CMDI 1.2 - Improvements in the CLARIN Component Metadata Infrastructure
- 1. CMDI 1.2 Improvements inthe CLARIN Component Metadata Infrastructure Mitchell Seaton Center for Language Technology Department of Nordic Research, UCPH [email protected] 10 May 2016 CLARIN Centre Meeting, Utrecht
- 2. CMDI • Schema Component-based modelling,flexible Specification language (CCSL) • Metadata CMDI records (instances) are based on CMD profiles • Standardisation CMD model ISO 24622-1:2015 • Semantics Consists of semantic annotations (concepts, CCR) 2
- 3. CMDI Framework 3 Concept definitions Profile1 Profile 2 Profile 3 Comp B Metadata record Metadata record Metadata record instanceOf contains links to Comp A Comp C
- 4. CMDI Taskforce https://trac.clarin.eu/wiki/Taskforces/CMDI • Resp:Writing specifications, documentation, software development, and integrations. • Contact: [email protected] • Coordinators: Twan Goosen (CLARIN ERIC) Menzo Windhouwer (Meertens Institute, MPI for Psycholinguistics) 4
- 5. CMDI 1.2 NewFeatures • Lifecycle Management • Open/External Vocabularies • Cues for Tools • Value Derivation • Mandatory Attributes 5
- 6. CMDI 1.2 Improvements 6 •XML-compliance, cleaner XSD schema, enhanced validation (schematron assertions) • CMD Record envelope changes from 1.1 <IsPartOfList> moved Stricter <ResourceRelation> Single ResourceProxy reference (@cmd:ref) on Components • Namespaces Reserved attrs (@cmd:ref, @cmd:componentId) Profile-specific payload namespace (cmdp) • Documentation Multilingual (@xml:lang) Component and Attribute levels (CCSL)
- 7.
- 8. Lifecycle Management 8 • Definedin Header elements for component spec • <Status> ('development', 'production', 'deprecated') • <StatusComment> (for deprecated) • <Successor> • <DerivedFrom> TextCorpus v1 Status: deprecated Successor: DerivedFrom: TextCorpus v3 Status: development DerivedFrom: v2 TextCorpus v2 Status: production Successor: v3 DerivedFrom: v1 Corpus Status: production DerivedFrom: v2 derived
- 9. Namespaces 9 • Global namespaces CMDI instance (general/envelope) namespace cmd - http://www.clarin.eu/cmd/1 Cues for tools namespace cue - http://www.clarin.eu/cmd/cues/1 • Profile-specific payload namespace (cmdp) cmdp - http://www.clarin.eu/cmd/1/profiles/{profileId} {profileId}: Identifier of the profile from which the schema is derived
- 10. Open Vocabularies 10 • CLAVAS(OpenSKOS) vocabulary service https://openskos.meertens.knaw.nl/clavas/ Hosted by Meertens (NL) • Defined in new <Vocabulary> element • External vocabulary (referenced by @URI attribute) May use @cmd:ValueConceptLink attribute in a metadata record as link to vocabulary entry • Controlled vocabulary (imported items, sub-set) • Localisation (@ValueLanguage attribute)
- 11. Cues for tools 11 •New XML Namespace (cue): http://www.clarin.eu/cmdi/cues/1 • Use cases: Enhanced documentation capabilities (localisation) Presentation hints (editors/browser tools)
- 12. Auto Value Derivation 12 •Optional <AutoValue> CCSL element on CMD Elements and Attributes • Derive content for an Element/Attribute from other values • Value may give information about a derivation function • No provided set of derivation functions • @cmd:AutoValue attribute used in the generated schema definition
- 13. CMDI 1.2 SpecificationDraft 13 • Formal specification draft • Review-cycle https://trac.clarin.eu/wiki/CMDI%201.2/Specification
- 14. Conclusion 14 • Development/review ofCMDI 1.2 specification draft • Implementation of CMDI 1.2 • CMDI 1.2 rollout/release