Karen Huff and Viktor Lesser. A plan--based intelligent assistant that supports the software development process. In Peter Henderson, editor, rd ACM SIGSoft/SIGPLAN Symposium on Practical Software Development Environments, pages 97--106, November 1988.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....are needed. Project goal definitions help in selecting the appropriate processes to match the given goals and also motivate project personnel in terms of what is expected. Algorithms are used to guide the performance of project members in terms of what they should do. With the exception of [17], too little emphasis has been placed on explicitly modeling project and process goals. In the large versus in the small models: Projects are typically performed by teams of developers. We distinguish between process modeling in thesmall (describing the process aspects concerning an individual ....

....project that will be sufficient to perform the work successfully. There will always be unforeseen problems that must be handled during project performance. Therefore, static (unchangeable) project plans are not acceptable for guiding software projects. Little previous work (again an exception is [17]) has addressed the need for changeable plans. 3.1.1 The Process Modeling Language MVP L In the MVP Project, we developed the process notation called MVP L [18] Our notation was designed for modeling processes in the large, including building comprehensible specifications and designs of ....

Karen Huff and Viktor Lesser. A plan--based intelligent assistant that supports the software development process. In Peter Henderson, editor, rd ACM SIGSoft/SIGPLAN Symposium on Practical Software Development Environments, pages 97--106, November 1988.

....level of description, describe a potentially non linear sequence of actions. These actions Non linear sequence indicates a partial ordering of nondeterministic or potentially iterative incremental actions. In other places, these action sequences are called task chains [6, 18, 40, 42] or p aris [431. 361 affect some concrete or abstract transformation of a soft ware BSO, product or partition. For example, in the task of implementing a software system design as a program, the creation of a successfully compilable program component is a necessary action. Other actions in the sequence ....

HUFF, K.E., and LESSER, V.: 'A plan-based intelligent assistant that supports the software development process'. Proc. Third Syrup. on Software Development Environments, ACM SIGSOFT, Boston, Massachusetts, November 1988, pp.

....of PML paradigms. For instance, 23] identifies five different paradigms: 1. Programming models, e.g. APPL A, PADM PML. 2. Functional models, e.g. HFSP. 3. Petri net (and state transition) models, e.g. SPADE 1, Melmac [42] and ProcessWeaver [29] 4. Plan based models, e.g. GRAPPLE [44]. 5. Quantitative models, e.g. system dynamics [1] Another classification is proposed in [53] where six different classes are identified: graphical, net oriented, procedural, object oriented, rule based, and multi paradigms. In general, the issue is still debated in the community (see the ....

K. Huff and V. Lesser. A plan-based intelligent assistant that supports the software development process. In Proceedings of the Third Software Engineering Symposium on Practical Software Development Environments, 1989.

....Major exceptions include Arcadia [8] which uses an imperative notation called appl a [13] based on Ada, hfsp [9] which uses an extension of attribute grammars, and Melmac [3] Slang [1] and Process Weaver [4] which use a form of Petri nets. Important examples of rule based PCEs include Grapple [7] and Merlin [11] We would like to compare our experience against the following two concerns: How (if at all) have the implementors of PCEs been able to support reengineering of their own code What processes (if any) have been constructed to assist and guide users in their reengineering efforts ....

Karen E. Huff and Victor R. Lesser. A plan-based intelligent assistant that supports the software development process. In Peter Henderson, editor, ACM SIGSOFT/SIGPLAN Software Engineering Symposium on Practical Software Development Environments, pages 97--106, Boston MA, November 1988. ACM Press. Special issue of Software Engineering Notes, 13(5).

....violate problem constraints (a specific type of design test case) simulating portions of the design and expediting design decision making and evaluation. Huff has exploited the structure of plans and their ability to relate disparate goals in several applications in software engineering, e.g. [12] in process engineering and [13] in software adaptation. Rist represented different levels of functionality and goals in programs using a plan representation [25] his PARE system extracted the abstract plan structure from PASCAL programs to aid in program design and re use. Planning has been ....

Karen Huff and Victor Lesser. A plan-based intelligent assistant that supports the software development process. In ACM SIGSOFT/SIGPLAN Software Engineering Symposium on Practical Software Development Environments, Nov. 1988.

....ultimately into individual rules. The resulting hierarchical plan specifies the control flow that the EPOS execution manager then assists users in carrying out. This approach breaks down when a (sub)task fails to achieve its goal, but an incremental replanning mechanism has been proposed. Grapple [8] plans software processes off line, for manual execution by process participants. In principle, it could also monitor in progress processes as input to plan recognition. StateMate [9] models the traditional who, what, where, when and how of a process and simulates the dynamic or behavioral ....

Karen E. Huff and Victor R. Lesser. A plan-based intelligent assistant that supports the software development process. In ACM SIGSOFT/SIGPLAN Software Engineering Symposium on Practical Software Development Environments, pages 97--106, Boston MA, November 1988.

....by constraining the messages allowed to be exchanged among the entities that compose the system. Formally, a law is a restricted Prolog program. Darwin itself is a constraint satisfaction system implemented as a prototype written in Prolog. A completely different Prolog based tool is GRAPPLE [84]. It is a plan based process assistant, that includes primitive environment operators that correspond to atomic actions within the environment. Some primitive actions correspond to tool invocations, while other primitive actions correspond to predefined scripts. These operators may be combined in ....

....inside which every process participant works. For instance, the following rule initializes the working context of a programmer: do working context(R,W,X,Item) 26 Product Specification [2, 9, 43, 47, 52, 58, 62, 68, 102, 114, 121, 124, 128, 139, 143, 182, 186, 191] Process Management [3, 13, 33, 35, 85, 77, 84, 146, 155, 196, 163] Validation and prototyping [24, 53, 56, 71, 82, 98, 99, 104, 115, 120, 133, 148, 170, 176, 177, 179, 188, 197] Design [8, 4, 38, 44, 63, 106, 109, 118, 147, 160, 162, 202] Editing and Compiling [10, 18, 48, 36, 78, 158, 136, 150, 194] Debugging and testing [21, 42, 46, 59, 67, 74, 76, 81, 92, ....

K. Huff and V. Lesser. A Plan-Based Intelligent Assistant that Supports the Software Development Process. In Proc. 3rd ACM SIGSOFT Symp. on Software Development Environments, volume 13:5 of ACM SIGSOFT Software Engineering Notes, pages 97--106, 1988.

....for process refinement, that follows at least two approaches, that we will call explicit and implicit . The implicit approach can be loosely characterized as the attempt to exploit Artificial Intelligence techniques in software process modelling; it is exemplified by Articulator [20] Grapple [14], EPOS [9] and PEACE [3] The emphasis is on the problem solving capabilities of the process engine, that enables the enactment of the process by refining an overall goal to simpler ones. Not always these approaches succeed in providing simple process models, nor it is always easy to assess them. ....

K. Huff and V. Lesser. A Plan-based Intelligent Assistant that supports the Software Development Process. In P. Enderson, editor, ACM SIGSOFT '88, pages 97--106, Boston MA, 1988. ACM Press.

....of their corresponding artifacts that are important to the processes that manipulate the artifacts. The processes cannot themselves access the actual artifact, except through invoked 9 tools, and tools cannot access the proxies. This approach has been adopted in Marvel, Merlin, SLANG, and GRAPPLE [20]. Other languages support the definition of artifacts at any level of granularity. For example, APPL A, Adele [6] AP5 [8] and Pleiades [36] are based on general purpose programming languages, and they provide rich type models that support the fine grained definition and manipulation of ....

....of data, and then effect process evolution by modifying the data representations. Examples of this approach are EPOS [10] in which production processes are represented as a combination of task networks and rules, SLANG [4] which represents production processes as modified Petri nets, GRAPPLE [20], which represents production processes as plans, and Merlin [22] which represents processes as mutable rules. For our purposes, the approach of combining representation of production processes and meta processes complicates the representation of production processes. It yields process ....

Karen E. Huff and Victor Lesser. A plan-based intelligent assistant that supports the software development process. In ACM Symposium on Practical Software Development Environments, pages 97 -- 106, 1988.

....decomposition in HFSP [26] module decomposition in Arcadia s APPL A [24, 38] or net decomposition in SPADE s SLANG [4] to cite a few. None has an associated refinement calculus. Besides, there are attempts to exploit Artificial Intelligence techniques, like Articulator [30] Grapple [21], EPOS [12] and PEACE [3] The emphasis is on the problem solving capabilities of the process engine, that enables the enactment of the process by refining an overall goal to simpler ones. These approaches not always succeed in providing simple process models, nor it is always easy to assess ....

K. Huff and V. Lesser. A Plan-based Intelligent Assistant that supports the Software Development Process. In P. Enderson, editor, ACM SIGSOFT '88, pages 97--106, Boston MA, 1988. ACM Press.

....in which the process currently is, in order to change that situation to the desired new one. In this respect, the contextual approach has some similarity with the planning paradigm that has emerged from Artificial Intelligence and with projects based on the planning paradigm such as GRAPPLE [Huf89]. Since the contextual models adopt the notion of a decision, all the properties of decision oriented models mentioned earlier are applicable to them. Further, due to the strong relationship between the situation and the decision, only those decisions which are appropriate in the situation at hand ....

: K. E. Huff, V. R. Lessor, A plan-based intelligent assistant that supports the software development process, Proc. of the 3rd Software Engineering Symposium on Practical Software Development Environments, Soft. Eng. Notes, 13, 5, 1989, pp97-106

.... main activities in the process life cycle: process modeling, process analysis, process enactment, and process execution status monitoring and control [2] 5] Previous approaches to providing automated support for the process lifecycle have focused attention on subsets of the lifecycle [8] 6] [4]. The SMART architecture represents the first attempt to provide comprehensive support for the entire process life cycle [2] SMART, however, does not provide robust support for execution status monitoring and process redesign based on feedback status analysis driven feedback. This paper presents ....

K. E. Huff and V. R. Lesser, "A Plan-based Intelligent Assistant that Supports the Software Development Process", Proceedings of the 3 rd ACM SIGSOFT/SIGPLAN Symposium on Practical Software Development Environments, ACM Press, New York, 1988, pp. 97-106.

....applicators. Many aspects of the environment were weakly analogous to what one would expect to find if using Petri net formalisms [4] as in the Role Interaction Net (RIN) language [5] The environment might also be viewed as a graphical design by constraint language comparable to GRAPPLE [6]. To help the environment scale up to handle large processes, aggregate abstractions could easily be created for common groups of objects (e.g. related collections of files) and entire circuits could be abstracted into new, user defined objects to use in building a process model. The environment ....

Huff, K. E., and V. R. Lessor. "A Plan-Based Intelligent Assistant that Supports the Software Development Process." Software Engineering Notes 13,5, 1989.

....not at present well suited for upstream process life cycle activities, such as incremental process modeling or simulation. In contrast, the PSS project in England [BPe91] has developed a process modeling and enactment language based on an object oriented knowledge representation notation. Grapple [HL88], on the other hand, relies on a set of goal operators and a planning mechanism to represent software processes. These are used to demonstrate goal directed reasoning about software processes during modeling and enactment. AP5 [BN93] developed at USC ISI, and Marvel [KF87] developed at Columbia, ....

K.E. Huff and V.R. Lesser. A Plan-Based Intelligent Assistant That Supports the Process of Programming. ACM SIGSOFT Software Engineering Notes, 13:97-- 106, Nov 1988.

....the goal of the Raven Configuration Management System. The RCMS is part of the Raven Project at the University of British Columbia. The goal of the Raven Project is to explore a variety of issues in distributed object oriented computing. The RCMS draws its inspiration from systems such as GRAPPLE[HL88] Conic[KMS87] MKS89] and Darwin[Min85] It provides a unified, dynamic, and distributed approach to configuration management. Configuration information is represented as a series of constraints expressed in first order predicate logic. Related constraints are grouped together into units referred ....

....are still under development. 5 Conclusions The RCMS represents a new approach to the problem of configuration management. Although there is a great variety of systems addressing particular issues in configuration management [BVT89] All80] KMS87] RSS88] PF89] Sim89] TS89] TBC 88] HL88] WSY91] few systems have attempted to encompass the features of the RCMS: 1) Active Configuration Management: The RCMS automates configuration management through the active monitoring and adjustment of all software components in a distributed system. Only Conic [MKS89] offers a similar degree ....

Karen E. Huff and Victor R. Lesser. A plan-based intelligent assistant that supports the software development process. ACM SigSoft, 13(5):97--106, 1988.

....and encoding for the process described above. The major elements of the structure are four roles. 3 The term plan is used here in the sense of something which might be familiar to a project manager. There are obvious similarities with the technical AI planning paradigm such as described in [2]. However, the relationships between such interpretations and a change model such as PMMS are not considered in this paper. 1. managing. This is the hub role. It receives establishes terms of reference and acts as the initiator of PMMS behaviour. 2. technology. This role is responsible for ....

Karen E. Huff and Victor R. Lesser. A Plan-based Intelligent Assistant That Supports the Software Development Process. In Peter Henderson, editor, Proceedings of the ACM SIGSOFT /SIGPLAN Software Engineering Symposium on Practical Software Development Environments, 28-30 November, 1988, Boston. ACM Press 1988.

....of operations and their effects. This work centers on using a rule base and goals to derive a generalized execution flow from a specific process history. By having enough rules, they show that a complete and correct process fragment could be generated from execution data. ffl Huff and Lesser [17] describe plan recognition, whereby a plan and its goal are inferred from a sequence of actions and an initial state. The use of this is to help interactively guide software developers along action sequences that achieve the correct goal, rather than to infer a model of the process. ffl Garg et ....

K.E. Huff and V.R. Lesser. A Plan-based Intelligent Assistant that Supports the Software Development Process. In SIGSOFT '88: Proceedings of the Third Symposium on Software Development Environments, pages 97--106. ACM SIGSOFT, February 1989.

....of a particular design history which can be used to solve several, analogous problems. In the process programming domain, the idea of viewing process programming as a mapping between between explicitly represented goals and a sequence of environment actions was first proposed by Huff and Lesser [16]. Their system, GRAPPLE, uses the planning paradigm for planning (constructing a sequence of actions to achieve a given goal) as well as plan recognition (inferring a plan and its goal from a sequence of actions) The domain knowledge in GRAPPLE consists of a library of pre defined higher level ....

Karen E. Huff and Victor R. Lesser. A Plan-Based Intelligent Assistant that Supports the Software Development Process. In Peter Henderson, editor, Proceedings of the ACM SIGSOFT/SIGPLAN Software Engineering Symposium on Practical Software Development Environments, pages 97--106, Boston, Massachusetts, 1988.

....critical details involved in model composition. Finally, we conclude and add some additional remarks regarding our current efforts. 2 Background In the domain of large scale software development, many new product, process, and decision support models have been proposed [DBe90, EJe91, GP 92, HL88, Kai88, MS90, OB92, RD91, SPW92, WS88] These models are used to facilitate different aspects of software development when employed within support systems we call process centered software development support environments (PSDSEs) These models also facilitate the archiving and reuse of ....

....SDM resource webs to achieve a better or best match The task of evolving a candidate SDM resource web into a configuration that better matches the one bound to the target resource infrastructure is the essence of the composition planning problem. This is a generative process planning problem [HL88] This means that the plan must be generated to incrementally transform the candidate SDM into the desired target resource configuration, given available resources. The transformations may include adding or removing resources, restructuring resource configuration (web) relations, revising ....

K.E. Huff and V.R. Lesser. A Plan-Based Intelligent Assistant That Supports the Process of Programming. ACM SIGSOFT Software Engineering Notes, 13:97-- 106, Nov 1988.

....we, in this chapter, will focus on the problems of simulating and analyzing business processes. We propose the use of a novel logic programming language as a tool that will support the analysis phase in process reengineering. Unlike other areas of process modeling, e.g. software process modeling [HL88] where the initial emphasis has been put on the parts of processes that can be automated so that they can be machine executable, business process modeling attempts to capture phenomena enacted by humans rather than machines. This creates additional requirements, specifically for modeling actions ....

....there as well. Perceptual actions are dealt with in [LLL 94] In the area of information systems development and in process modeling in particular, the frame and ramification problems have traditionally been either ignored or bypassed by means of explicit assumptions (see, e.g. DDBDP94] HL88] In [BMR93] a systematic solution to the frame and ramification problems for determinate transaction specifications is presented. We will employ this solution for strengthening process specifications at design time to guarantee the maintenance of constraints by any implementation that ....

K. Huff and V. Lesser. A Plan-based Intelligent Assistant that Supports the Software Development Process. In Proceedings of SIGSOFT-88, pages 97--106, 1988.

....research includes the formulation of languages and paradigms for formally describing software processes. Paradigms that have been used in formal process modeling include programming [115] graph based control such as state machines [80] and Petri nets [11, 12, 40] rulebased [14, 96] and planning [68]. Huff [67] gives a good overview of modeling and languages. Each paradigm has its own strengths and weaknesses; a rule based model will be able to easily incorporate exceptions into its behavior, but will not be able to give a process engineer a good overall picture of the model s likely ....

K.E. Huff and V.R. Lesser. A plan-based intelligent assistant that supports the software development process. In Proc. 3rd ACM SIGSOFT/SIGPLAN Symposium on Practical Software Development Environments, pages 97--106. ACM Press, February 1989.

No context found.

Karen Huff and Viktor Lesser. A plan--based intelligent assistant that supports the software development process. In Peter Henderson, editor, rd ACM SIGSoft/SIGPLAN Symposium on Practical Software Development Environments, pages 97--106, November 1988.

No context found.

[Huff 89] Karen E. Huff , Victor R. Lesser, A plan-based intelligent assistant that supports the software development, ACM SIGPLAN Notices, v.24 n.2, p.97-106, Feb. 1989

No context found.

Huff, K.E., Lesser, V.R. "A Plan-Based Intelligent Assistant That Supports the Software Development Process", Third Symposium on Software Development Environments, ACM, 1988.

No context found.

Karen E. Huff and V. R. Lesser. A Plan-Based Intelligent Assistant that supports the the Software Development Process. In Proc. of the 3rd ACM Symposium on Software Development Environments, pages 97--106, Boston, Massachusetts, November 1988.

First 50 documents

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