Temporal Capabilities in Support of Conceptual Process Modeling using Object- Role Modeling Daniele Gianni, d.gianni@unimarconi.it, Guglielmo Marconi University, Italy Paolo Bocciarelli and Andrea D’Ambrogio paolo.bocciarelli@uniroma2.it, dambro@uniroma2.it University of Rome TorVergata, Italy Keywords: Object-Role Modeling, Temporal Modeling, Conceptual Data Modeling, Process Modeling. Abstract Conceptual data modeling languages must be provided with temporal capabilities to support the data evolution throughout the execution of a conceptual process model. Asides from supporting the storage of historical data, temporal capabilities must also provide the means for verifying the consistency between the data temporal properties and the data modification resulting from the process execution. The Object-Role Modeling (ORM) language is a conceptual data modeling language that is based on the concepts of Fact (i.e. true statements on the represented world), Fact Type, and Fact Base (i.e. the set of all the Facts). Currently, the ORM language does not address the specification of Facts temporal properties, and therefore does not also support the verification of Facts variations during a process execution. The paper introduces an initial ORM overlay methodology that aims to laying the foundation of the conceptual modeling structures that can support the verification of temporal evolution of conceptual data models (i.e., whether a Fact can be asserted or retracted, depending on its temporal properties). Moreover, the overlay methodology also defines a temporal visual notation and an initial semi-formal temporal verbalization that eases the use of the methodology to the ORM modelers. A simple example illustrates the potential application of the overlay methodology. 1. INTRODUCTION Temporal modeling has been a widely researched area in support of historical data storage in database systems [1]. The Object-Role Modeling (ORM)—the most popular conceptual data modeling language [2]—has also been provided with methods for the representation of temporal concepts within a business domain [2]. However, when using ORM for conceptual process modeling, a different use scenario is to be considered as a conceptual process model can be executed and the execution affects the represented world (e.g. assertion or retraction of Facts in the Fact Base). As a result, two new needs arise: 1) ensuring the consistency between the process specification and the data specification; and 2) supporting the monitoring of the process execution. Concerning the temporal aspects, the consistency can be supported by classifying Facts according to the possible temporal properties indicating their temporal validity, thus constraining the assertion or retraction of Facts. Differently, the monitoring can be achieved with the definition of verbalization capabilities that verbalize the Fact Base populations depending on the current time and on the time in which the Facts were asserted or retracted from the Fact Base. The paper presents the initial steps for the definition of an overlay methodology for temporal modeling capabilities in support of conceptual process modeling using ORM. The overlay methodology consists in the definition of temporal versions of key Fact-based Modeling (FBM) (e.g. Fact, Fact Type, Fact Base, and Schema), introducing the relationships with the concepts of temporal validity time. Moreover, the temporal version of the Fact Base is also associated to current time concept for maintaining a unique identifier for respective temporal snapshots. A visual notation is also provided to hide the complexity of the overlay methodology to the modelers, who can more easily overcome the limitations of the ORM language related to the definition and verification of Facts temporal properties. In addition, the overlay methodology provides an initial and semi-formal verbalization of the newly defined temporal Schema and Fact Base. The temporal Schema verbalization aims to contribute to ease the use of the temporal method to ORM modelers. Differently, the Fact Base verbalization, inspired by the Constellation Data Language [1], specifically addresses the above mentioned issue of monitoring the process execution. However, the overlay methodology is constrained by a number of simplifying assumptions which may be potentially overcome in future developments or better tuned to the specific business domains. Currently, the overlay methodology assumes that 1) time intervals are atomic and non-overlapping; 2) Facts have unchangeable temporal classification; 3) time point represents the “fixed” level of granularity for the temporal modeling. Although not formally demonstrated in the paper, these assumptions can be intuitively (but not trivially) overcome with the definition of overarching modeling structures and the respective derivations rules, as outlined in the paper. The paper is structured as follows. The Related Work section compares the overlay methodology with existing Proceedings of the Symposium on Theory of Modeling & Simulation-DEVS Integrative, Tampa, Florida, US, April, 2014