A retrospective of a pioneering project. earlier than XML, Other than SGML, Still going: CuProS metadata for deeply nested relations and navigating for retrieval in RAFFAELLO

Abstract

The spread of XML vindicated a project that the present authors developed earlier, independently of SGML, the parent-language of XML (SGML was originally devised for communicating structured documents). Our own project was the Navigation component of the Raffaello retrieval system. The latter was originally devised for managing the deeply nested, flexible relations that constituted the lexical database of the Onomaturge expert system for Hebrew word-formation. Whereas Raffaello within Onomaturge was a simpler version, such that retrieval was done by means of retrieval functions implemented in Lisp and reflecting the actual structure of the nested relations as known beforehand, the version of Navigation that was implemented by El-Sana under Nissan’s supervision used a metadata schema. The syntax for describing metadata in Raffaello was in a language defined by Nissan, CuProS (short for Customization Production System). Whereas a few articles about Raffaello were published, none explained CuProS and the Navigation tool as implemented by El-Sana. The unabated interest in XML, Web technology, and ontologies (an area relevant for the database of Onomaturge) have vindicated Nissan’s contention that nested relations (a research area in database design) should be allowed not only unlimited depth of nesting, but also extreme flexibility of structure. This is now taken for granted because of how metadata are defined for XML, but the feasibility of the idea was shown by El-Sana’s managing to implement retrieval as based on metadata description, along the lines of CuProS syntax. In this article, apart from explaining the syntax of CuProS and also describing El-Sana’s implementation of retrieval, we also illustrate the approach to metarepresentation through an exemplification from the structure of lexical frames in Onomaturge. In particular, we discuss variants of a given derivational pattern of wordformation, and we also discuss the evolution of terminology for given lexical concepts, throughout historical strata of Hebrew. We show how this is handled in nested relations, but a fuller discussion is provided in appendices. The approach is also exemplified based on a project applied to Italy’s regional constitutions.