The ABCDEFG schema is very granular and allows a detailed description of data gathered in geoscience (see technical schema documentation at http://www.geocase.eu/efg). It comprises a broad range of properties, including

geological and geomorphological observations;

As fossil objects share a considerable number of properties from both biology and geology, it is a great advantage to establish a geoscience schema as an extension of a biological data standard, rather than creating a completely separated standard for geosciences. Furthermore, another area of overlap are the data elements that refer to the collection or collection event itself. The EFG extension is not applicable without its backbone, the ABCD schema itself (Fig. 1). Elements describing the collection, the holding institution, contact persons, and any additional information on the dataset are placed within ABCD. Furthermore, details on the gathering event (location, date, collector, etc.) and links to object related multimedia material including their respective licensing or any related publication can be specified in the ABCD standard.

ABCDEFG is an XML-based schema and thus can be easily processed by software and is readable by humans. The schema is hierarchically structured and the descriptive elements are aggregated in thematic concepts. All elements are arranged according to their contextual relation to other elements and their cardinality (Holetschek et al., 2012). The full potential of ABCDEFG is unlocked in combination with the free and open-source BioCASe Provider Software (BPS; http://www.biocase.org/products/provider_software/). It consists of a generic XML wrapper software for relational databases and a custom XML-based protocol for communication, and it requires a schema definition (XSD) for each supported data standard.

The BPS XML wrapper is able to connect to a wide array of database management systems (e.g., MySQL, PostgreSQL, Microsoft SQL, Oracle, as well as Excel spreadsheets) and provides a user interface for mapping the database fields to the data standards' elements. This generic approach of mapping the individual data models to the standard schema is the essential functionality for standardized data mobilization. Once an institution has set up the mapping for its database, the released content (including revisions, updates and new records) is exposed in the same manner through the BPS. Thus, only changes in the data model itself necessitate modifications in the mapping.

Although mapping of data is straightforward, the (potentially heterogeneous) content needs to be checked on a regular basis to verify that the results are appropriate within the data standard's context. This practice is supported by the BioCASe technology, as different datasets and sources provided in ABCD and ABCDEFG can easily be monitored using the BioCASe Monitor Service (Glöckler et al., 2013). By using the BioCASe data interface and protocol in the backend and a comprehensible user interface as a frontend, the monitoring tool facilitates checks of the structure, plausibility, and completeness of datasets. Furthermore, it verifies compliance of provided data for transformation into other target schemas and can be used for summaries and simple descriptive statistics across different datasets. Thus, data aggregators like the Geosciences Collection Access Service (GeoCASe) can make use of the tool in the supervision of the progress in data provision. A full implementation of the BioCASe Monitor Service for summarizing the data provision and checking the minimum requirements of data mapped to ABCDEFG can be found on the GeoCASe website (http://geocase.eu/partners_and_providers).

For several geoscientific disciplines, data recommendations exist and structures provide guidance on what kind of data should be recorded in what form during field work or which research information should be publicly available. The Evolution of Terrestrial Ecosystems Program (ETE; Damuth, 1997) was designed in the late 1990s and provided the basis for later developments of various data structures (Reed et al., 2015). At about the same time the Paleobiology Database (PBDB; https://paleobiodb.org) was founded. The PBDB is a public resource of collection-based occurrence and taxonomic data for fossils of all geological ages. The PBDB considerably inspired the development of the EFG schema. Other important data conventions for paleontology are the FAUNMAP and MIOMAP initiatives, focusing on fossil occurrences in North America from different time periods (Carrasco et al., 2005; Graham and Lundelius Jr., 2010).

All these data structures (ETE, PBDB, and FAUNMAP/MIOMAP) perfectly manage localities and faunal lists, but cannot be applied on specimen data in a straightforward fashion. Gilbert and Carlson (2011) established a data dictionary for paleoanthropology, including a specific table schema for specimens as the central unit of organization, and localities as clusters of specimens. This structure facilitates the data collection in the field, the curation of objects in institutional collections, and the publication of specimen related information. Along with the International Geo Sample Number (IGSN) for the registering of samples to the System for Earth Sample Registration SESAR two metadata sets were defined. These metadata schemes comprehend core metadata used for the registration itself (registration metadata) as well as basic characteristics of physical samples and collections (descriptive metadata and latest version of both sets are available via https://github.com/IGSN/metadata/wiki). EarthChem (http://www.earthchem.org/), the NSF (National Science Foundation)-funded data facility for solid earth geoscience data, provides suggested and controlled vocabularies for geochemical, geochronological, and petrological data in their EarthChem Systems (http://www.earthchem.org/resources/vocabularies) in order to help standardizing these types of data in general. In addition, recommended properties for a detailed description of minerals and their structure can be derived from Mindat (http://mindat.org), the most comprehensive and world's largest public database for mineral information. However, none of the above-mentioned recommendations has yet reached the status of a ratified standard.

In biodiversity science and related disciplines, Darwin Core (DwC) and ABCD are the best-known and most commonly used standards. In parallel with ABCD, Darwin Core evolved over time and its development is mainly driven by the community (Wieczorek et al., 2012). Although both schemas aim at facilitating the mobilization of biodiversity data, they differ in their purpose. ABCD is a highly structured schema and was intended to be applicable to a great variety of data in various degrees of granularity, including their relations and cardinality. It is almost exclusively tied to the XML format. Darwin Core in contrast aims at providing a flexible vocabulary for sharing information about biological diversity and can be represented using various technologies including XML, RDF, or plain CSV files. It is used to describe the recorded occurrence of a species in space and time, and to link related information and evidence to this declaration. The Darwin Core standard was ratified in 2009 by TDWG and since then it is the most common standard for publishing datasets in the Global Biodiversity Information Facility (GBIF; http://gbif.org).

The usage of Darwin Core for geoscience is limited. Although specimens can be classified as “FossilSpecimen” (http://rs.tdwg.org/dwc/terms/FossilSpecimen) and basic information about the chrono-, litho-, and biostratigraphy can be accommodated in the GeologicalContext terms (http://rs.tdwg.org/dwc/terms/GeologicalContext), appropriate terms for geological collection items are still lacking in Darwin Core. Thus, important measurements and facts like the host rock of a specimen, the original stratigraphic association of allochthonous material and metasomatic, metamorphic, or diagenetic alterations of the rock cannot be shared and distributed using this data standard. The Darwin Core Paleontology Extension (DarwinCoPE) was proposed as a community standard in 2005, but even then it was described as less detailed than the evolving ABCDEFG (Theodor, 2006). Furthermore, the PaleoCore initiative established a list of common terms for the publication of paleoanthropological datasets (http://paleocore.org/standard/) and combined elements from Darwin Core with some of the Dublin Core Metadata Initiative (DCMI, http://dublincore.org/documents/dcmi-terms/). However, the possibilities of accommodating comprehensive geoscientific data are still limited.

In summary ABCDEFG is the most comprehensive data standard for geoscientific collection data. Although there have been different approaches, none of them provide a comparable detailed and atomized schema for both paleontological and geological data.

ABCDEFG is already used for data aggregation and provision to a number of thematic data portals. The following examples illustrate different data portals and aggregators showing collection related information from the same data source endpoints mapped to the ABCDEFG standard. The Global Biodiversity Information Facility (GBIF, http://www.gbif.org) is a web portal for the access to biodiversity data. It accepts several different data standards, including Darwin Core, ABCD, and ABCDEFG. As the paleobiological data provide important information about ancient biodiversity, fossils from natural history collections and in the published literature are important data sources for GBIF. Therefore, different institutions provide occurrence data of not only extant species but also of fossil specimens.Examples:Fossil (Mammuthus primigenius), http://www.gbif.org/occurrence/1099276603Fossil (Dorygnathus banthensis), http://www.gbif.org/occurrence/1099085192 Within the SYNTHESYS project task Geosciences Collection Access Service (GeoCASe) a data portal (www.geocase.eu/access) was developed to make data from the GeoCASe network freely available on the internet. The portal uses the BioCASe technology to perform a distributed query on the original provider's databases. In contrast to paleontological specimens, data on geoscientific collection objects are not available via GBIF. GeoCASe thus represents a complementary portal for non-biological data, while both portals overlap in data provision on fossil objects. GeoCASe originally had its scope on geoscientific objects (fossils, minerals, and rocks) from European institutions. However, it is now open for worldwide earth science collectionsExamples:Fossil (Mammuthus primigenius), http://geocase.eu/portal/?unitid=MB.Ma.13543Fossil (Dorygnathus banthensis), http://geocase.eu/portal/?unitid=MB.R.3661Mineral (azurite), http://geocase.eu/portal/?unitid=MFN_MIN_1999_2270 The German Federation for Biological Data (GFBio; http://www.gfbio.org) takes care of the management and standardization of biological research data during the entire data life cycle (DataOne, 2017) and uses ABCDEFG for fossil data in order to standardize heterogeneous datasets and to publish the data on the national GFBio portal. The GFBio data centers are responsible for the sustainable management, long-term accessibility, and archiving of their client's data. Therefore, they build upon domain-specific data standards like ABCDEFG that are actively used, well documented, and accepted by the community.Example:Fossil (Jurassiphorus cailliaudanum), https://www.gfbio.org/data/search?q=MB.Ga.524Species (Mammuthus primigenius), https://www.gfbio.org/data/search?q=Mammuthus primigeniusEuropean digital library Europeana. The European digital library Europeana (http://www.europeana.eu) displays the cultural heritage of Europe. The EU-funded project OpenUp! and its successor project Europeana DSI provide more than 3 million images and other multimedia objects from the natural history domain (Berendsohn and Güntsch, 2012; http://open-up.eu). Project partners are delivering their data in the ABCDEFG schema using the BioCASe technology and making their multimedia files (mainly images) freely available, including those of fossils, minerals, and rocks. As Europeana focuses on media objects, specific filter requests exclude all those records without any multimedia files. But again, none of the data providers need to set up a separate schema mapping, parallel to those of the above-mentioned data portals. Furthermore, Europeana compiles data from cultural heritage collections that use their own data standards. In order to harmonize these with the data from the natural history domain, a subset of elements from the ABCDEFG schema was identified to meet the requirements of the Europeana data model (ESE/EDM; Zágoršek et al., 2012). As ABCDEFG is XML-based, the transformation to the Europeana data model is straightforward.Examples:Fossil (Dorygnathus banthensis), http://www.europeana.eu/portal/de/record/11622/_MFN_FOSSIL_VERTEBRATES_V_MFN_GERMANY_MB_R_3661.htmlMineral (azurite), http://www.europeana.eu/portal/de/record/11622/_MFN_MINERALOGY_MFN_GERMANY_MFN_MIN_1999_2270.htmlObject landing pages of the Museum für Naturkunde Berlin. Recently, the Consortium of European Taxonomic Facilities (CETAF; http://cetaf.org/) adopted the approach of using actionable, web-based uniform resource identifiers (http-URIs) for physical objects (Güntsch et al., 2017). The advantage of this approach is the immediate linkage of the physical object with its digital representation. As the Hypertext Transfer Protocol (http) is a well-known syntax, people tend to “click” on the URI and expect to find information about the physical object. Thus, the http-URIs ideally resolve to web landing pages with a compilation of the objects' media and collection data. At the Museum für Naturkunde Berlin these landing pages make use of ABCDEFG structured data to use the standard's controlled and well-documented terms for exposing the information.Examples:Fossil (Dorygnathus banthensis), http://coll.mfn-berlin.de/u/MB_R_3661Mineral (azurite), http://coll.mfn-berlin.de/u/MFN_MIN_1999_2270

Table 1 summarizes the number of institutions using the ABCDEFG schema and the number of objects published in different portals.