During the past decade, much research has been performed in the areas of method engineering and process improvement. As a result of this research, we are developing the online method engine (OME). The OME is a knowledge management system that provides support during process improvement initiatives, using a set of assessments, an extensive method base, and automatic method assembly mechanisms. The system is designed around four main acivities: knowledge dissemination, method assessment, method improvement, and method enactment. In this paper, we demonstrate the core components and the main scenarios for the usage of the online method engine.
In an attempt to describe and reuse software development practices, several researchers
have proposed method bases to capture this knowledge [
The online method engine (OME) is the implementation of scientific concepts in the
field of method engineering (ME) and software process improvement (SPI), in
combination with concepts from the field of knowledge kanagement (KM)). The main goal
of the OME is to provide a system that facilitates the sharing of knowledge related
to method fragments for software development, assessing methods that are being used
in practice, improving methods based on assessment results, and adopting (improved)
methods in practice. The main philosophy behind the OME is that incremental (or
stepwise) improvement of processes reduces risks related to process change and improves
the change of success [
Following the four goals of the OME, the system is based on four layers of functionality; knowledge dissemination, method assessment, method improvement, and method enactment. Each layer increasingly relies on the previous ones, and all of the layers rely on a central method base, which forms the backbone of the system. The quality of the system is ensured by a constant process of validation and review.
Traditionally, method bases contain fragments of method knowledge, which reside on
the M2 level (according to the MOF framework) [
Complimentary to the situational indicators, the method base contains a set of capabilities that can be used to characterize a specific method implementation in a specific domain. The capabilities relate directly to a domain model, which captures the relevant process areas in a specific domain.
The situational indicators and the capabilities are related through a domain-dependent
mechanism called the situational factor effect. A situational factor effect describes how
a specific value of a situational indicator typically influences the relevance of a
capability [
Online Method Engine
Method Enactment Layer Method Improvement Layer Method Assessment Layer Knowledge Dissemination Layer
Process/method modeling
Workflow integration
Template generation
Data authoring Method fragment selection
Method fragment integration
Improvement roadmapping
Method base improvement Situational profiling
Maturity profiling
Maturity benchmarking Process/method administration
Method base search
Development Database - Operational data Method Base - Method fragments - Capabilities - Situational Factors - Assemby Rules - Situational Factor Effects - Experience
The final core element of the method base consists of a set of method fragments.
Method fragments describe a coherent piece of method knowledge that facilitates
reaching a specific goal during software development. A method fragment is described in
terms of activities and deliverables. Within the OME, method fragments are currently
modeled using process deliverable diagrams (PDDs) [
Scenario A: Method Assessment
In general, a process improvement effort starts with an assessment of the current
processes. The assessment approach employed within the OME is based on the
situational assessment method (SAM) [
The SAM is realized through two forms within the OME. The first form is used to
capture the organizational context. It consist of 24 questions spread out over 5 pages.
Each question has a short description, a set of answers, and possibly some help text to
indicate the type of answer expected. The second form is used to capture the current
capabilities. This form contains 68 questions, which are spread out over 4 pages. Each
page represents a business function, i.e. a layer from the SPM competence model [
Scenario B: Method Discovery
The results of the method assessment activity indicate the areas within the current process that are open for improvement. This is done in the form of a set of capabilities that are divided over various process areas, such as requirements prioritization or product roadmapping, and ordered based on an associated maturity level. This makes it possible to match these missing capabilities, or a subset of them, to existing methods that are stored in the method base.
The OME facilitates method discovery by providing tools to search the method
base based on several filters. These filters include the focus area, the business function,
the relevant capabilities, and keywords. The result is a set of method fragments that
fit these filters. Each result can be expanded into a detailed description of the method
fragment. The main components of these descriptions are a textual summary of the
method fragment, a detailed description that can contain both text and illustrations, a
list of the relevant situational factors and capabilities, a PDD, descriptions of the main
activities and deliverables, and a list of reference. An example excerpt of a method
description for the Echo approach [
Within the OME, the user is aided in this discovery process as much as possible. Elements within the PDD and items within the situational factor and capability lists are hyperlinks to more detailed descriptions. This makes it possible to obtain a quick overview, while allowing the user to go into more depth or to find related method fragments. 3.3
Scenario C: Method Improvement An important goal of the OME is to assist during a process improvement effort. In many cases, there are various solutions available to solve a specific problem, such as requirements prioritization. Not all of these solutions are applicable to any given situation, and it is hard for the method engineer to determine which solutions are likely optimal. The knowledge available within the method base allows for an automated approach to selecting appropriate method fragments. This selection can be based on the structural aspects of the method fragment (activities, deliverables), required capabilities, and situational factors.
The OME can propose a set of relevant method fragments based on the requirements of the user. It is possible to select method fragments that focus on one or more specific business functions, that implement capabilities up to a certain maturity level, or for the entire process. The selected fragments are presented to the user so that they can be reviewed using the tools described above (under method discovery). A prototype of the selection step is shown in Figure 4.
Once the user has selected the method fragments that he deems relevant, these fragments need to be assembled. In most cases, this means that they need to be integrated within the existing process. This activity can be partially performed automatically based on the structural aspects of the method fragments. In many cases, conflicts will arise during this assembly. For instance, multiple method fragments can include similar or incompatible deliverables. A report of these issues is presented to the user for further review.
The overarching goal of the OME is to allow for incremental improvement. This
is a very important characteristic, as the implementation of many process changes
at the same time is often unrealistic and unfeasible within an organization. Process
changes are always prone to resistance among employees, unforeseen complications,
and changes within the environment. Therefor, the system generates a series of
implementation plans [
The generated plans are presented to the user as a set of timelines. Each timeline contains the proposed increments with a summary of their contents. Once again. the user can review these changes to gain a detailed understanding of their contents and impact (see FIgure 5 for an example of these timelines). 3.4
Scenario D: Method Enactment The final activity within a process improvement effort, apart from the review, is the enactment of process changes. Although this is an activity that is mainly dependend on social and managerial aspects, it is possible to support parts of it through automated means. Within the context of the OME, enactment support focuses on the generation of templates and the automated migration of development tools. 3.5
Scenario E: Method Administration
All functionality within the OME is based on the contents of the method base. This
method base consists of method fragments, capabilities, situational factors, situational
factor effects, and experience reports. For the creation of method fragments, we employ
specialized tools instead of developing functionality within the OME itself. MetaEdit+
[
Both the conceptual design as well as the initial prototype of the OME have been
validated in earlier studies [
On the method improvement layer, we have realized partial planning functionality. This makes it possible to generate a set of improvement plans based on a goal method. However, more research is needed to incorporate the removal and replacement of method fragments, and to support more complex situations including improvement based on an existing method.
For the method enactment layer, no functionality has been implemented sofar. Techniques are currently under development to translate method changes into concrete enactment actions.