Experience Factory

Experience base

Knowledge management

Visual Query Interface

eWorkshop

 The Experience Factory (EF) approach defines a framework for building Learning Organizations. A concrete manifestation of Knowledge management concepts and tools, the EF content and structure are often referred to as the Experience base. FC-MD has developed support tools to manage content and structure and execute the needed procedures, as well as to help capture, store, integrate, analyze, synthesize and retrieve experience. The Visual Query Interface and eWorkshop are examples of tools that support the analysis and synthesis of organizational experience.

Technology Maturity

Goal Question Metric

Project Measurement and Data Collection

Empirical Studies

Model Building

Management by Data

 Technology Evaluation involves experimenting with various technologies to determine their maturity and suitability for use in specific environments. Any evaluation begins with setting study goals and objectives, and defining the context data and measures to be collected. Empirical studies can be performed as part of the evaluation process to quantitatively assess technologies and build models to enable better management and technical practices.

Assessments

Process Modeling

Risk Management

CMM(I)

Consortium Management

Training and Education

 FC-MD helps organizations to achieve their software process improvement goals through baseline assessments, process modeling, action planning assistance, periodic consulting support, and auditing services. Staff expertise in risk management and lessons learned for process improvement in small organizations and non-traditional software environments plays a significant role in the delivery of these services. Staff are certified by the Software Engineering Institute in performing Software Capability Evaluations and are experienced in assisting organizations to achieve compliance with the Capability Maturity Model® (CMM) and Capability Maturity Model-Integration® (CMMI).

 We also coordinate and manage several software-related consortia to provide a software engineering resource to member companies in advancing the practices of system and software engineering and improving the quality of their software-related products and services. These consortia integrate research and experience into practical improvement, create opportunities to develop and disseminate improvement practices, enhance the competitiveness of member companies, accelerate new software technology adaptation, leverage member company experience, promote inter-corporate cooperation of member organizations, and provide training and education.

Dependability

Reliability

Security

Architectural Evaluation

 There are many aspects to evaluating product quality. One approach is to model a system’s dependability from a user perspective. This involves defining a methodology for characterizing software systems’ dependability and an iterative approach for defining and quantifying dependability attributes, such as reliability. This methodology takes into account the multiple facets of dependability and the system’s different stakeholders. We are developing an experience base of models that describe, assess, and predict software’s dependability properties. 

 In the information security arena, there are many techniques that improve specific aspects of security, but there is little empirical evidence of their effectiveness with respect to various classes of vulnerabilities in different contexts. We are developing a cost-benefit function of how much security a given level of investment can provide.

 Product quality is also affected by design degeneration. Programmers working within time and budget constraints sometimes add and change code without fully understanding the architecture of the system they are modifying.  The changes that they make are often magnified when developers who were not part of the original design team further alter the system. Re-engineering is costly and time consuming and also delays the implementation of new features. Architecture evaluation is a form of defect reduction that addresses this problem, facilitating maintenance and evolution by detecting unnecessary complexity, incorrectly implemented software solutions, and dead code.  FC-MD’s process for architecture evaluation reveals architecture violations in a clear, systemic way, making it easer to address problems and incorporate implementations according to the original architecture. Our systematic process,  supported by a software tool, is flexible, cost-efficient and can be tailored to meet different levels of design constraints.

Reading/Inspections

Agile Technology

Using COTS in Development

Product Line Development

 Software inspections ensure that software artifacts, created during the software life cycle, possess the required quality characteristics. For instance, inspections improve design and code quality by increasing defect removal during development in a cost efficient manner, thus ensuring that the software artifacts necessary for its construction correctly reflect the needs of stakeholders.

 FC-MD has continued its work on the research and application of “software reading techniques,” which increase the effectiveness of software inspections by providing guidelines that inspectors can use to examine (or “read”) a given software artifact and identify defects. Empirical evidence demonstrates that software reading is a promising technique for increasing software quality for different situations and document types and is not just limited to source code. Software reading can be performed on all documents associated with the software process and is an especially useful method for detecting defects, since it can be applied as soon as documents are written. FC-MD is engaged in a number of collaborations for the purpose of refining reading techniques for different stages of the life cycle.

 Agile Software Development Methods are software development practices designed to efficiently produce software and reduce overhead costs. Although interest in Agile Methods is increasing, very little empirical evidence exists to support anecdotal evidence about its usefulness and effectiveness. FC-MD collaborates with experts and practitioners to characterize Agile practices by

•   conducting a series of eWorkshops.

•   designing and conducting experiments.

•   developing a measurement framework to help practitioners and experimenters collect metrics and better understand these practices.

 The COTS Lesson Learned Repository (CLLR) is part of the Center for Empirically-Based Software Engineering (CeBASE) and provides support for the software development community in using COTS products.  The CLLR allows practitioners to share insights and solutions that can reduce risk and improve the industry’s quality and productivity. It contains descriptions of the lesson (summary, risk, or issue addressed, type of data supplied—qualitative or quantitative), recommended audience (program or project manager, developer), and the story behind the lesson. The context where the lesson was learned is also incorporated, with reference to details regarding the type of system, organization, number and types of COTS products, and life cycle plans. Many of the lessons-learned are also relevant for product-line developers.

Project Management

Software Acquisition

System Development

Requirements Management

Configuration Management

Dynamic Simulation

GUI Testing

Web Development

 FC-MD supports all aspects of software engineering activities and processes, from system acquisition and project management  to development and maintenance, including requirements management and configuration management.

 FC-MD’s modeling capabilities feature both static and dynamic simulation models which are calibrated specifically to meet the needs of different organizations and their projects, taking into account their specific problems, questions, and decisions. By focusing on relevant variables that must be monitored and controlled, FC-MD identifies significant factors inherent in the development process and tailors metrics collections to improve tracking and planning.

 FC-MD employs the GUITAR model (Graphical User Interface Testing Framework) to automate some of the crucial aspects of testing a GUI more logically and thoroughly than current tools that employ capture and relay functionality. GUITAR enables developers to create ‘projects’ for each GUI application under test, consisting of artifacts (files/folders) that are produced or consumed by tools and implemented as GUITAR’s plug-ins.  The test designer can implement new tools and integrate them with GUITAR via plug-ins. FC-MD can define and set up a test environment for your GUI-based software products.

© 1999 - 2008 Fraunhofer Center MD
For suggestions and comments, please contact webmaster@fc-md.umd.edu