# Introduction ## Objective The objective of this document is described in @charta2022 as following: > Assessment of standards (i.e. formal, accessible, shared and broadly applicable language for knowledge representation) > for the creation of mappings (e.g., SSSOM, OAEI), the formal documentation of mappings and their metadata in > collaboration with the Terminology Services WG, as well as their (re-)evaluation and publication. To some degree this also include tools, procedures, frameworks and recommendations for the creation, documentation, validation and publishing of mappings. ## Glossary - **semantic artifacts** (SA) subsume all kinds of formal ontologies, terminologies, vocabularies, lists of terms, and similar knowledge organization systems (KOS). In the best case they are available in RDF and hosted in a terminology registry. Many ways exist to classify SA by type and boundaries are fuzzy. One classification used in the report is this: - An **ontology** is a semantic artifact expressed in RDF/OWL format. - A **schema** is a semantic artifact primarily used as data format. It can better be expressed with a specific schema language such as XML Schema. - A **terminology** is a semantic artifact that can (at least roughly) be expressed in SKOS format. This includes all SA except ontologies and schemas. - A **name authority list** ... - Individual elements of a semantic artifact (concepts, classes, items, terms...) are referred to as **entities**. - A **mapping** is a managed link between two entities from two semantic artifacts (1-to-1 mapping) or between one entity from one semantic artifact and multiple mappings from another one (1-to-n mapping). The mapping includes additional information such as mapping type, provenance etc. - **mapping source** is a semantic artifact an entity is mapped from - **mapping target** is a semantic artifact an entity is mapped to - A **concordance** (also known as crosswalk or alignment) is set of mappings between two semantic artifacts. Concordances can help for data integration and enrichement by bridging the gap between different data models and vocabularies. - A **transformation** is an executable function that converts a record with entities from a mapping source into a record with entities from a mapping target. ## Examples ~~~{.ttl} @prefix skos: . @prefix wikidata: . @prefix gnd: . @prefix iconclass: . wikidata:Q11451 skos:exactMatch gnd:4034402-2 . wikidata:Q11451 skos:exactMatch iconclass:47I1 . ~~~ # Review ## Not best practice - Visual graphs (mostly with pen and paper): useful for thinking but not formal - Source code of applications that apply transformations: better than a standard without implementation but not scalable - Spreadsheets: the plague but unavoidable and handy ## Standards In no particular order: - RDFS and OWL properties - [SKOS mapping properties](https://www.w3.org/TR/skos-reference/#mapping) - [JSKOS mappings](https://gbv.github.io/jskos/#concept-mappings) and [Concordances](https://gbv.github.io/jskos/#concordances) - [Simple Standard for Sharing Ontological Mappings (SSSOM)](https://mapping-commons.github.io/sssom/) - [LinkML-map](https://linkml.io/linkml-map/) - [Expressive and Declarative Ontology Alignment Language (EDOAL)](https://moex.gitlabpages.inria.fr/alignapi/edoal.html) # Instructions and recommendations This document may eventually contain recommendations on: - creation of mappings - validation of mappings - documentation of mappings - publication of mappings - mapping registries - [mapping tools](#tools) To get there, we will: - Find instructions and have specialist discussions on from the field of computer science. - Refer to existing guidelines - Assess and discuss how to use existing standards in NFDI ## Where are mappings managed 1. As **part of** a semantic artifact. An example is the [Integrated Authority File (GND)](http://bartoc.org/en/node/430) with embedded links to LCSH, MeSH, AAT, ORCID and other systems. 2. In **plain files** (Spreadsheet/CSV, JSKOS, SSSOM, RDF/Turtle...), possibly managed in a version control system 3. With a [mapping editor](#mapping-editors) and/or in a mapping registry 4. As part of executable **transformations** ## Tools A comprehensive list of tools for management of semantic artifacts is provided by @Voss2025^[Online at .]. This report includes some mapping editors, generators and converters listed below. Some generic KOS editors also allow to manage mappings from the semantic artifact being managed to other KOS. ### Mapping editors - Cocoda - MSCR - Amalgame More specialized mapping editors for a limited kind of SA include [OpenRefine](https://openrefine.org/) and [Mix'n'match](https://mix-n-match.toolforge.org/). Several more tools and frameworks have been created or proposed in context of the Ontology Mapping & Alignment Initiative (OAEI), for instance [AgreementMaker Ontology Matching System](https://agreementmaker.github.io/). These tools, however have rarely become more than a short-lived research demo, and they mostly aim at automatic creation of mappings instead of their intellectual curation. ### Mapping converters - jskos-cli - sssom-py, sssom-java, sssom-js ## Databases - knowledge graph database management systems [Semantic MediaWiki] and [Wikibase]. - [Alignment API and Alignment Server](https://moex.gitlabpages.inria.fr/alignapi/) ## KOS editors If mappings from a semantic artifact are primarily managed by the editors of this SA, it makes sense to combine management into one tool. This ensures good overview of the mapping source but mapping targets are covered less so.