| Abstract As a community, we have begun to recognize that software cannot be produced with a standard technology, but needs to be developed with technologies tailored to the goals and characteristics of particular projects. Consequently, software engineering research needs to be performed in an experimental context that allows us to observe and experiment with the technologies in use, understand their weaknesses and strengths, tailor the technologies for the goals and characteristics of particular projects, and package them together with empirically gained experience to enhance their reuse potential in future projects. Several software engineering research groups have made the paradigm shift to an experimental, empirical software engineering view. The purpose of this network is to encourage and support the collaboration and exchange of results and personnel among these groups. Specific emphasis is placed on experimentation and empirical studies with development technologies in different environments; the repetition of experiments across environments; and the development and exchange of methods and tools for model building, experimentation, and assessment. The long-term expectation is that such cooperation will enable the abstraction and unification of environment-specific results and knowledge with the objective of generating the basic components of our discipline. The founding ISERN members chose the Quality Improvement Paradigm as the reference model to provide a common terminology for their cooperation. The QIP is an experimental framework for software development, based on the scientific method and instantiated in the TAME project at the University of Maryland. It views measurement as essential to the capture and effective reuse of software experience, and assumes the process is a variable based on the characteristics and goals of the project and organization. This framework views software engineering as a laboratory science which must be supported by the effective cooperation between academia and industry in order to achieve significant improvements. ISERN is open to other academic and industrial groups world-wide which are active in experimental software engineering research and are willing to adopt the experimental framework. There is no membership fee. The individual network members are responsible for funding collaboration through existing local or future joint grants. Purpose and Focus Software engineering is a relatively new and immature discipline when compared to the other engineering disciplines. As we mature, we have been evolving an experimental view of research allowing us to observe and experiment with technologies, understand their weaknesses and strengths, tailor technologies to the goals and characteristics of particular projects, and package them together with empirically gained evidence to enhance their reuse potential in future projects. Several academic and industrial research groups world-wide made the paradigm shift to an experimental software engineering view. Each of these groups have produced and will continue to produce software engineering models valid within their local laboratory environments. In order to take the next step towards building the basic models and components of the entire software engineering discipline, we have to be able to abstract from the characteristics of specific environments. No single research group is able to provide the laboratory environment necessary to learn about variations in the effects of technologies across multiple environments and influential factors. The purpose of ISERN is to provide a practical solution to this scale-up dilemma. Through various forms of joint activities the network intends to allow different research groups to learn from each other as well as elevate the maturity of experimental software engineering by providing access to a network of laboratory environments. Common Research Framework Within the TAME project at the University of Maryland, an evolutionary quality improvement paradigm (QIP)has been applied to foster the necessary maturing of software engineering technology. The term quality is used here in a broad sense referring to all attributes of software products and processes. It is, however, recognized that quality has a different meaning to different people involved in development/maintenance/use of software and needs to be defined precisely. The QIP assumes that software development is experimental and, therefore, needs to be conducted accordingly. It suggests that projects within an organization are based on the continuous iteration of characterization, goal setting, selection of improved technology, monitoring and analysis of its effects to correct projects on-line, post-mortem analysis to understand what could be done better in future projects, and packaging the newly learned lessons so they can be reused effectively in future projects. These activities can only be performed in an engineering-like manner if supported by measurement and performed in a laboratory-style environment. The ISERN framework is defined by the QIP and the following infrastructure technologies needed for its practical instantiation: TOP-DOWN MEASUREMENT: Software measurements must be tied to goals. The Goal / Question / Metric (GQM) paradigm has been proposed to support the definition of quantifiable goals and the interpretation of collected measurement data. EXPLICIT MODEL REPRESENTATION: Notations for explicit (process, product, quality) model representation are needed. Various approaches are currently being evaluated. REUSE ACROSS PROJECT BOUNDARIES: The Experience Factory concept has been proposed to support the organization of models for reuse across project boundaries Activities Joint ISERN activities focus on one or more of the following categories: 1. exchange of tools and/or people 2. further development of experimental infrastructure technologies for model building, experimentation and measurement/assessment approaches 3. joint empirical studies 4. definition and use of a common terminology 5. compilation and maintenance of a common annotated bibliography One can think of several reasons for joint ISERN experiments: 1. sharing of results across sites 2. performing families of experiments across sites 3. replicating specific experiments across sites 4. unifying results and knowledge whenever possible to generate the basic components of our discipline 5. reducing necessary resources in each organization (i.e., share cost) 6. identifying and understanding the effects of context variables The founding ISERN members agreed to adopt the TAME (GQM, QIP, Experience Factory) model as a common framework for joint activities. Communication ISERN provides several means for communication between the geographically distributed member sites: 1. Electronic Communication: An effective electronic exchange supports easy access to research results, reports, and prototype tools. 2. Working meetings: A yearly meeting to discuss results and revise future agendas. This meeting is not run as a workshop with refereed papers. Instead, it is run like an internal project meeting. The individual members will take turns in organizing this yearly meeting, currently in conjunction with Empirical Software Engineering International Week (ESEIW). 3. Exchange of researchers: It is anticipated that ISERN will ease the exchange of researchers for short-term visits as well as long-term postdoc- or sabbatical-style visits. The software engineering discipline is seeing an increasing number of postdoc appointments. We believe these appointments are important to the field; in an engineering discipline, a fresh PhD needs to gain experience in a different environment before being buried by the responsibilities of a regular professorial position. These appointments are difficult in a discipline like software engineering because of the overhead involved in adjusting, not only to the different culture in a foreign country, but to the different research frameworks and of different groups in a discipline as immature as ours. It is easier to adjust to any environment which shares a common research framework such as the one that will be provided by the ISERN collaborations. Members ISERN was initially created by: Nara Institute of Science and Technology (Japan), (contact: Prof. Dr. Koji Torii) University of Kaiserslautern, (Germany), FB Informatik, AG Software Engineering (contact: Prof. Dr. Dieter Rombach) University of Maryland at College Park (USA), Department of Computer Science (contact: Prof. Dr. V.R. Basili University of New South Wales, Sydney (Australia), School of Information Systems (contact: Prof. Dr. Ross Jeffery) University of Roma at Tor Vergata (Italy), Laboratory for Computer Science (contact: Prof. Dr. Giovanni Cantone) VTT Electronics, Technical Research Centre of Finland (contact: Dr. Markku Oivo now Univ. of Oulu, Finland) Numerous other academic and industrial organizations have joint ISERN since. A list of all current ISERN members is attached. Membership Application ISERN is open to other academic and industrial groups world-wide that are active in experimental software engineering research and willing to adopt the experimental research framework. There is no membership fee. ISERN members are pairs of (organization, contact person). Anyone affiliated with member organizations may participate in ISERN activities (see special rules for meetings). The contact person is responsible for ensuring adherence to the ISERN Research Framework by all participants from his/her organization. If the contact person leaves the organization, the organization must reapply for membership. Interested organizations may apply by sending an informal request to andreas jedlitschka at iese fraunhofer de. They will get the electronic application form asking them to describing their past experience in experimental software engineering research as well as their expectations from a future ISERN membership. The application will be evaluated by the ISERN Steering Committee. Membership is granted according to a 3-step procedure: 1. Attending as invited observer at an annual ISERN meeting. 2. Attending as an invited candidate at the following ISERN meeting (at this meeting the candidate gives a presentation; the candidate is supported by one senior member of ISERN nominated by the steering committee; all current ISERN members vote after the meeting per e-mail; each member has one vote; the candidate is admitted with 2/3 majority of all members). 3. Attending the following meetings as a full ISERN member. Membership is revoked if 3 out of 5 consecutive ISERN meetings have been missed, or no signs of cooperation (e.g. exchange of people/tools, papers with multiple authors, joint experiments) exist for 2 years. Meetings and Meeting participation ISERN meetings take place, once a year. In order to have efficient meetings the following participation rules have been defined: 1. For ISERN member organizations, max. two (2) persons are allowed to participate, the official ISERN contact and a second representative. If the offical ISERN contact is not participating, only one person is allowed to participate. This person has preferably participated in ISERN meetings before. 2. For ISERN applicants, either observers or candidates, only one person is allowed to represent the organization. In case of doubts or for exceptions, please contact the current program chairs. Benefits Benefits from Network Experimental software engineering research requires the cooperation of industry and academia. We anticipate the participation of industrial as well as academic groups in ISERN. The following sections list a number of benefits each type of participant could expect from involvement in ISERN. Benefits to Researchers Researchers in experimental software engineering frequently suffer from the lack of access to industrial data, lack of resources for replicating interesting experimental research results, reuse research results and tools from other environments, and exchange junior researchers in order to broaden their experience. Each academic ISERN member has its own research focus and local laboratory environment. However, the contributions towards improving models of our discipline are limited in such local environments. The necessary globalization requires cooperation with other groups. ISERN provides a common framework for such cooperation in the area of experimental software engineering research. Expected benefits are the lowering of hurdles for (a) communicating effectively, (b) reusing each others results (especially the experimental results from each others local laboratory environments and tools), (c) forming joint initiatives for replicating experiments, (d) submitting joint proposals, and (e) establishing a program for the exchange of researchers/developers (especially post-docs) across ISERN groups. Benefits to Companies Companies typically suffer from the inability to affect the direction of software engineering research and, therefore, have tremendous problems with identifying, tailoring, and transferring research results into practice. Each industrial ISERN member has its own interests. Expected benefits are the involvement of companies in the setting of research goals, the participation in experiments, the reception of technologies packaged with experience according to their goals, the delegation of personnel to research groups, and the hiring of highly experienced research personnel. |
Research, Innovate, Repeat. Together.