FTP, or stored at the user’s ‘DILVE mailbox’ within the platform. DILVE includes a simple yet powerful HTTP-based API that allows integration in automatic processes.
Data extractions can be programmed, with different book selection criteria and allowing the definition of automatic periodic incremental data extractions.
Providing up-to-date and rich metadata to the book supply chain is a requirement for publishers.
Different studies show that the quality and completeness of the metadata has a direct impact on book sales (see http://www.nielsenbookdata.co.uk/uploads/3971_Nielsen_Metadata_white_paper_A4(1).pdf). The need for adequate metadata is growing as a consequence of the increasing importance on the Internet in book sales. In online stores, the only representation of a book is through its metadata. Buying decisions a taken solely based on the provided metadata. The inclusion of cover images, adequate book descriptions, 'look inside' facilities, links to similar books or works by the same author, etc. are crucial for selling the book. This applies both to e-books and printed books, since access to the physical book is only possible after the buying decision has been taken.
However, maintaining and distributing rich metadata is technically complex and requires skills and facilities which are beyond the possibilities of most publishers. The large numbers of users that need this information, coupled with a wide assortment of required formats and communication channels makes this task almost impossible, even for large firms. On the other hand, for most actors of the book chain consolidating metadata coming from several hundred different publishers would be an insurmountable hurdle.
A solution is needed that will help publishers in their task of maintaining and distributing metadata and, on the other side, that will help consumers of metadata in their task of gathering and consolidating metadata originated in diverse sources and formats.
If all publishers have to deal with all users needing their metadata and all users have to deal with all publishers, the complexity of their interrelations grows exponentially. The same problem would exist with physical books if publishers were required to deal individually with each point-of-sale.
The basic concept of DILVE is that of a 'distributor'. Publishers only need to deal with one entity, as well as metadata users. A central distributor redices an " n x n" problem to "1 x 1". Each actor in the value chain has to deal with only one entity for all their metadata needs. DILVE acts as a bulk data provider. Users are supposed to keep local databases in their premises, either with the totality of titles managed bay DILVE, or with subsets based on the particular interests of each user. DILVE provides the functionality needed to automatically keep these local databases updated and synchronized with the publisher's catalogues.
DILVE includes a number of input interfaces that allows data feeds by the following mechanisms: interactive input form (this used bay a large number of small publishing hoses), manual upload of files via an online dialogue, or via FTP.
The following formats can be used for upload: ONIX for Books (both versions 2.1 and 3.0 are allowed, with their long and short tags variations) and CSV files with a user-defined field map (the user can choose from a repertoire of more than 150 fields).
All information is checked for data integrity and converted to ONIX when stored. Thus, although DILVE is based in ONIX, publishers are not forced to use this format to manage their catalogues.
On the output side, the user can choose from the following mechanims for accessing infrmation: mail, FTP feeds to their own systems, DILVE's 'mailbox' (this is a per-user FTP directory tree in DILVE's server), DILVE's API (this is a simple yet powerful language-agnostic HTTP-based API that allows integration in any programming environtment) and in SINLI format (this is a local standard for data interchanges in the Spanish book sector, it defines both the interchange mechanics and file formats).
The user can choose between ONIX (either 2.1 or 3.0, with long or short tags, and different character encodings), CSV files with user-defined maps (each user can define an ilimited number of CVS maps from the 150+ available fields) or SINLI format.
Also, each user can define an illimited number of 'extraction programs' specifying the record selection criteria, the format of the resulting file, the sending mechanism and the periodicity (ranging from twice a day to monthly). It is possible to select only new records or records that have changed since the last run of a data extraction.
The role of technology
DILVE uses the following technologies:
- Client-Server architecture with an Oracle-based (Standard One license) database;
- Apache + JBOSS infrastructure;
- Java based programming;
- Lucene indexation for text rich searches;
- ONIX-based core;
- HTTP-based Application Programming Interface
The platform is in constant revision in order to adapt it to the fast changing needs of the book supply chain. An average of over one monthly code revision has been implemented in the last 24 months.
The first version of DILVE was open to the public on October 2006 and it is still an ongoing project. Tight integration with the Spanish ISBN Agency (also managed by the FGEE) is scheduled within the next 18 months.
DILVE is now used by more than one thousand publishers who maintain metadata of over than 540.000 titles. Among other users, over 320 distributors and more than 1.500 booksellers and online store use DILVE.
Since 2006, DILVE has served over 618 million bibliographical records (including new entries and updates), in different formats, to its users.
DILVE has helped Spanish publishers to dramatically improve the quality of their metadata and has made possible its effortless distribution to all actors of the book supply chain.
At present only Spanish publishers (or non-Spanish publishers with a strong presence in the Spanish market) can handle their catalogues through DILVE. However, this is not a technical restriction. DILVE source code was used as the base for a similar service (CANAL) implemented in Brazil for by the Camara Brasileira do Livro (http://www.canal.org.br/dilvenoejb/dilveweb/dilve_english.jsp)