The Speech & Language Data Repository (SLDR, http://sldr.org) is a CLARIN-C centre aimed at pooling and preserving linguistic/oral resources. It is currently hosted on a virtual machine in the workspace of ORTOLANG at INIST.
The main features of this repository are:
- It is open to the scientific community regardless of geographical and institutional boundaries. To this effect, navigation tools and descriptive metadata management are multilingual: four navigation languages (English, French, Spanish, Chinese) plus an optional language for metadata. In February 2015, SLDR enlisted 557 participants from 50 different countries.
- Generic processing of items as ‘packages’ (according to OAIS terminology). No classification plan is imposed on data producers. Item types (primary data/corpus, secondary data, tools) are determined by their own metadata.
- Access rights are consistent with the provisions of the French Code du patrimoine. Curators do encourage the Open Data approach, starting with scientific usage with the permissions of participants. During periods recognized by legal derogations, producers may restrict access to certain categories of users, or individuals and institutions eligible for privileged access.
- For all data available after login to the site, the distribution is subject to the acceptance of a SLDR licence plus, if necessary, a customised licence stating the conditions for use of the resource.
- For items in open access, all categories of Creative Commons licences are available and their use is strongly recommended.
- The production of persistent identifiers (PIDs) ensures access to objects and their contents (with an adjustable granularity) via permanent URLs. This production follows a deterministic algorithm (see presentation) that enables producers to prédict the identifier of each document, thereby facilitating its use in automated processing or publications.
- Any resource downloaded by registered users provides access to its User community for the sharing of experience about its use and interest. Public (moderated) comments are also associated with each item distributed by SLDR.
Current work on SLDR aims at ensuring the continuity of deposit, curation and dissemination of data, until features of the new ORTOLANG platform allow a full migration of its service and data/metadata. Details on SLDR features and a follow-up of its migration may be found here.