CDIF Goals

The primary goal of CDIF is to let modelling tools understand each other. This includes interoperability between and among CASE tools, other related modelling tools and repositories.

In order to do that, CDIF defines an integrated meta-model (i.e. a common, integrated understanding of what the models are the tools are talking about), and ways to communicate models or specific aspects of models among software product, such as through a file format or bindings to distributed object technologies.

CDIF covers both the semantics and graphics of models, but they are not merged, so semantic concepts can have zero or more graphical representations, or there could be models which are graphics only.

CDIF focuses on the description of the information to be transferred, not on data management. CDIF presently does not cover configuration management or security enforcement, for example, although it may contain concepts to describe that data.

CDIF is a pragmatic technology dealing with Software and Systems Engineering techniques supported by commonly available CASE and related modeling tools (a somewhat loose interpretation of the term "CASE"). Some examples of these techniques are

• Data Flow Modeling (such as Gane/Sarson, Yourdon/DeMarco, but also Shlaer-Mellor)
• Data Modeling (as in common conceptual, logical and physical information modeling)
• State Event Modeling (as in Mealy/Moore, but also Harel and object-oriented techniques)
• Object-Oriented Modeling (as in OMT, Booch, UML, Shlaer-Mellor, OOIE etc.)
• Computer Aided Control Systems design (as in CACSD block diagrams)
• Project Management etc.

Because the CDIF Technical Committee is well aware that they will not be able to cover the entire life cycle of systems nor all flavors of all techniques in the standardized CDIF Integrated Meta-model, non-anticipated information can be represented in a CDIF transfer using Extensibility. Extensibility is a concept inherent to CDIF: Tool vendors can define their own extensions to CDIF and represent data that uses these extensions seamlessly with data that uses the standardized CDIF meta-model only.

CDIF Design Principles

CDIF strives to provide a representation for a given area of interest within Software Engineering / Systems Engineering adhering to the principle of "highest common denominator". This is a synthesis of elements in common use between tools covering a particular area. All these "areas" (in CDIF terms: Subject Areas) are in fact "views" of an integrated representation of modeling information (the CDIF Integrated Meta-model).

For example, the concept of an Attribute is defined only once but used in several Subject Areas, such as Object-oriented Analysis and Design, Data Modeling, Data Flow Modeling, State/Event Modeling. On a larger scope, State/Event Modeling (i.e. States, Transitions, Conditions, Actions, State decompositions etc.) is defined only once within the CDIF Integrated Meta-model, but used by tools doing SA/RT, or Object-oriented Analysis and Design. In CDIF, a tool would use the Data Flow Model Subject Area with the State/Event Modeling Subject Area to support SA/RT, versus the Object-oriented Analysis and Design (Core) Subject Area with the State/Event Modeling Subject Area to support an object-oriented technique.

CDIF Objectives

• To provide a precise, unambigous definition of information representation. This applies to both static file transfer and repository content definition.
  
• To support multiple physical representations of the same model. For example, in a file using a clear text representation (see: SYNTAX.1 and ENCODING.1), or, in a repository such as PCTE, or IRDS, or in a distributed-object context.
  
• To provide the importer with sufficient information to enable the importer to enable the reproduction of the transferred data consistent with the original sense.

CDIF attempts to preserve the highest level of fidelity of models transferred from one tool to another. This includes both semantic and presentation information.

CDIF's principle of Maximum Information Transfer specifies that

• Exporters shall be built without knowledge of any prospective destination CASE tool.
  
• An exporter shall transfer all available information within the intended scope of the transfer.
  
• Each importer may then discard any information for which it has no use.

When using the CDIF Integrated Meta-model as a schema for a repository, the principle of Maximum Information Transfer allows tools with different capabilities to work on the same data, without knowing which other tools are also involved.

Last modified: 97-05-09 by CDIF Webmaster.Copyright (C) 1997 EIA/CDIF. All rights reserved.
CDIF is a registered trademark of the Electronic Industries Association. All other trademarks are trademarks of their respective owners.