Potentially Relevant Interoperability Standards, Protocols, Application Profiles, Best Practices
The following questions can be used to organize classes of metadata standards (i.e., what is the purpose of the standard?)
- What do we need to know about an object?
(A separate issue, possibly relevant: standardized data dictionaries and vocabularies)
- Identifier standards
- Descriptive metadata
- Rights metadata
- Structural metadata
- Technical metadata
- How is that information encoded?
- Record format/markup language
- Metadata Wrappers
- How is that information transmitted and accessed?
This is a gray zone: although not strictly metadata standards per se, many APIs also require that data be encoded in certain markup languages. Transformations come into play here.
- What do we need to know about a Collection? Relevant resources (putting this here as a placeholder until CI subgroup 1 forms its pages; 12/20/2010 Claire Stewart)
- Asset Action project (view and get stuff, object-specific? but can make assertions at the collection level, hopefully? about which actions are supported throughout)
- IESR collection description (mostly descriptive about a collection, but permits entry of OAI information, Z39.50 including location of ZeeRex explain file, indication that other methods like FTP, OpenURL, SRU, SRW, etc. are supported)
- RSLP collection description
- RIF-CS Registry Interchange Format - Collections and Services collection description
- Collection Description Specification , draft NISO Z39.91
- Making Collections Web Friendly (informal notes. Added by E. Kansa 1/7/11)
- In addition to the metadata standards etc. described above, there needs to be some emphasis placed on best-practices for Web-architecture. This means measures to promote interoperability / use of collections in the context of the open Web.Some things to consider:
- URL/URI design, linked data.
- Content negotiation, publication of multiple representations of resources (JSON, XML, etc.) to facilitate use with mobile devices, ebooks, mashups, etc.
- Other attributes of RESTful design.
- Serviced based on the Atom Syndication Format
- For data portability. Serialize query results as paged Atom-feeds. Useful for retrieving subsets of collections that may be of interest (one can then further annotate resources in such subsets according to whatever ontology is needed for a given research question)
- For data preservation / archiving. Digital libraries can crawl feeds to get lists of resources (and metadata about such resources) to archive.