Paper Search Console

Home Search Page Alphabetical List About Contact

Journal Title

Title of Journal:

Search In Journal Title:


Search In Journal Abbravation:


Springer, Cham

Search In Publisher:



Search In DOI:


Search In ISSN:
Search In Title Of Papers:

Modeling with a UML Profile

Authors: Jugurta Lisboa Filho, Cirano Iochpe,

Publish Date: 2016
Volume: , Issue:, Pages: 1-12
PDF Link


A spatial database management system (SDBMS) provides storage structures and basic operations for spatial data manipulation, whereas geographic information systems (GIS) provide the mechanisms for analysis and visualization of geographic data (Shekhar and Chawla, 2003). In this way, geographic databases (GeoDB) are collections of georeferenced spatial data, stored by SDBMS and manipulated by GIS.GeoDB, as any database, must be designed following the traditional database design methodology that includes the conceptual, logical and physical design phases (Elmasri and Navathe, 2000). To draw up a data schema during the conceptual phase, a conceptual modeling language must be used. A strong tendency exists in computer science to adopt the Unified Modeling Language (UML) (OMG, 2007) as a system modeling standard based on the object-oriented paradigm, and more specifically the UML class diagram for database design. However, for GeoBD design, it is necessary to extend UML with new elements that enable the modeling of spatial-temporal characteristics of geographical phenomena. UML is a naturally extensible language, in other words, it has its own constructs allowing its extension. The stereotype concept, one of the UML extension mechanisms, allows the definition of new specific model elements generating a profile tailored for a particular problem domain (OMG, 2007). There are some UML extensions for GeoDB modeling (Bédard et al., 2004; Borges et al., 2001; Lisboa Filho and Iochpe, 1999). To exemplify a spatial UML profile, described here is the Spatialtemporal UML-GeoFrame modeling language, which extends the UML, generating a profile of stereotypes to support the GeoDB conceptual modeling.GIS were originated outside the computer science field, unlike most software technologies developed in the last decades such as the operating systems based on windows, DBMS, fourth generation languages, CAD, CAM and CASE tools, Office Automation Systems (OIS), and more recently the World Wide Web (WWW) with the software revolution due to the explosion of internet use.One of the consequences of this historical origin is that most GIS application designers are their own users, who have the evolutionary approach as their main software developmental methodology, and whose main focus of attention is geospatial data acquisition and analysis. Thus, the old raster-vector debate (Couclelis, 1992) prevailed for a long time as an important theme in GIS conferences. Consequently, methodologies developed in the software engineering field are frequently not used in GIS application design, causing great losses in the quality of the produced systems and high maintenance costs.An alternative to reduce these problems is the use of a database design methodology. Consequently, during the 1990s, several extensions of specific conceptual modeling languages for GIS applications were proposed in the literature. Initially these modeling languages were based on the entity-relationship model (ER), proposed by Peter Chen (1976) or one of its extensions (e.g., Merise and Enhanced Entity-Relationship, EER). A few modeling languages were based on the semantic data model IFO (Abiteboul and Hull, 1987). At that time, the use of the object-oriented paradigm in system development was becoming more and more popular. Accordingly, several authors used as their base object-oriented design methods, such as OMT (Rumbaugh et al., 1991) and OOA (Coad and Yourdon, 1991), proposing extensions for the modeling of spatiotemporal aspects of geographical phenomena (Bédard et al., 2004; Borges et al., 2001; Lisboa Filho and Iochpe, 1999).With the aim of standardizing the different existent graphic notations and defining a basic group of model constructs for software systems, in 1996 three great experts on object-oriented modeling joined their approaches to create the UML (Booch et al., 1998). Consequently, by 1999 some UML extensions to GeoDB modeling came out, some of them supported by CASE tools (e.g., Perceptory Bédard et al. 2004, ArgoCASEGEO Lisboa Filho et al. 2004).The UML-GeoFrame modeling language (Lisboa Filho and Iochpe, 1999) is described here, to exemplify a spatial UML profile, and show how UML can be naturally extended by its own extension mechanism, named stereotype. Unlike other conceptual modeling languages that seek constructs’ completeness, so as to consider almost all modeling possibilities of geographical phenomena in different dimensions (descriptive, spatial and temporal), the UML-GeoFrame has as its inspiration the simplicity of the ER model and proposes the smallest possible group of stereotypes to assist the main requirements of GeoDB modeling, but at the same time allowing understanding by nonspecialized users, through a quite simple and instinctive graphic notation.A conceptual data modeling language provides a formal base (notational and semantics) for tools and techniques used in data modeling. Data modeling is the abstraction process where only the essential elements of the observed reality are emphasized, the nonessential elements being discarded. The process of conceptual database modeling comprises the description of the possible data content, besides structures and constraints applicable to them. This database description is based on the semantic constructs provided by a conceptual data modeling language.The UML-GeoFrame, originally presented in Lisboa Filho and Iochpe (1999), is based on a hierarchical class structure that makes up the conceptual GeoFrame framework. The GeoFrame provides the fundamental elements present in any GeoDB, whereas the UML class diagram provides the semantic constructs for a conceptual modeling language. This integration enables GeoDB design in a graphic language easily understandable by the users.The result of the modeling process is a conceptual data schema that expresses “what” will be stored in the database and not “how” the data will be stored. A conceptual data schema becomes therefore an abstraction of the real world that is being modeled (miniworld). Consequently, every element of the reality to be modeled in the conceptual data schema must be stored in the GeoDB. In the same way, every object stored in a GeoDB must have been represented in the conceptual data schema, but this does not often happen.



Search In Abstract Of Papers:
Other Papers In This Journal:

Search Result:

Help video to use 'Paper Search Console'