Previous Next Methods Allocation  
Annex 1  
5.6 Category of Methods "Estimation Models" (EMOD)  

  SMOD - Schätzmodelle

Contents  
  • 5.6.1 Overview
  • 5.6.1.1 Appraisal Criteria
  • 5.6.1.2 Operation Criteria
  • 5.6.1.3 Individual Methods
    •     
      
  • 5.6.2 Individual Descriptions
  • 5.6.2.1 Function Point Method (FPM)
  • 5.6.2.2 Constructive Cost Model (CoCoMo)

    • 5.6.1 Overview

    In the Methods Standard, method Review For the effort estimation required within the scope of the preliminary and detailed planning of the Project Management, the category of methods Estimation Models is defined in the Methods Standard. An individual method, a definite estimation procedure as e. g. the Function Point Method or CoCoMo is only selected within the scope of the operationalization of the Methods Standard.

    It is not possible to specify a definite procedure already within the Methods Standard. Nevertheless, in spite of the large number of estimation procedures exists none that might be suggested as the "standard method". Each one of the procedures has certain faults. Besides, the selection of a procedure is influenced by a variety of marginal conditions and factors. In addition the Methods Standard should be open to new developments at this point.

    Primarily, the category of methods Estimation Models comprises all procedures applicable for the effort estimation. Even if-in accordance with the appraisal criteria mentioned below-a procedure is basically suited for the application in the V-Model, it cannot as a rule be universally applied since effort estimation is particularly dependent on and influenced by external factors, such as the organizational and project-specific environment. Therefore, the operation criteria must be defined for each basically applicable estimation procedure.

    In this connection, the appraisal criteria are constant values referring to the estimation procedure itself. The operation criteria are variable values referring to the individual project-specific environment and are thus different in every project. Only after the results from the two investigation steps is it possible to select a certain estimation procedure for an actual project environment. And even then an adjustment to the software development process according to the V-Model is required for most of the remaining methods, provided they are not only supporting a macro estimation.

    5.6.1.1 Appraisal Criteria

    Next, the criteria are listed that can be used to appraise the basic suitability and usability of an estimation method in the V-Model/Project Management. The listed appraisal criteria should be used as a measuring device, in particular for new developments in this field.

    There exists hardly any estimation method able to meet all these criteria (it also contains some "desired criteria", in particular the last three items); in order to be classified as suitable for the application in the V-Model, though, the predominant part should be evaluated as positive.

    5.6.1.2 Operation Criteria

    Next, the criteria are listed that are used to define the application field for an estimation method since not all estimation methods are equally suited for each individual application.

    5.6.1.3 Individual Methods

    It is not possible to list all existing individual methods in this documentation. However, each estimation method can be traced to one or to a combination of basic types. According to /Noth, 86/, the basic types are: The two best known and-based on its versions and further developments-most frequently applied individual methods for the effort estimation are which are introduced next. Many of the effort estimation methods presented in literature and implemented in tools use one of these two approaches. Apart from that, also the less known method is considered as relevant; the utilization of an experience database is the core of this method and therefore permits interesting perspectives.

    5.6.2 Individual Descriptions

    5.6.2.1 Function Point Method (FPM)
    5.6.2.2 Constructive Cost Model (CoCoMo)

    Note:

    (1) INVAS /Noth, 1986/ was developed within the scope of a research project of Cologne University in cooperation with Siemens AG. INVAS is no completely defined method. It rather suggests a procedure which has to be specified in various ways for various development environments.

    INVAS separates the effort estimation into a quantity and into a productivity estimation. In this connection, the project is structured into subproducts and milestone results. These are evaluated according to their qualitative influence factors (e. g. system size, processing mode, degree of innovation, experience, technology) and special measuring quantities (e. g. requirements, functions, program lines, documentation pages, interfaces, test cases). After that the productivity based on the values stored in an experience database can be estimated. In this connection, subproducts with similar influence factors and measuring are utilized.

    Previous Next GDPA Online Last Updated 01.Jan.2002 Updated by Webmaster Last Revised 01.Jan.2002 Revised by Webmaster