Database Reference
In-Depth Information
ing (Savinov, 2009a) which is intended to solve a wide spectrum of problems by reducing them to the
following three structural principles distinguishing it from other data models:
•
Duality principle
answers the question
how
elements exist by assuming that any element is a
couple
of one identity and one entity (called also reference and object, respectively)
•
Inclusion principle
answers the question
where
elements exist by postulating that any element is
included
in some domain (called also scope or context)
•
Order principle
answers the question what an element
is
, that is, how it is defined and what is its
meaning by assuming that all elements are partially ordered so that any element has a number of
greater and lesser elements
Formally, the concept-oriented model is described using a formalism of nested partially ordered sets.
The syntactic embodiment of this model is the concept-oriented query language (COQL). This language
reflects the principles of COM by introducing a novel data modeling construct, called
concept
(hence
the name of the approach), and two relations among concepts,
inclusion
and
partial order
. Concepts are
intended to generalize conventional classes and inclusion generalizes inheritance. Concepts and inclu-
sion are used also in a novel approach to programming, called concept-oriented programming (COP)
(Savinov, 2008, 2009b). Partial order relation among concepts is intended to represent data semantics
and is used for complex analytical tasks and reasoning about data.
The concept-oriented model and query language are aimed at solving several general problems which
are difficult to solve using traditional approaches. In particular, the following factors motivated this work:
•
Domain-specific identities.
In most existing data models elements are represented either by
plat-
form-specific
references like oids or by weak identities based on entity properties like primary
keys. These approaches do not provide a mechanism for defining strong
domain-specific
identi-
ties with arbitrary structure. Concepts solve this problem by making it possible to describe
both
identities and entities using only one common construct. This produces nice symmetry between
two branches: identity modeling and entity modeling.
•
Hierarchical address spaces.
Elements cannot exist outside of any space, domain or context but
existing data models do not support this abstraction as a core notion of the model. A typical solu-
tion consists in modeling spaces and containment like any other domain-specific relationship. The
principled solution proposed in COM is that all elements are supposed to exist within a hierarchy
where a parent is a space, context, scope or domain for its child elements. Thus inclusion relation
between concepts turns an element into a set of its child elements. Since identities of internal ele-
ments are defined relative to the space they are in, we simultaneously get a hierarchical address
space for the elements. Each element within this hierarchy is identified by a domain-specific hier-
archical address like a conventional postal address.
•
Multidimensionality.
Dimension is one of the fundamental constructs which is used to represent
information in various areas of human knowledge. There exist numerous approaches to multidi-
mensional modeling which are intended for analytical processing. The problem is that there exist
different models for analytical and transactional processing which rely on different assumptions
and techniques. The goal of COM in this context is to rethink dimensions as a first-class construct
of the data model which plays a primary role for describing both transactional and analytical
aspects. Data should be represented as originally existing in a multidimensional space and dimen-
sion should be used in most operations with data.
