Stanley M. Sutton, Jr. and Leon J. Osterweil. The design of a next-generation process language. In Proceedings of the Joint 6th European Software Engineering Conference and the 5th ACM SIGSOFT Symposium on the Foundations of Software Engineering, pages 142--158. SpringerVerlag, 1997. 7  Home/Search   Document Details and Download   Summary   Related Articles   Check

....and the lack of flexibility in its enactment. The purpose of this report is to present the most important features of a PML called PROMENADE (PROcess oriented Modelling and ENActment of software DEvelopments) intended to be a part of the second generation of PMLs (as it has been named in [SO97]) and to make some contributions to the improvement of the referred issues. Remarkably, PROMENADE is based on UML notation, making it more attractive for software engineers (not another formalism to be learned) Section 2 shows the reasons for which PROMENADE has focused on those features. ....

Sutton, S.M.; Osterweil, L.J.: The Design of a Next-Generation Process Language. Proceedings of ESEC/FSE '97, Lecture Notes in Computer Science, Vol. 1301, M. Jazayeri and H. Schaure (eds.). Springer-Verlag, Berlin Heidelberg New York (1997), 142-158.

....fragments. 4 Process Modeling in UML To increase model understandability and to decrease model maintenance effort, process modeling should be supported on a very high level of abstraction. In contrast to other approaches introducing a special purpose process modeling language (EPOS [10] JIL [25], MADAM [21] this section will demonstrate how process models can be adequately defined in a standardized object oriented modeling language, the UML. 4,1 Structural Modeling The process specification models processes on the type level. Resulting process schemas thus abstract from a multitude of ....

S. Sutton and L. Osterweil. The design of a next generation process language. In M. Jaayeri and H. Schauer, editors, Proc. ESEC '97 LNCS 1301, pages 142-158, Zfirich, Switzerland, Sept. 1997.

....recent research efforts, some of which are discussed below. We explicitly acknowledge that this is not an exhaustive survey of the field, but rather a sampler of some representative research efforts. 4.1. 1 JIL Developed by the LASER research group at University of Massachusset at Amherst, JIL [70] is the result of Osterweil s reflection on the successes and failures of first generation research on process programming and PSEEs. It is an interesting and ambitious effort aiming at providing a semantically rich language, which features high level, process specific constructs. This research ....

S. M. Sutton Jr. and L. J. Osterweil, "The Design of a Next-Generation Process Language". In Proceedings of the Fifth ACM SIGSOFT Symposium on the Foundations of Software Engineering, Zurich, Switzerland, September 1997. LNCS 1301, pp. 142-158. - 26 -

....to system architecture. Most of the PSEEs developed during the last years adopts a standard client server architecture. Process enactment is centralized and tools communicate with the engine through point to point connections. Examples of such PSEEs are Adele [8] Arcadia [26] EPOS [11] JIL [25], Marvel [21] Merlin [23] Oikos [1] Process Weaver [19] Provence [7] SENTINEL [15] and SPADE [6, 5] Endeavors [9] adopts a complex architecture to distribute process enactment. Several process engines may coexist. They communicate with standard point to point connections. As mentioned, ....

M. S. Sutton and J. L. Osterweil. The design of a nextgeneration process language. In Proceedings of the Fifth ACM SIGSOFT Symposium on the Foundations of Software Engineering, number 1301 in Lecture Notes in Computer Science, pages 142--158, Zurich, Switzerland, Sep 1997. Springer-Verlag.

....systems for supporting automation of the process up to an acceptable level. This topic has drawn a special attention within the scientific community and, as a result, several PMLs have been developed (see [6] and [5] for a survey) Currently a second generation of PMLs is coming into existence ([22], 23] trying to fix some common drawbacks of most of former approaches: difficulty to express complex processes and to understand the resulting model; non visual models or too nave visual ones; difficulty to formalize the many facets of the process; use of one single language paradigm; etc. ....

.... than reacting to the rising of certain events (basically, task finalization) This latter fact is quite an important restriction since one of the features that have been recognised to be important for a second generation PML is its ability to combine both proactive and reactive control paradigms [22], in order to get less prescriptive and more expressive process models. Activity properties cannot be shown in activity diagrams. For instance, activity deadlines and duration, roles related to an activity (who should be informed about it. used resources and tools. Needless to say that ....

[Article contains additional citation context not shown here]

Sutton, S.M.; Osterweil, L.J.: The Design of a Next-Generation Process Language. Proceedings of ESEC/FSE '97, Lecture Notes in Computer Science, Vol. 1301, M. Jazayeri and H. Schaure (eds.). Springer-Verlag, Berlin Heidelberg New York (1997), 142-158

....Most of the most known PMLs we are aware of use just one single control paradigm for modeling the dynamic part of the process. Whether it is proactive or reactive (event driven) neither of them seem expressive enough to address the specificities of a real software process. Some PMLs, like JIL[SO97], provide both paradigms but the proactive one behaves in an imperative way which leads in practice to an event driven control. JH99] recognizes the necessity of a more flexible and high level modelling and therefore defines a control flow allowing different kinds of task dependencies (although ....

....of those relationships that can only be known at enaction time. This latter feature helps very much in the modeling of fine grained processes. 4 Modularity Although the need of a modular model construction when dealing with complex software processes is widely recognized (see [ARF97] and [SO97] for example) most of the approaches we are aware of deal with this aspect by relying on the modularity features of the paradigm they use (mostly object orientation) Using this approach, modularity in the construction of a new model is achieved mainly by the reuse of some already constructed ....

Sutton, S.M.; Osterweil, L.J.: The Design of a Next-Generation Process Language. Proceedings of ESEC/FSE '97, Lecture Notes in Computer Science, Vol. 1301, M. Jazayeri and H. Schaure (eds.). Springer-Verlag, Berlin Heidelberg New York (1997), 142-158.

.... responsibility [6] 67] 68] and hybrid models such as [71] There is also recent work, in this case motivated by modeling of complex software processes, on an agent coordination language that also shows much promise for being able to specify and implement complex agent interaction patterns [72]. Multiagent research has long been divided into two camps, one concerned with cooperative (benevolent) agents and the other concerned with self interested agents [14] There has been very little cross fertilization of ideas between these camps. Research on self interested agents is often based ....

# S.M. Sutton Jr. and L.J. Osterweil, "The Design of a NextGeneration Process Language," Proc. Joint Sixth European Software Eng. Conf. and Fifth ACM SIGSOFT Symp. Foundations of Software Eng., Springer-Verlag, Zurich, 1997.

....uniformly for any types of workflows. Since a posteriori flexibility requires human intervention, it implies an additional and very crucial design goal: a workflow modeling language on a high level of abstraction is needed which is easy to use and supports the visualization of its elements (cf. [SuOs97]) In particular, the trade off between high level formalisms, such as graph based modeling approaches, and low level mechanisms, such as rule based specifications which are hard to understand for humans and difficult to analyze but provide a great flexibility, has to be resolved. Beside ....

Sutton Jr., S.M.; Osterweil, L.J.: "The Design of a Next-Generation Process Language", in Jazayeri, M; Schauer, H (eds.), Software Engi- neering - ESEC/FSE'97, Proceedings, LNCS 1301, Springer, 1997; pp. 142-158.

....carrying out a process specified in Paddle involves executing the corresponding program, one disadvantage of this procedural representation of process knowledge is that it enforces a strict depth first left to right processing of the goal structure. The second generation process language JIL [42] is more flexible in this respect: not only proactive (i.e. fixed) but also reactive control specifications can be given. Representing processes as programs has the advantage that processes are formal objects on which reasoning can be performed. On the other hand, humans find it easier to follow ....

Stanley Sutton and Leon Osterweil. The design of a next-generation process language. In M. Jazayeri and H. Schauer, editors, Proceedings of the Fifth ACM SIGSOFT Symposium on the Foundations of Software Engineering, LNCS 1301, pages 142--158. Springer-Verlag, 1997.

....of software and that there are considerable benefits that will derive from basing a discipline of process development on the more traditional discipline of software development. As a consequence, he proposes process modeling languages based on the procedural programming language paradigm 6 Part I [Ost87, SO97]. Since processes involve a lot of creativity the term programming does not seem to be adequate to describe the development of a process. Osterweil s paper provoked many reactions expressing the unhappiness about the idea that a software organization can program its development process. This ....

....Arcadia research project focuses on large, complex software systems, particularly those with high reliability requirements. The used process modeling language APPL A bases as an extension of Ada on the procedural programming language paradigm. Experiences from this project are considered in JIL [SO97] announced as a next generation process language. The Adele 2 [BEM91] project uses a combination of Introduction 7 object oriented and rule based approach. These projects are all quite large, focus mainly on the process issues in software engineering and deliver powerful but rather complex and ....

Sutton, S., Osterweil, L.:"The Design of a Next-Generation Process Language", CMPSCI Tech. Rep. 96-30, January 1997

....from first generation languages now need to be consolidated and exploited through the development of new process languages. Here the chief difference seems to be between those who believe the most promising approach is a set of sub languages which can be factored together as and when required [23], and those like ourselves who are concentrating on a better core language [15] One common theme is the issue of managing concurrency. suspend activity and extract a script suspendedwc : decompose ( wcactivity ) suspendedChecker : suspendedwc.scriptvector[checkerIndex] revise script ....

.... : revisedChecker; restart the suspended activity wcactivity : activate( suspendedwc ) Figure 7 An example of stopping, modifying and restarting an activity using decompose Little JIL [21] is a sub language of the second generation process language JIL [23] which concentrates on the coordination of activities and agents. It has a visual syntax and is aimed at making it easier for practitioners to experiment with process programs. This emphasis on coordination is closely related to our view of processes as sets of mediated collaborations. In both it ....

Sutton Jr., S.M. and Osterweil, L.J. "The Design of a Next-Generation Process Language" in Proceedings of the Joint Sixth European Software Engineering Conference and the Fifth ACM SIGSOFT Symposium on the Foundations of Software Engineering, pp 142--158, Zurich, Springer-Verlag LNCS 1301, 1997.

....problem solving processes. TAEMS models are used in multi agent coordination research [5, 31] and are being used in many other research projects, including: cooperative information gathering [23] collaborative distributed design [7] intelligent environments [20] coordination of software process [9], and others [4, 29, 13] Typically a problem solver represents domain problem solving actions in TAEMS, possibly at some level of abstraction, and then passes the TAEMS models on to agent control problem solvers like the Design to Criteria scheduler. TAEMS models are hierarchical abstractions ....

Stanley M. Sutton Jr. and Leon J. Osterweil. The Design of a Next Generation Process Language. Umass CS TR-96-35.

.... projects, including: Cooperative InformationGathering [27, 25] collaborative distributed design [12] distributed situation assessment [6] surviveable systems [35] multi agent diagnoses [3] intelligent environments [24] hospital patient scheduling [10] and coordination of software process [22]. Typically a problem solver represents domain problem solving actions in TMS, possibly at some level of abstraction, and then passes the TMS models on to agent control problem solvers like the multi agent coordination modules or the Design to Criteria scheduler. 4 TMS models are hierarchical ....

S. Sutton Jr. and L. Osterweil. The design of a next-generation process language. In Proceedings of the Fifth ACM SIGSOFT Symposium on the Foundations of Software Engineering, pages 142--158, September 1997.

.... that a process model must move towards an ever more complete description of the organization that performs it; this requires a rich representation formalism that can capture the sequence of activities, the products or resources to which they apply, the human or software agents that perform them [26], and a myriad of relationships that link them. This can lead to powerful modeling notations, but unfortunately also to complicated models that are dicult to construct and understand. It also leads to self contained execution environments that assume all activities and artifacts are part of the ....

Stanley Sutton and Leon Osterweil. The design of a next-generation process language. In M. Jazayeri and H. Schaure, editors, Proceedings of ESEC/FSE '97, 1997.

....of correctness of a model enaction. The existence of such formal basis would provide a well established foundation to reason about model enaction. For these and other reasons (e.g. integration of different paradigms) a second generation of PMLs is actually coming into existence [WBG 98] as JIL [SO97] and others. Intended to be part of this generation, we present here a PML aimed at supporting these properties. The language is the kernel of our PROMENADE approach (PROcess oriented Modellization and ENAction of software DEvelopments) currently in progress; we call the language PROMENADE too. ....

Sutton, S.M.; Osterweil, L.J.: The Design of a Next-Generation Process Language. Proceedings of ESEC/FSE '97, Lecture Notes in Computer Science, Vol. 1301, M. Jazayeri and H. Schaure (eds.). Springer-Verlag, Berlin Heidelberg New York (1997), 142-158

.... joint responsibility [6, 67, 68] and hybrid models such as [71] There is also recent work, in this case motivated by modeling of complex software processes, on an agent coordination language that also shows much promise for being able to specify and implement complex agent interaction patterns [72]. Multiagent research has long been divided into two camps, one concerned with cooperative (benevolent) agents and the other concerned with self interested agents [14] There has been very little cross fertilization of ideas between these camps. Research on self interested agents is often ....

S. M. Sutton, Jr. and L. J. Osterweil, "The Design of a Next-Generation Process Language," Proc. Joint 6th European Software Eng. Conf. and 5th ACM SIGSOFT Symp. Foundations of Software Eng., Springer-Verlag: Zurich, Switzerland, 1997.

....to system architecture. Most of the PSEEs developed during the last years adopts a standard client server architecture. Process enactment is centralized and tools communicate with the engine through point to point connections. Examples of such PSEEs are Adele [18] Arcadia [19] EPOS [20] JIL [21], Marvel [22] Merlin [23] Oikos [24] Process Weaver [25] Provence [26] SENTINEL [27] and SPADE [28, 29] OZWeb uses standard web technologies to support distributed processes. The client server approach it adopts is quite traditional with respect to the PROSYT architecture. Endeavors [30] ....

M. S. Sutton and J. L. Osterweil, "The design of a next-generation process language," in Proceedings of the Fifth ACM SIGSOFT Symposium on the Foundations of Software Engineering, no. 1301 in Lecture Notes in Computer Science, (Zurich, Switzerland), pp. 142--158, Springer-Verlag, Sep 1997.

....Our experience is that it is not unusual for almost half the PWI PML code for a model to be dealing with general concurrecy issues rather than the specifics of the process involved. 4 B.C. Warboys et al. Recently Sutton and Osterweil have been working on a second generation process language [14], based on their extensive experience using APPL A. They identify a number of issues which they believe are key to second generation process languages: semantic richness, ease of use, appropriate abstractions, process composability, visualisation, and multiple paradigms. In their language JIL, the ....

S.M. Sutton Jr. and L.J. Osterweil.: The Design of a Next-Generation Process Language. Technical Report CMPSCI Technical Report 96-30, University of Massachusetts, (1997). in Proceeding of the Fifth ACM SIGSOFT Symposium on the Foundations of Software Engineering, Zurich (LNCS 1301).

....obtained directly from several Pleiades users. The client applications about which we obtained information were a reusable components library, the Arcadia language processing tool set, TAOS (Testing with Analysis and Oracle Support) 29] the Booch Object Oriented Design process program (BOOD) [34], FLAVERS (Flow Analysis and VERification System) 12] an agenda management system, and an avionics validation and verification system [21] The process we used to perform the evaluation was as follows. We constructed a questionnaire that included approximately fifty questions. The questions ....

S. M. Sutton, Jr. and L. J. Osterweil. The design of a next-generation process language. In Proc. 5th Symp. on the Foundations of Software Engineering, Sep 1997. (To appear.).

....network bandwidth, where the performance of an action using the resource may degrade based on the state of the resource and the degrees of degradation vary. plication domains. For example, in a current project we are interfacing our agent control technology with a higher level process view [20] of the task of sending robot teams into hazardous environments to perform unmanned exploration (e.g. damaged buildings to access structural conditions) This application requires different protocols and different behaviors than applications such as the coordination of agents in an intelligent ....

.... projects, including: cooperative information gathering [27, 26] collaborative distributed design [11] distributed situation assessment [4] survivable systems [34] multi agent diagnoses [3] intelligent environments [24] hospital patient scheduling [8] and coordination of software process [20]. Typically a problem solver represents domain problem solving actions in TAEMS, possibly at some level of abstraction, and then passes the TAEMS models on to agent control problem solvers like the multi agent coordination modules or the Design to Criteria scheduler. 2 TAEMS models are ....

Stanley M. Sutton Jr. and Leon J. Osterweil. The design of a next-generation process language. In Proceedings of the Fifth ACM SIGSOFT Symposium on the Foundations of Software Engineering, pages 142--158, September 1997.

....enacting concepts are at the center of workflow management. Support for heterogeneous processes (human centered and system centered) flexibility, reuse, and distribution are important challenges for the design of the nextgeneration process modeling languages and their enactment mechanisms (cf. [8, 25]) In particular, flexible and collaborative processes require human intervention. Therefore, a process modeling language on a high level of abstraction is needed which is easy to use and supports the visualization of its elements. In particular, the trade off between high level formalisms, such ....

....All of them have to be integrated within a single process model. Thus, for adequate process modeling support, different modeling paradigms (net based and rule based, proactive and reactive control flow specification, etc. have to be taken into account and have to be integrated (cf. [20, 25]) Flexibility: Flexibility of a WFMS comprises two fundamental aspects: 1) The specification of a flexible execution behavior to express an accurate and less restrictive behavior in advance: flexible and adaptable control and data flow mechanisms have to be taken into account in order to ....

[Article contains additional citation context not shown here]

Sutton Jr., S.M.; Osterweil, L.J.: "The Design of a Next-Generation Process Language", in Jazayeri, M; Schauer, H (eds.), Software Engineering - ESEC/FSE'97, Proceedings, LNCS 1301, Springer, 1997; pp. 142-158.

....its structure and its behavior. To increase model understandability and to decrease model maintainance effort, process specification should be supported on a very high level of abstraction. In contrast to other approaches introducing a special purpose process modeling language (EPOS [9] JIL [21], MADAM [19] this section will demonstrate how process models can be adequately defined in a standardized object oriented modeling language, the UML. The transition from analysis to specification will be discussed at the end of this section. Structural Modeling The process specification models ....

S. Sutton and L. Osterweil. The design of a next generation process language. In M. Jazayeri and H. Schauer, editors, Proc. ESEC `97, LNCS 1301, pages 142--158, Zurich, Switzerland, Sept. 1997.

....enacting concepts are at the center of workflow management. Support for heterogeneous processes (human centered and system centered) flexibility, reuse, and distribution are important challenges for the design of the next generation process modeling languages and their enactment mechanisms (cf. [ElNu96, SuOs97]) In particular, flexible and collaborative processes require human intervention. Therefore, a process modeling language on a high level of abstraction is needed which is easy to use, which supports the visualization of its elements, and which allows for the adequate and participatory design of ....

.... system requirements which are not satisfied by most commercial WFMS [GHS95] Ease of use: A very crucial design goal is, that the language is easy to use, i.e. it should allow for modeling of processes at a high level of abstraction, and it should support the visualization of its elements (cf. [SuOs97, DEA98]) this is naturally given in a net based approach; on the other hand, rule based specifications are hard to understand for people, but provide a great flexibility) Of course, sufficient expressive power of a language is a basic requirement, but without sufficient adequacy a language will be of ....

[Article contains additional citation context not shown here]

Sutton Jr., S.M.; Osterweil, L.J.: "The Design of a Next-Generation Process Language", in Jazayeri, M; Schauer, H (eds.), Software Engineering - ESEC/FSE'97, Proceedings, LNCS 1301, Springer, 1997; pp. 142-158.

No context found.

Stanley M. Sutton, Jr. and Leon J. Osterweil. The design of a next-generation process language. In Proceedings of the Joint 6th European Software Engineering Conference and the 5th ACM SIGSOFT Symposium on the Foundations of Software Engineering, pages 142--158. SpringerVerlag, 1997. 7

....process development, deployment, and use. The incorporation of developer roles into the development life cycle is part of the development model for EJBs. The incorporation of various stakeholders in various roles in the development cycle is also advocated for so called factored process languages (Sutton Osterweil, 1997). 5. Plans We are working on the design and implementation of a transactional business process server, and our first results are very promising. Our TBPS has been implemented as a set of Java classes (JavaBeans) to run on IBM s WebSphere middleware, assuring portability and middleware ....

S. Sutton, Jr. and L. J. Osterweil. The Design of a Next-Generation Process Language. In Proceedings of the Joint 6th European Software Engineering Conference and the 5th ACM SIGSOFT Symposium on the Foundations of Software Engineering, Springer-Verlag, Berlin, Heidelberg, pp 142-158, 1997.

....a process language that attempts to resolve the apparently conflicting objectives of supporting this wide variety of abstractions and creating a language that is easy to use and understandable by non programmers. Little JIL is strongly rooted in our past research on process programming languages [27, 28], but it makes some important breaks with this earlier work. Of primary importance for this paper is the focus on the coordination of activities and agents. Coordination, as defined by Carriero and Gelernter is the process of building programs by gluing together active pieces and is a vehicle ....

....the familiar problem of trip planning. While this process is not as complex as the software engineering processes for which Little JIL is designed, it serves as an effective vehicle for demonstrating the language features. 2. Design Principles Little JIL draws on the lessons of our earlier work [28] by retaining the step as the central abstraction and scoping mechanism but refines the features in terms of which a step is defined. The design of Little JIL features was guided by three primary principles: Simplicity: To foster clarity, ease of use, and understandability, we made a concerted ....

S. M. Sutton, Jr. and L. J. Osterweil. The design of a nextgeneration process language. In Proc. of the Joint 6th European Software Engineering Conf. and the 5th ACM SIGSOFT Symp. on the Foundations of Software Engineering, pages 142--158. Springer-Verlag, 1997.

....the execution engine from having to implement the semantics required to define those details. Some recent process programming languages have carried this idea still further, by demonstrating modular approaches to incorporating broader semantics while reducing execution platform complications [4, 19, 17]. In programs written in a modular language, different types of language semantics are addressed by different language modules, with the runtime behavior of the process realized through the coordination of separate components. For example, a process language might modularize the description of ....

....that are separate from each other and from details of execution sequencing and concurrency. It has been argued that this approach benefits users of the language because they can comprehend their processes more readily by being able to focus their attention separately on different narrower aspects [17]. 2 This approach has the additional advantage of allowing the different semantic modules of the language to be defined in different languages, each of which can presumably be selected to be more appropriate for the particular semantic feature. This modularization of the process description ....

S. M. Sutton, Jr. and L. J. Osterweil. The design of a next-generation process language. In Proc. of the Joint 6th European Software Engineering Conf. and the 5th ACM SIGSOFT Symp. on the Foundations of Software Engineering, pages 142--158. Springer-Verlag, 1997. Zurich, Switzerland.

....activities, artifacts, resources, events, agents, and exceptions, which can easily make a language large and complex, and creating a language that is easy to use and understandable by non programmers. Little JIL is strongly rooted in our past research on process programming languages [25, 26], but it makes some important breaks with this earlier work. Of primary importance for this paper is the focus on the coordination of activities and agents and the premise of process language factoring. Process language factoring is the separation of various semantic elements of a process so that ....

....extending a conventional programming language with process motivated constructs (APPL A [25] This work suggested that it would be preferable to develop a new special purpose, high level language designed specifically for process programming. This new language, JIL, has been described elsewhere [26]. Preliminary evaluation of JIL has suggested: 1) the value of high level, process oriented semantics, 2) the appropriateness of the step as a central abstraction, 3) the use of the step construct as a scoping construct for other features, and 4) the possibility of a factored language design. ....

S. M. Sutton, Jr. and L. J. Osterweil. The design of a nextgeneration process language. In Proc. of the Joint 6th European Software Engineering Conf. and the 5th ACM SIGSOFT Symp. on the Foundations of Software Engineering, pages 142--158. Springer-Verlag, 1997.

....these ways. A major obstacle in presenting demonstrably effective processes has been the lack of languages that are adequate vehicles for this. In our past work we have presented languages designed to support the clear and precise exposition of process. Most recently we have proposed that the JIL [16] process programming language incorporates a mix of features and abstractions that seem to offer promise of supporting the clear and precise exposition of complex This work was supported by the Advanced Research Projects Agency under grant F30602 94 C 0137 and F30602 97 2 0032. software process ....

....progress of the design process, to assist in the assignment and reassignment of steps to agents, both human and automated, and to manage contention for resources among the agents. The value of these capabilities increases with more complex designs and larger design teams. 3 Overview of JIL JIL [16] is a process programming language intended to support the development of high level process abstractions through a collection of powerful and easy to use semantic constructs. JIL represents processes as compositions of steps. The specification of a step is defined in terms of a number of ....

[Article contains additional citation context not shown here]

Stanley M. Sutton, Jr. and Leon J. Osterweil. The design of a next-generation process language. In Proceedings of the Joint 6th European Software Engineering Conference and the 5th ACM SIGSOFT Symposium on the Foundations of Software Engineering. Springer-Verlag, 1997. To appear.

....practical utility of simplified languages has been limited. In this paper, we present Little JIL, a language that attempts to resolve these two apparently conflicting objectives, semantic richness and ease of use. Little JIL is strongly rooted in our past research on process programming languages [19, 20], but it makes some important breaks with this earlier work. Of primary importance for this paper is the focus on the coordination of activities and agents and the premise of process language factoring. Process language factoring is the separation and potentially independent treatment of various ....

....extending a conventional programming language with process motivated extensions (APPL A [19] This work suggested that it would be preferable to develop a new special purpose, high level language designed specifically for process programming. This new language, JIL, has been described elsewhere [20]. Preliminary evaluation of JIL has suggested: 1) the value of high level, process oriented semantics, 2) the appropriateness of the step as a central abstraction, 3) the use of the step construct as a scoping construct for other features, and 4) the possibility of a factored language design. ....

[Article contains additional citation context not shown here]

Stanley M. Sutton, Jr. and Leon J. Osterweil. The design of a next-generation process language. In Proc. of the Joint 6th European Software Engineering Conf. and the 5th ACM SIGSOFT Symp. on the Foundations of Software Engineering, pages 142--158. SpringerVerlag, 1997.

....we have found that it is quite beneficial to also reduce it to operational practice. We have done this by developing a formal process definition for a specific software design method, namely Booch Object Oriented Design [1] expressing the process definition in the JIL process programming language [5]. This process definition work has had a number of significant benefits. First, because defining a process by means of a programming language provides a precise semantics, we are able to be quite rigorous in the definition of the method. This enables us to provide details of the method that were ....

Stanley M. Sutton, Jr. and Leon J. Osterweil. The design of a next-generation process language. In Proceedings of the Joint 6th European Software Engineering Conference and the 5th ACM SIGSOFT Symposium on the Foundations of Software Engineering. Springer-Verlag, 1997. To appear.

....adaptation. 3.2 Representing a Linear Regression Process In this section we illustrate the linear regression process using the Little JIL process language. Little JIL is a visual language derived from a subset of JIL, a process language originally developed for software development processes [22]. Little JIL focuses on coordination of agents in the performance of process activities in a wide range of processes. Little JIL represents the activities of a process as steps, where each step can be decomposed into substeps. Substeps within a step can be invoked either proactively or reactively. ....

Sutton, Jr., S.M. and Osterweil, L.J. The Design of a Next-Generation Process Language, in Proc. of the Joint 6th European Softw. Engg. Conf. and the 5th ACM SIGSOFT Symp. on the Foundations of Softw. Engg. (1997), Springer-Verlag, 142-158.

....able to support significant dynamism and user adaptability, though we have not yet experimented with these capabilities. 4. 2 Evaluation of a Generated AMS The Grapevine prototype was evaluated by using it to generate an AMS to support the execution of process programs written in a subset of JIL [11]. JIL programs are executed by human and software agents, and we coordinated these agents with a Grapevine generated AMS. The utility of the created AMS was evaluated in the context of execution of a rudimentary process program for a phase of the Booch Object Oriented design process. In this ....

Stanley M. Sutton Jr. and Leon J. Osterweil. The Design of a Next-Generation Process Language. To appear in the Proceedings of the Fifth annual conference on the Foundations of Software Engineering.

No context found.

S.M. Sutton Jr. and L.J. Osterweil.: The Design of a Next-Generation Process Language. Technical Report CMPSCI Technical Report 96-30, University of Massachusetts, #1997#. in Proceeding of the Fifth ACM SIGSOFT Symposium on the Foundations of Software Engineering, Zurich #LNCS 1301#.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC