- A Use one common system
- Requires: (Re)classify/(re)index all documents (manually or with automated support)
- Requires: Select an established or create a new system
- Suited for a university level TD or a subject TD database
- Exception: existing common national subject indexing solutions (Germany, DINI recommendations)
- Monolithic rather than flexible solution
- High costs
- Suitable automatic tools missing
- Systems not in major use (Sweden)
- Creation of new systems: history of CERIF
- Local and national structure of academic institutions and educational systems
Keyword indexing
- Few candidate vocabularies. LCSH, SWD suitability
- Automatic indexing and keyphrase extraction
- Methodological complexity of mapping
- Complementary approaches (VT, NDLTD research)
- B Map to a common "switching" system
- Map heterogeneous systems for integrated subject access
- Broad, deep, well-used and maintained system needed
- No need to operate at record level or to change local system
- Switching structure can be used as integrated browsing system. Different browsing views could be mapped on top of it
- Limited loss of precision
- Mapping to subject specific systems more worthwhile
- Example Renardus. Vascoda
- Semi-automatic mapping
- Future effort sharing. Mapping infrastructure
Potential actions, e.g.:
- Agree on minimum subject access principles and features
- Investigate browsing system/classification and granularity
- Support document-type filtering (TDs) and subject browsing and filtering in existing systems
- Carry out end-user usability studies
- Stimulate development of (semi-)automatic classification methods
- Stimulate development of (semi-)automatic classification mapping methods
- C Support indexing/classification in different systems
- Provides increasing overlap and partial interoperability "from below". Option for international services such as NDLTD and XTCat
- Source for semi-automated classification or multilingual access efforts in other services
- Subject specific and secondary systems in Germany
- Interoperability with other systems if mappings available or directly
Potential actions, e.g.:
- Issue general recommendation to keep all subject indexing information in TD records and carry it with the records into new contexts.
- Issue related recommendation to classify (and index) as detailed as it can be afforded as close to the source of the publication as possible to improve later options of the documents to be re-used in both general and specialised services.
- D Do nothing. Social tagging
- Rely on fulltext indexing, citation indexing from Google and similar and on success with social tagging and other discovery approaches
- Awareness in several communities that more quality solutions and features are needed
- Social tagging limitations and characteristics of TDs. Watch CiteUlike
Potential actions, e.g.:
- Start experimenting with "social" subject tagging using controlled vocabulary in a subservice for ETDs or attached to subject based services and evaluate performance and usability.