Interoperate with whom? formality, archaeology and the semantic web

‘Interoperability’ is often cited as a desirable end-goal for information systems, but the highly abstract nature of this apparent benefit sits uneasily with the task-oriented realities of data-management. The approach most frequently advocated is to increase the formality of the system, which facilitates system-integration yet also raises additional barriers to entry that reduce the potential pool of systems to interoperate with. The Semantic Web initiative in particular has faced accusations that difficulties associated with its adoption can outweigh the perceived benefits of data-sharing. This issue will be discussed in reference to current doctoral research being undertaken in Humanities data integration. It will argue that technologies that either heavily front-load or defer dealing with semantic complexity are unlikely to be viable across the producer spectrum. Recourse to altruistic arguments suggests a tacit acceptance that application of such technologies may not be in the immediate interest of the curator. An approach which offers multiple ‘pay-off points’ is inherently more attractive to potential adopters. In particular, we focus on means by which data-driven microproviders - owners of the small but important datasets that tend to form the ‘long tail’ of academic data in the Humanities - can participate in semantics-driven datasharing.We propose that (at least) five escalating levels of semantic formalization can be identified, each with differing requirements and benefits for the implementer: i. Literal Standardization, ii. Instance URI generation, iii. Canonical URI mapping, iv. RDF generation, and v. Database-schema-to-Ontology mapping. We note that Linked Data - hitherto seen as the simplest semantic approach - is relatively advanced in this scheme. We argue that data providers should be encouraged to migrate towards full semantic formalization only as their requirements dictate, rather than all at once. Such an approach acts as both a short and long-term investment in semantic approaches, in turn encouraging increased community engagement.We also propose that for such processes to be accessible to data-curators with low technical literacy, assistive software must be created to facilitate these steps. We have been developing a prototype package targeted specifically at archaeologists that enables them to produce valid, globally-integrated RDF from unnormalized excavation data with minimal technical knowledge. This takes the form of a Wizard that inspects legacy relational data, and provides predictive mapping and association which users can confirm or amend. A secondary program uses the resulting output in conjunction with the original data in order to produce valid RDF/XML as desired. Using this software we have demonstrated that archaeological excavation data encoded in a variety of formats, languages and schemas can be successfully integrated by its curators