Kaiser G.E., Barghouti N.S., Feiler P.H., Schwanke R., "Database Support for knowledgeBased Engineering Environments", IEEE Expert, Summer 1988, pp 18-32.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....(e.g. SADT [17] ffl graph net analysis oriented formalisms (e.g. Petri nets [2] Funsoft nets [9] or statecharts [13] ffl process programming enforcement oriented languages (e.g. Arcadia APPL A [21]or IPSE 2. 5 PSS [4] ffl rule based enforcement oriented languages (e.g. Marvel MSL [12]) Data model Definition The sub model of a softwareprocess model which describes the data types being manipulated, their structure, and properties. Comments Very different kinds of descriptions can be found according to the intended use of the model. For instance, an artifact can be ....

....as time or budget. The latter ones deal for instance with execution properties of actual tools of the domain support technology, such as their role as transformers of data objects, their kind (interactive or automatic) their appropriate paradigms of invocation (see the envelope concept in Marvel [12]) Role model Definition The sub model of a softwareprocess model which describes the roles types, their structure, and properties. Comments Some role properties are more or less relevant according to the intended use of the model# for instance, access control properties are more important for ....

Kaiser G.E., Barghouti N.S., Feiler P.H., Schwanke R., "Database Support for knowledgeBased Engineering Environments", IEEE Expert, Summer 1988, pp 18-32.

....exist outside the MOO (presumably in the file system) every artifact to be manipulated within the process is represented by a corresponding object within Promo. The Promo object contains a URL [5] that acts as a pointer to the actual artifact. Like Marvel s use of the file system as its database [33], the use of URLs within Promo permits the easy storage and manipulation of potentially complex data. By contrast with Marvel, using URLs rather than the local file system permits the distribution of artifacts in a manner that makes sense for the process. 1 While the data of the artifact resides ....

G.E. Kaiser, N.S. Barghouti, P.H. Feiler, and R.W. Schwanke. Database Support for Knowledge-Based Engineering Environments. IEEE Expert, 3(2):18--32, Summer 1988.

....conventional Unix tools into the tailored PSDE rather than building specialized tools; and (3) supporting cooperation and coordination among multiple developers within a single software process. Many of the details of the first two aspects of Marvel have been presented in previous papers [KFP88, KBFS88] In this paper, I concentrate on the third aspect. I present the design and implementation of the concurrency control mechanism in Marvel, which significantly improve on our initial ideas regarding concurrency control [BK90] I also describe the constructs of the Marvel process modeling language ....

Gail E. Kaiser, Naser S. Barghouti, Peter H. Feiler, and Robert W. Schwanke. Database Support for Knowledge-Based Engineering Environments. IEEE Expert, 3(2):18--32, Summer 1988.

....(presumably in the file system) every artifact to be manipulated within the process is represented by a corresponding object within Promo. The Promo object contains a URL [BernersLee et al. 1994] that acts as a pointer to the actual artifact. Like Marvel s use of the file system as its database [Kaiser et al. 1988], the use of URLs within Promo permits the easy storage and manipulation of potentially complex data. By contrast with Marvel, using URLs rather than the local file system permits the distribution of artifacts in a manner that makes sense for the process. 1 While the data of the artifact reside ....

KAISER, G. E., BARGHOUTI, N. S., FEILER, P. H., AND SCHWANKE, R. W. 1988. Database support for knowledge-based engineering environments. IEEE Expert 3, 2 (Summer), 18--32.

....ENUM activity, product, decision, context 5.2.2 Granularity Most traditional process models are large grained descriptions of the product lifecycle. On the other hand, there are very fine grained models. For example specifying that after a source code file is edited, it should be recompiled [Kai88]. Recently, hybrid formalisms that use different notations for large grain and smallgrain aspects of process such as PROCESS WEAVER [Fer91] have been developed. The nature of granularity needed is dependent on the situation at hand. Granularity affects the kind of guidance, explanation and trace ....

.... for E3 [Fin94] various Prolog dialects for EPOS [Jac92] Oikos [Amb91] and PEACE [Fin94] PS Algol for PWI [Fin94] Different computational paradigms have also been used, for example, Petri nets in EPOS [Jac92] and SPADE [Ban93] rule based paradigm in MERLIN [Emm91] ALF [Ben89] Marvel [Kai88], EPOS [Jac92] and triggers in ADELE [Bel89] and MVPL [Fin94] There is a relationship between the construction technique and the form facet in the system world. Indeed, languages are typically related to process programs whereas instantiation techniques have been used to construct process ....

[Article contains additional citation context not shown here]

G. E. Kaiser, N. S. Barghouti, P. H. Feiler, R. W. Schwanke, Database Support for Knowledge-Based Engineering Environments, IEEE Expert, 3(2), 1988, pp18-32.

....However, until they have actually lived up to their promise, a mix of file based object storage and a relatively small database or managing relationships seems to be the compromise of choice. Even very recent contributions from the software engineering community such as the Marvel environment [KBFS88] and the hypertext system DIF [GS90] follow this hybrid approach. The most frequently quoted limitations that render current databases unsuitable for software environments include the lack of complex objects, the need for integrating external tools, the extensibility in terms of available ....

....functions in addition to data structure, whereas Arcadias CACTIS [HK87] offers selective re computation of data derived by such functions. Tool integration is sometimes addressed by socalled envelopes that represent and activate the functionality of tools by literals in a rule language [KBFS88]. The many object oriented or next generation database manifestos indicate that there is no obvious consensus on a common data model. This lack of consensus and the quest for open environments has led many to believe that extensibility is an essential ingredient for specification ....

[Article contains additional citation context not shown here]

Kaiser, G.E., Bhargouti, N.S., Feiler, P., Schwanke, R.W. (1988). Database support for knowledge-based engineering environments. IEEE Expert 4, 2, 18-32.

....process. ADELE [BE87] uses different kinds of predicative integrity constraints to recognize consistency problems. But the user has to attach these constraints to the right objects (otherwise the system will behave erroneously) and the resulting corrective actions are hand programmed. MARVEL [KBFS88] uses rules as a representation for external tool integration, with opportunistic rule invocation in either forward or backward fashion. There are no specific features for the configuration management problem. Among other reasons, the moderate scope of these approaches may also be traced to the ....

Kaiser, G.E., Bhargouti, N.S., Feiler, P., Schwanke, R.W. (1988). Database support for knowledge-based engineering environments. IEEE Expert 4, 2, 18-32.

....was feasible and useful and also to give insight into the ongoing design of such Process Modelling Languages. Keywords:Process Modelling Languages, Formal Specification 1 Introduction Several formal approaches for process modelling have been developed. Some interesting examples can be found in [1], 2] 3] and [4] These different approaches are generally so because they are based on different underlying models rather than for methodological reasons. In this paper we review a small case study which we recently undertook to investigate the application of an existing formal framework to ....

G. E. Kaiser, N. S. Barghouti, P. E. Feiler, R. W. Schwank, "Database Support for Knowledge-Based Engineering Environments" IEEE Experts, 1988.

....of the software process. We, therefore, compare our approach with other approaches which also use a rule based formalism. The main advantages and key distinguishing features of the Merlin approach as compared to other rule based process centered software development environments like MARVEL [KBFS88] or OIKOS [ACM90] are the persistent storage of a process combined with its efficient and flexible interpretation of backward and forward chaining rules. Furthermore, the unique feature of combining different rule syntax based on different underlying execution paradigms represents a first step ....

....Thus, the concept of a working context and user friendly display of all possible activities before the user actually makes a (possibly failing) try, does not exist in MARVEL. Concerning data base support, the MARVEL system uses a special purpose adaption of the UNIX directory structure (cf. [KBFS88]) This enables to keep pointers to potentially executable rules analogous to the ECorrRule edges in Merlin (cf. section 5) However, to our knowledge, there is no support to manage variables in global variable lists or manage other very fine grained information. In addition, the MARVEL solution ....

Kaiser G.E., Barghouti N.S., Feiler P.H., Schwanke R.W., Database Support for Knowledge-Based Engineering Environments, IEEE Expert, Summer 1988, pp. 1832.

....compare our approach with other approaches which also use a rule based formalism. First of all, the feature of different syntactical definitions for rules which are differently interpreted, distinguishes our approach from other rule based approaches in process modelling like the MARVEL system [KBFS88] or OIKOS [ACM90] MARVEL also employs an alternating use of forward and backward chaining which is called opportunistic processing. However, the process engineer has not always explicit control about the switching points. Only in some cases he she can define rules called hints which can only be ....

Kaiser G.E., Barghouti N.S., Feiler P.H., Schwanke R.W., Database Support for Knowledge-Based Engineering Environments, IEEE Expert, Summer 1988, pp. 18-32.

....for instance, to choose a bottom up or top down design approach [3] or to decide if an entity should be retyped, decomposed or subtyped. At the other end of the spectrum, there are models extremely detailed. For example specifying that after a source code file is edited, it should be recompiled [29]. Recently, hybrid formalisms that use different notations for large grain and small grain aspects of process such as PROCESS WEAVER [20] have been developed. We believe that a single process modelling formalism should accommodate a wide range of model granularity in a homogeneous fashion. The ....

....process performance. Consequently, current CASE tools are essentially passive; they help in capturing, storing manipulating and documenting products but offer very little to support the largely intellectual process of product construction. Most process centred software environments [25] 4] 2] [29] provide active guidance [16] which consists of strict enforcement of a process model. In fact, as indicated by [45] process models are used to describe tools activity and to allow automatic invocation of tools. Process models are used to guide tools usage and not to support the application ....

G.E. Kaiser, N.S. Barghouti, P.H. Feiler, R.W. Schwanke, Database Support for Knowledge-Based Engineering Environments, IEEE Expert, 3(2), 1998, pp18-32.

....the query mechanism for the program views. A further important requirement for CARE (as well as for CASE) is the concurrent access to the software system by different reengineering programmers. With this concern in mind, we intend to set our system on the rule based development environment MARVEL [4]. 7 Appendix A The following definition defines the syntax of COBAPA. The definitions of logical expression, expression, list of variables, list of expressions, newline, line number and variable name are considered to be as they are generally known. We show the syntax of the control structure ....

Gail E. Kaiser, Naser S. Barghouti, Peter H. Feiler, and Robert W. Schwanke. Database support for knowledge-based engineering environments. IEEE Expert, 3(2):18--32, Summer 1988.

....based, includes some support for management of persistent data; Marvel 2.6 is supported on Sun OS 4.0 with X 11 windows (also Ultrix 3.1 AIX 2.2.1) Marvel has been extensively described in the literature. Particular reference may be made to the key papers: Kaiser, Feiler Popovich (1988) and Kaiser, Barghouti, Feiler Schwanke (1988). 4 How does Marvel work To construct a Marvel environment the developer must produce a data model and a process model. The data model describes the objects to be managed during the process of software development and their properties. The process model describes the activities carried out on ....

Kaiser G., Barghouti N., Feiler, P. & Schwanke, R. (1988); Database Support for Knowledgebased Engineering Environments; IEEE Expert; V3, N3, pp 18-32.

....make integrated SDEs practical in the same way that research results in database and operating systems have achieved widespread commercial application. I describe an admittedly incomplete model that is a step towards this goal, and apply this model to our MARVEL software development environment [1, 2]. This extended transaction model addresses concurrency control, consistent release and user control through a semantics for commit serializability supported by two new transaction processing operations, split transaction and join transaction, that we recently introduced [3] Previously we ....

Kaiser, G.E., Barghouti, N.S., Feiler, P.H. and Schwanke, R.W. Database Support for Knowledge-Based Engineering Environments. IEEE Expert 3(2):18-32, Summer, 1988.

....We discuss in another paper [19] the specification of alternatives, iteration and synchronization through the conditions and effects of rules; the process is not in any sense limited to a deterministic sequence of steps. Possible chains are compiled into a network when the kernel is tailored [18]. The process engine chains among rules with different or multiple parameters by inverting local bindings [16] An automation predicate is enclosed in parentheses ( and may optionally be preceded by one of the chaining directives. A consistency predicate is enclosed in square brackets ....

Gail E. Kaiser, Naser S. Barghouti, Peter H. Feiler and Robert W. Schwanke. Database Support for Knowledge-Based Engineering Environments. IEEE Expert 3(2):18-32, Summer, 1988.

.... which becomes the parameter (predicate 1a) Chaining is prohibited into or out of the compile rule from another instance of the compile rule, due to the no chain directive (there is also a no backward directive) Additional details about the rule formalism and its chaining engine are given in [4, 11, 10]. Conventional file oriented tools are integrated into a Marvel process without source modifications, or even recompilation, through an enveloping language [7] The rule activity indicates the tool and envelope name, with input literals and attributes to be supplied as arguments and output ....

Gail E. Kaiser, Naser S. Barghouti, Peter H. Feiler, and Robert W. Schwanke. Database support for knowledge-based engineering environments. IEEE Expert, 3(2):18--32, Summer 1988.

....chaining problem. In this paper, we describe the chaining problem and explain how it is solved in the current implementation of Marvel. We describe Marvel, detailing only those aspects necessary to understand the chaining problem and our solutions; for a more complete description of Marvel, see [12, 13, 11]. We next explain the chaining problem in detail with a motivating example. We describe a heuristic approach to solving this problem and discuss its implementation in the current Marvel system. We then extend the example to show the limitations of this approach, 1 For this paper, we restrict ....

Gail E. Kaiser, Naser S. Barghouti, Peter H. Feiler, and Robert W. Schwanke. Database Support for Knowledge-Based Engineering Environments. IEEE Expert, 3(2):18--32, Summer 1988.

....with predecessor successor relations that can be directly represented in the process state task server. Although Oz s process modeling language tends to obscure such relationships from a human readability standpoint, they are visible in the internal rule network compiled from the process model [34]. A less restrictive option would be to only post tasks with already satisfied conditions, or prerequisites in some other non rule based process modeling paradigm, but this prevents posting of tasks obligated for eventual completion [50] but not currently enabled. A better approach might be to ....

Gail E. Kaiser, Naser S. Barghouti, Peter H. Feiler, and Robert W. Schwanke. Database support for knowledge-based engineering environments. IEEE Expert, 3(2):18--32, Summer 1988.

....principles, but employ a rooted filesystem for byte stream data items. The byte stream data items are typically those required by COTS tools, and it is not necessarily the case that all data items are represented in the filesystem. Examples of form 3a include the Smile and Marvel [KBFS88] data repositories. Since the filesystem can provide little help in determining the set of data items to be moved, or relations among these items, the SDE s user interface must provide an iteration or navigation facility that reaches all data items. Form 3b Form 3b is the same as form 3a, except ....

....Marvel 2.6, consists of about 45,000 lines of C code, provides both a command line user interface and an X11 graphical user interface, and runs on SunOS 4.0.3, Ultrix 3.1 and AIX 2.2.1. Further details of Marvel, and our experience using it, are described in previous papers [KF87a, BK88, KFP88, KBFS88, KB88, Sok89, KBS90] This paper is concerned only with Marvel s data model and object base as it relates to the implementation of our two immigration tools. 5.2 Marvelizer: Base Case Immigration Marvelizer is an implementation of the base case of extraction. It is implemented by the marvelize ....

Gail E. Kaiser, Naser S. Barghouti, Peter H. Feiler, and Robert W. Schwanke. Database Support for Knowledge-Based Engineering environments. IEEE Expert, 3(2):18--32, Summer 1988.

....to the inherent difficulty of distributing an object oriented database. A special consideration is given to the specific problems of programming environments. The second part describes partial implementation of a subset of this model, namely a single server implementation, for MARVEL. MARVEL [7] [21], is an SDE that combines object oriented data modeling with rule based process modeling. MARVEL is considered as an instance of the software process environments [9] which are characterized as having a programmable process model. Thus, the implementation for MARVEL had to consider its specific ....

Gail E. Kaiser, Naser S. Barghouti, Peter H. Feiler and Robert W. Schwanke. Database Support for Knowledge-Based Engineering Environments. IEEE Expert 3(2):18-32, May, 1988. 81

....employs existing external tools to carry out the steps of a software process. Since tools are expensive to develop in terms of both time and cost, a major objective is to employ existing tools without modifying them. We achieve this by encapsulation of external tool interactions within envelopes [2]. The envelope concept was introduced in the ISTAR environment [8] Each step of a process is represented in the process model by an activity. An activity may involve building an executable, running a test suite, or simply invoking an editor. An activity s execution often involves the invocation ....

Kaiser, G.E., Barghouti, N.S., Feiler, P.H., and Schwanke, R.W., "Database Support for Knowledge-Based Engineering Environments," IEEE Expert 3(2):18-32, Summer 1988.

No context found.

G.E. Kaiser, N.S. Barghouti, P.H. Feiler, and R.W. Schwanke, "Database Support for Knowledge-Based Engineering Environments," IEEE Expert, Summer 1988 .

No context found.

G.E. Kaiser, N.S. Barghouti, P.H. Feiler, R.W. Schwanke, Database Support for Knowledge Based Engineering Environment, IEEE Expert, 3:3, 1988.

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