This article is based on the experiences in the VLIRED Cuba programme with interoperability between the metadata of di erent platforms and the reusability of the data in di erent information silos. VLIRED Cuba, a VLIR-UOS network programme, started with implementing speci c platforms for educational, library and information management and research output, resp. Moodle, ABCD, DSpace and VIVO. The nal goal is to create an integrated research, education & information network. The authors de ned the basic concepts metadata, authority systems and interoperability based on a literature study. \Interoperability is enabling information that originates in one context to be used in another in ways that are as highly automated as possible". But the overall success of such data structures being applied by a wide variety of metadata producers will depend on the quality of the metadata content they provide. VLIRED Cuba chooses for open source software to develop the network: ABCD for the library, DSpace and VIVO for research information and Moodle for e-learning. But the platforms uses di erent metadata formats, respectively MARC, Quali ed Dublin Core, a group of ontologies and LOM. To make the di erent systems interoperable the content should be described in standardized way. The project focused therefore on the key elements to standardize: people, organization and keywords. A basic principle is that these entities should be de ned by a unique value. This has to be integrated in the di erent metadata formats. Authority systems will deliver standard content: for people ORCID, Cuban ID or local VIVO id's - for keywords controlled vocabularies like CCS, AGROVOC and MESH. Finally VLIRED Cuba is developing tools like EsFacil and EsFacil Authority to create standardized metadata in di erent platforms. EsFacil Authority is an authority system that can manage all type of authority lists and make then searchable through a REST API. EsFacil will collect and enrich publication metadata and make it available for systems with di erent metadata format.
1
This article is focusing on the problems of interoperability between the metadata
of di erent platforms and the reusability of the data in the di erent information
silos. The use of standards in metadata and authority control systems is
central in this approach. The authors are involved in VLIRED Cuba, a VLIR-UOS
network programme, which is creating a virtual research and educational
environment in Cuba using open source software solutions. The development of the
VLIRED Cuba is a cooperation of six Cuban universities (Camaguey, Holguin,
Pinar del Rio, UCI, UCLV + University of Oriente) and the Flemish universities
coordinated by VLIR-UOS with the support of the Ministry of Higher Education
of Cuba and other local stakeholders [
VLIRED Cuba started with implementing speci c platforms for educational, library and information management and research output, resp. Moodle, ABCD, DSpace and VIVO. The nal goal is to create an integrated research, education and information network. It is not the goal of VLIRED Cuba to create one global platform. Services should be available in the platforms were they are relevant. Teachers and students will use primarily the e-learning environment, but when relevant will nd their information from the other platforms. Researchers will use the research platform(s) and get the relevant information there as well.
It is therefore essential to standardize the approach to metadata. Metadata
is a key part of the information infrastructure necessary to help create order
in the chaos of the Web, infusing description, classi cation, and organization to
help create more useful stores of information [
This part focuses on the concepts of metadata, authority system and interoperability as de ned in literature.
There are many de nitions of metadata in the existing literature. The
basic de nition of metadata is data about data. Erik Duval added: Metadata are
basically descriptive data. As such, metadata are at the heart of more general
developments in the area of digital libraries. Basic metadata elements indicate the
title, author, year of publication and similar simple bibliographic data. Richer
metadata structures also cover technical features, copyright properties,
annotations and so on. The purpose of metadata is \to facilitate search, evaluation,
acquisition, and use" of resources [
New descriptive metadata schemes such as Dublin Core, o er data structures for interoperability that are a promise for bibliographic control and access issues. But the overall success of such data structures being applied by a wide variety of metadata producers - with an equally wide variety of motivations, concerns and skills - will depend on the quality of the metadata content they provide. Authority control is central to create quality metadata.
Authority control is the process of selecting one form of a name and recording
it, its alternatives, and the data sources used in the process. It is an important
tool that boosts recall and precision in the retrieval of information resources.
It provides consistency in the form of access points used to identify persons,
families, corporate bodies, and subject headings [
Interoperability is enabling information that originates in one context to be
used in another in ways that are as highly automated as possible [
Finally, the semantic web developments and linked data are a way to
implement interoperability in a highly automated way. Linked Data is simply about
using the Web to create typed links between data from di erent sources. These
may be as diverse as databases maintained by two organisations in di erent
geographical locations, or simply heterogeneous systems within one
organisation that, historically, have not easily interoperated at the data level.
Technically, Linked Data refers to data published on the Web in such a way that it is
machine-readable, its meaning is explicitly de ned, it is linked to other external
data sets, and can in turn be linked to from external data sets [
The use of unique identi ers for metadata elements is essential in the approach of VLIRED Cuba in the development of interoperability between the di erent information silos and platforms. An analysis of the di erent platforms and their use of metadata is the following step. 3
The network platforms and their metadata formats.
VLIRED Cuba has implemented the following platforms: ABCD for library management, DSpace and VIVO for research output management and Moodle as e-learning platform.
1. ABCD 3.0
The higher education authorities in Cuba decided nearly 8 years ago to use open source software ABCD as the library system in every university. Since 2011 the UCI was called by the MES to support the implementation of ABCD in the Cuban universities. Since 2013, VLIRED Cuba, worked on the development of ABCD 3.0, a JAVA version based J-ISIS. A national development team was created, with members from all the network universities and coordinated by UCI and as special collaborators some specialists from INFOMED and IDICT. ABCD 3.0 was released in January 2017 and installed in the network and all MES universities during the same year.
As every traditional library system, the main metadata format of ABCD
3.0 is MARC. MARC was the rst computer based standard for bibliographic
data (1960s) developed Library of Congress. It is since the 70's an international
standard in the library world. There are several versions of MARC in use around
the world, the most predominant being MARC 21. Still it is based on technology
from the 60's and therefore outdated. MARC is, at its heart, a data format built
to contain catalogue records; bibliographic items are described via the catalogue
records rather than directly via the structured MARC data [
The structure of MARC elds includes indicators and sub eld codes as shown in Figure 2.
Other interesting elds with possibilities of authority control and implementing interoperability are : 110 - Main Entry{Corporate Name and 650 Subject Added Entry.
2. DSpace as repository software
All universities of the network are using version 6., but customized to a VLIRED Cuba format. Version. They are availability on Internet or through the national Cuban university network. (UHOLM, UC, UCLV and UPR).
DSpace basic metadata format is still Quali ed Dublin Core (QDC), even when this format has been replaced by DCTerms. It is possible to add other metadata formats but it is only possible following the element-quali er structure. Implementing a richer application format like BIBO or BIBFRAME is therefore not realistic.
Metadata granularity is therefore limited. For example QDC has only
identi er.citation and DCTerms bibliographicCitation as element. Source, volume,
pages and/or article id cannot be described separately, not to mention the
description of conferences for proceedings papers. Di erent extensions have been
realized to get a more granular metadata without a standardized approach [
DSpace added from version 1.6 the interesting option of an identi er for an
element. Already it is possible to link author names to ORCID. This functionality
has been used by AgriOcean DSpace to store the unique id of terms out of ASFA
and AGROVOC thesauri [
With DSpace, VLIRED Cuba is confronted with two major challenges to resolve the problem of metadata quality and interoperability with other platforms: metadata granularity and implementation of authority systems. Therefore a group was set up to develop a speci c submission module, EsFacil including metadata auto-extraction functionalities and the management of authority systems.
3. VIVO: a research information system
Research output is larger than publications. To manage the broader scope of research information VLIRED Cuba evaluated the possibilities of using a CRIS system, nally choosing to use VIVO, an open source software. Its semantic web approach was another reason for its selection. The universities of the project are experimenting with the software, speci cally how to ingest and manage existing content.
For its (meta)data management VIVO uses a collection of ontologies to represent the research of an organization. The Integrated Semantic Framework ontology modules for VIVO (the VIVO-ISF ontology) provide a set of types (classes) and relationships (properties) to represent researchers and the full context in which they work. VIVO integrates existing ontologies in its Integrated Semantic Framework. To describe publications it uses for example the following ontologies bibo, foaf, vcard, skos and vivo.
VIVO follows the principles of Linked Data. All concepts are de ned by and in their relations. All concepts (people, organisations, projects, publications) are uniquely identi ed by URIs. 4. Moodle as a e-learning environment
The Cuban Higher Education Ministry (MES) has a national strategy for Virtual Learning Environments (VLE) to support education at every university and at distance, with MOODLE as the preferred platform. VLIRED Cuba has implemented the Moodle platform in the di erent universities of the network. On the level of the programme a central Moodle has been installed including the doctoral school courses and the English course (http://redtic.uclv.cu/moodle/).
Moodle is using LOM, the IEEE 1484.12.1 { 2002 Standard for Learning
Object Metadata, as metadata format. It is an internationally recognised open
standard (published by the Institute of Electrical and Electronics Engineers
Standards Association) for the description of \learning object". The LOM data
model speci es which aspects of a learning object should be described and what
vocabularies may be used for these descriptions; it also de nes how this data
model can be amended by additions or constraints [
Interoperability between the di erent platforms of the information, education and research network VLIRED Cuba
The di erent platforms of the network - Moodle for e-learning, ABCD as a library systems, DSpace as repository and VIVO as research information system - have speci c functionalities which are valuable for their users. The project does not want to create an extra layer over the existing platforms but want to make exchange of metadata and data possible between them.
ABCD, DSpace, Moodle and VIVO are using di erent metadata formats and
ontologies, because of the culture and necessities of the library, repository and
educational communities respectively. But as Erik Duval stated `Interoperability
is enabling information that originates in one context to be used in another in
ways that are as highly automated as possible [
A unique metadata format for all the platforms is not realistic. But there is clearly a need for shared convention for identifying and encoding values before metadata can be exchanged.
The VLIRED Cuba workgroup on research information systems evaluated the elements and elds relevant for reuse in the di erent platforms. In the table below, you nd the major elements relevant for reuse. In red are the ones that can be de ned uniquely.
Over the di erent platforms the project focuses on the main elements that de ne persons, organizations and keywords. How can they be de ned in the metadata format of the di erent platforms? Which authority les and controlled vocabularies can be used to standardize the content? How can unique identi ers be implemented?
Central identi cation systems for researchers and scienti c authors are now
implemented. The most important development is since its creation in 2010
ORCID. The core mission of ORCID is to provide a registry of persistent unique
identi ers for researchers and scholars. ORCID is an open initiative where the
ORCID identi er needs to be integrated into research work ows and linked to
information on research activities such as publications, grants, patents, and
datasets [
Cuban researchers and academics can register in ORCID, simply to be part of that international community. The project is promoting ORCID in the di erent universities. Still, registration is a personal action.
To create an authority list of the authors of the network universities and in a broader context of Cuban authors, a more proactive approach has been developed: by collecting data from di erent resources: Moodle and DSpace of the network universities, ASSETS, IDICT databases, the Tocororo project from UPR and international databases like Web of Science, Scopus, Google Scholar, PUBMED, Researchgate, Scielo, . . . . . The challenge is to collect this data and curate it. As a test in Web of Science the project group could download already 1508 publications from UCLV researchers. The idea is to upload this data in VIVO after curation. VIVO automatically generates an ID for every object, speci cally for every person and organization. Other ID's can be added, in the rst place the ORCID and if developed a Cuban ORCID.
The platforms have to use the same authority le for researchers and academics. A unique ID should be added to the metadata formats. In the table 8 a proposal of implementation is presented.
A way to de ne content unequivocally is by using controlled vocabularies. Di erent thesauri are translated to SKOS. The project already uses CCS, AGROVOC, MESH. Other thesauri can be added. If they are not available in SKOS format, they can be transformed using tools like VocBench.
Metadata elements for subject should be using the same thesauri as authority les.
EsFacil Authority is the tool developed by the project to manage the access to di erent vocabularies and authority systems locally. EsFacil Authority tool is fully implemented based on Free and Open Source Software (FOSS), which allows to share EsFacil Authority with the international community as FOSS through GitHub: https://github.com/ELINF-Cuba-Network.
EsFacil Authority aims to create an applications ecosystem enabling authority control capacities for external applications, by reusing semantically structured data shared by di erent institutions. It aspires to facilitate the processing of authority data in a standardized fashion, following the principles of Linked Data.
EsFacil Authority exposes vocabularies in SKOS format such as AGROVOC, covering all areas of interest of the Food and Agriculture Organization of the United Nations with over 32000 concepts, CCS vocabulary for Computer Sciences and MESH for Medicine and Life Sciences.
It also can integrate data from VIVO, which can be used as authority system for persons and organizations. EsFacil Authority queries to VIVO are done through VIVO's SPARQL endpoint.
The EsFacil Authority interface has four main functionalities exposed as REST web services: { Search for personal authors information. { Search for corporate authors information. { Retrieve registered controlled vocabularies list. { Search for an authorized term on a speci ed controlled vocabulary.
External applications, ABCD, DSpace, VIVO can send requests to EsFacil
Authority, then EsFacil Authority queries its available information sources and
retrieves the requested information structured as a XML or JSON. Figure 10
shows an answer to a query on the term \database" on the ACM Controlled
Vocabulary.
Two main elements are sent as answer in this case, the identi er of the term
in the requested vocabulary and the authorized term by itself. The identi er of
the term is computer oriented for uniquely identify it by using an URI and the
authorized term is what the person using the system sees [
A logic step after the development of the authority system is the development of EsFacil as a submission module to create rich metadata. This is still work in progress but the main components, automated metadata extraction with DarkAIV and authority control (EsFacil Authority) are realized. EsFacil is conceived as an independent module that will ingest records and les in repositories and information systems, in the rst place in DSpace.
The idea is to store rich metadata and translate it to the format of the target applications. Internally, EsFacil uses a nosql approach using the JSONB type of Postgresql, with CSL JSON as metadata format.
The Citation Style Language (CSL) is an XML-Based stylesheet language for formatting of citations and bibliographies. The choice for CSLJSON, , also known as Citeproc JSON implementation, is based on the fact that it is a rather simple but rich format and that it is already used by services like Zotero, Mendeley, CiteProc, Crossref and Islandora. Metadata in CSL JSON for items have a format as described in gure 12. Crossref added an ORCID and an a liation tag. EsFacil proposes to extend it with a URI tag. Another important element is subject. The tag subject can only contain the terms. To enrich it a new tag is created `subject-id' which is enriched with a key as well as with the source.
CSL is very exible. By extending the functionalities of CSL EsFacil still can import existing metadata from various resources like like Crossref, Zotero and Mendeley and at the same time using the EsFacil Authority tool to enrich it and present it through web services to the information platforms of VLIRED Cuba: ABCD, DSpace, VIVO and Moodle. 4
The simplest solution to create interoperable metadata in di erent platforms is to use the same formats everywhere. In reality library, education and research have developed over time speci c formats adapted to the methodologies of their domain. The platforms in VLIRED Cuba uses MARC, Quali ed Dublin Core, LOM and di erent ontologies. The approach of the project is to search for the main common elements used by the di erent platforms. The project workgroup identi ed persons, organizations and keywords as main common elements. The quality of the content depends on the same de nition of the elements, preferably by using unique identi ers for every entity.
Modern authority systems deliver unique id's. For persons ORCID is an international standard, but national or local id's delivered by the research unit are
Fig. 12. Examples of CSL JSON and CSL JSON with EsFacil extension relevant as well. There are initiatives to register organizations initiated by ORCID and GRID (Global Research Identi er Database - https://www.grid.ac/). For keywords there is growing availability of linked data based controlled vocabularies. VLIRED Cuba is developing the tools EsFacil and EsFacil Authority to create standardized metadata in di erent platforms, speci cally ABCD, Moodle, DSpace and VIVO. But the tools can be integrated in other software packages too. These e orts guarantees standardized metadata in the di erent platforms. The next step in the project is to create services that integrates the data from the di erent platforms. In the philosophy of the project this will not replace the existing platforms because each of them have their functionalities and users.
Di erent strategies are possible. VIVO for example, ingest all relevant information in its database. More in general the metadata will have to be exposed to other systems, preferably using a Linked Data approach. Finally in the project some universities are working to create relations between di erent platforms. The Moodle research group at the University of Pinar del Rio has for example developed modules in Moodle that can access DSpace. More developments on services are expected to be realized in the following years.