After two years of preparation, specification and implementation, the task force of CLARIN's centre committee has now made available a new version of CLARIN's metadata framework: CMDI 1.2. Any of the metadata records complying with the CMDI 1.1 standard of which there currently are hundreds of thousands, can from now on be converted to the CMDI 1.2 version of the same record, and be supported as such within the core of CLARIN's metadata infrastructure.
The new version of CMDI adds a number of improvements and new features compared to its predecessor and provides better compliance with XML standards and conventions. Also, it comes with extensive documentation in the form of a specification (final version expected to be publish this autumn) which describes in detail how CMDI records and components may (and may not) be defined, and includes many helpful examples providing guidance for metadata creators, modellers and application developers. The newly released CMDI 1.2 toolkit provides scripts, utilities and libraries useful to for instance developers that would like to adapt their software and repository managers interested in converting their existing metadata. A ready to use batch conversion tool for CMDI records can be downloaded separately as well - see the CMDI 1.2 page for details.
With the recent 'rollout' of CMDI 1.2, the core metadata infrastructure has been adapted to support both CMDI 1.1 (as it has done for many years) and its successor. This adaptation was designed to have as little impact on the existing infrastructure as possible. As a result, all existing CMDI components, profile and records remain accessible and have kept their validity. In fact, deprecation of CMDI 1.1 has explicitly not been put on the roadmap yet. Although in the long run support for older version(s) of CMDI will most likely be diminished, this will certainly not happen until most, if not all actively developed tools in the CLARIN ecosystem have been extended with support for CMDI 1.2. We hope for many of the tools that are developed at centres to pick up such support soon, so that ultimately the end users can benefit from its improvements.
A detailed overview overview of the new features and other changes in CMDI 1.2 can be found in the annotated version of the 2014 executive summary of changes in CMDI 1.2. The following are the most salient:
- Fields can be associated with external vocabularies (from CLAVAS), providing options for both closed and open controlled vocabularies
- Attributes on elements and components can be made mandatory, whereas they are always optional in CMDI 1.1
- The documentation options for components and profiles have been improved, now allowing for multilingual documentation on all levels
- Each profile has its own XML namespace
- The way in which relations among resources within a CMDI record can be expressed has been improved
In addition there have been numerous cleanups to the language used to define components (the Component Specification Language CCSL), and the language has been prepared for future extensions in the domains of display information, automated value derivation and lifecycle management (including deprecation and succession) for components and profiles. Not all of the mentioned features have been implemented into the Component Registry yet, which means they cannot be applied by metadata modellers at the moment. However, the next release of the component registry (expected autumn 2016) will make most of the new functionalities available through its user interface.
The overall aim of these changes has been to facilitate the creation of higher quality metadata, in line with one of the main focus areas of the CLARIN-PLUS project. For example, the usage of semantically annotated open vocabularies leads to fewer spelling or naming variations and more reliable value mapping in the context of metadata based search; a lifecycle framework for components and profiles makes it easier for metadata modellers to propagate fixes and other relevant changes while preventing metadata creators from accidentally using outdated profiles; multilingual documentation increases the likelihood that the semantics of metadata fields and their values will be interpreted correctly by metadata creators and end users of repository software.
Although it is possible and perfectly safe to keep using CMDI 1.1, you may be interested in switching to CMDI 1.2 so that you can benefit from its new features and other improvements - wether you are a metadata modeller, creator, developer, end-user or responsible for a metadata repository. We have collected answers to questions regarding the usage of and migration to CMDI 1.2 for various user groups on the special CMDI 1.2 page on clarin.eu. The task force will be adding up-to-date information over the coming time period, and will also be available to answer any questions regarding component metadata via firstname.lastname@example.org.