Wiederhold, G., Wegner, P., and Ceri, S. Towards Megaprogramming. Communications of the ACM 33, 11 (Nov. 1992), pp. 89 -- 99.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....and we will use the word theory for large grain types at the module level. Some History The paradigm for module composition described in this paper is based on parameterized programming and hyperprogramming [Gog89, Gog90b] the term megaprogramming has been used in the darpa community [BS92, WWC92]) Because this approach involves modules for both specification and code, both kinds of module must be composed. Specifications are used as headers for code, and are combined with simple set theoretic operations. At the code level, composition may be done with intermediate compiled code, which ....

Wiederhold, G., Wegner, P., and Ceri, S. Toward Megaprogramming. Communications of the ACM, 35(11):89--99, 1992.

.... architectures [GK94] In an information sharing model, for example, agents advertise about their capabilities and needs, so that other agents can use this information for their own tasks ( EW94, LMM94] We take an alternative approach, organizing agents in groups, called federated systems [GK94, WWC92] Figure 2 illustrates an example federation: In such a system, agents communicate by means of Agent Resource Broker Agent A Agent B Agent C Agent D Agent E Federation entry point Figure 2. An example federated system a broker agent called facilitator. Agents tell their needs and ....

G. Wiederhold, P.Wegner, and S. Cefi. Toward megaprogramming. Communications of the ACM,33(11):88-- 1992. 14

....and data, allowing for model instance evolution and reflection, in addition to more traditional data access. Second, high level abstractions are first class and highly scalable. These benefits are of particular importance when the problem being modeled falls in the domain of megaprogramming [25]. 3. THE OSM MODEL In this section, we introduce Object oriented System Modeling (OSM) a logical system model that provides a solid foundation for the system development approach we are proposing. 3.1. Overview OSM is an object oriented model for systems analysis, specification, design, ....

G. Wiederhold, P. Wegner, and S. Ceri, "Toward Megaprogramming," Communications of the ACM, 35, 11, pp. 89-99 (November 1992).

....work while affiliated with NC State University. In reality, the components emerging today are based on standards that may compete and conflict with each other: OMG s CORBA [3] Sun s Java [4] Microsoft s COM [5] A software component considered in this paper has the granularity of a megamodule [6]: Megamodules are internally homogeneous, independently maintained software systems [ Each megamodule describes its externally accessible data structures and operations and has an internally consistent behavior. i.e. such components are stand alone programs, installed and maintained ....

....operations and has an internally consistent behavior. i.e. such components are stand alone programs, installed and maintained on a specific host on the network. In contrast to developing a megamodule composition, programming architecture, and language compilation environment as envisioned in [6], the widely practiced approaches to composition of stand alone modules into a single program rely on scripting languages [7] Scripting is widely used by programmers to deliver packaged multi component applications whose interface may range from a simple command line to complex GUI in workflows, ....

C. Wiederhold and P. Wegner and S. Ceri. Toward Megaprogramming. Communication of ACM, 35(11):89--99, 1992.

....under one, and only one, request broker architecture. Though this is extremely necessary for application development, architectural needs go unfulfilled. CORBA is more like a detailed requirement specification, defining in detail the needs of a particular application domain. Megaprogramming [WWC92] extends the call mechanism to an asynchronous messaging paradigm between large components called megamodules. The communication between two megamodules is defined with language structures like setup, estimate, invoke, extract, and examine. Languages, like Rapide [LV96] extend the interface ....

G. Wiederhold, P. Wegner, and S. Ceri. Towards megaprogramming. Communications of the ACM, 35(11):89--99, 1992. 91

....that legacy component A need not be modified; but unless A was designed with future interoperability in mind, this is unlikely. 2 PolySPIN s approach addresses the third (and therefore the first and second as well) and most difficult of the three scenarios, sometimes called megaprogramming [14, 73], in which both components A and B already exist. We want to modify 2 Moreover, even if legacy component A were designed with interoperability in mind, that is no guarantee that it is an appropriate kind of interoperability to use with future component B. 51 Megaprogramming Common case ....

Wiederhold, Gio, Wegner, Peter, and Ceri, Stefano. Toward megaprogramming. Communications of the ACM 35, 11 (Nov. 1992), 89--99.

....school of thought and be given sets of links that are customized to that perspective. Another user with a different perspective would find a different set of links for the same corpus. Research in thesaurus merging (Rada, 1985) scheme merging (Nica Rundensteiner, 1995) and ontology definition (Wiederhold, Wegner, Ceri, 1992) may eventually be helpful here. Second, scholars should be encouraged to create pathfinders guides to themes or topics that not only give pointers to information resources but also critical commentary and interpretations about those resources. Since it is likely that we see the continued ....

Wiederhold, G., Wegner, P., & Ceri, S. (1992). Toward megaprogramming. Communications of the ACM, 35(11), 89---99.

....offer possibilities to perform fine tuning of its parameters to meet application requirements. A single system that bridges all of the underlying heterogeneities is a challenging task, and will have to find the right tradeoff between hidden and visible integration. The new concept megaprogramming [15] applies well to the MDBS problem. Megaprograms are expected to coordinate so called megamodules, comparable to LSYSs. Each megamodule exports a (self) description of its ontology . The megaprogramming language should offer constructs to supply parame ters to modules, initiate computations, and ....

Gio Wiederhold, Peter Wegner and Stefano Ceri. "Toward Megaprogramming". Communications of the ACM . Vol. 35, No. 11, November 1992.

....execution, it di ers signi cantly from MARS in that is is a language speci cation, whereas MARS provides a software infrastructure for MDA execution. Thus MARS may even be used as the compilation target for Opus. Another runtime approach to providing MDA support can be found in the CHAIMS project [4,5]. While similar in goals to MARS, CHAIMS operates at a much higher level of abstraction. Rather than providing a low level infrastructure to support legacy module interactions, CHAIMS takes the Corba route of providing objects that know how to perform some task and tell others who might inquire ....

G. Wiederhold, P. Wegner and S. Ceri, Towards Megaprogramming,Communications of the ACM, November 1992.

....can reasonably be expected to interoperate in a federated environment. We exclude processes and programming language objects from the federation discussion, because the mechanism is far too heavy for such detailed integration. Objects of the same granularity are elsewhere called megamodules [251]. Also megamodules encapsulates besides procedures and data also types, knowledge and ontology (concepts and interpretation paradigm) In contrast to programming language objects and processes of a computing system, the federated system objects can encapsulate not only data but also ....

....Both approaches attach system size, persistence of the systems over evolution steps in system components, and diversity of various aspects, like software production concepts and tools, communication facilities, and interfaces. An application in a federated environment can be seen as a megaprogram [251]. However, the evolution of megacomponents is not controlled by the megaprogram programmer, but the federation mechanism is used for creating a temporary commitment between the components. The megaprogramming tools trust in a component database and a shared ontology. In the federated environment, ....

WIEDERHOLD,G.,WEGNER,P.,AND CERI, S. Towards Megaprogramming. Communications of the ACM 33, 11 (Nov. 1992), 89 -- 99.

....in CodeFinder prevents placing other types of artifacts such as larger grained components or supporting documentation. While it is desirable to support higher level reuse, such as subsystems [Wirfs Brock, Johnson 1990] architectures [Shaw, Garlan 1996] or megaprogramming [Biggerstaff 1992; Wiederhold et al. 1992] the functionality and level of granularity described here is characteristic of current software reuse repositories, which are designed to find and re use functions and objects to create new programs. Evolution of Reuse Repositories 11 23 99 10 Henninger Given the use of non words and ....

Wiederhold, G., Wegner, P., Ceri, S. "Toward Megaprogramming," Communications of the ACM, 35(11), pp. 89-99.

....useful systems. 1. location and retrieval of computational objects by users. There is a great deal of ongoing work on information retrieval and broking on the Web. These are clearly related to our work and need to be adapted to our context as well. REFERENCES 14 2. model management research [Kri93, Bec94, Muh92, WWC92] methods will now truly have large model libraries distributed over the Internet to work with. This can be used to validate existing research methods as well as stimulate new lines of inquiry. 3. Finally, there are a range of issues related to making method check in, method selection and ....

G. Wiederhold, P. Wegner, and S. Ceri. Toward megaprogramming. Communications of the ACM, 35(11):89--99, 1992.

....M to the World Wide Web. This approach heads for what could be called large scale model management, where the aim is high level integration of a heterogeneous world of model implementations. A similar goal is apparent in the work of Muhanna [Muh92] The proposals for a megaprogramming language [WWC92] go into the same direction, although they focus more on the related database and software engineering aspects. In the sequel, we will present MMM, a distributed model management system for time series analysis and prediction. Examples of time series range from the irregularity in a heartbeat to ....

....Ypsilon encapsulates methods in function classes, which each have an input field, an output field, and a method called evaluate to call the routine for computing the output from the input. This design is similar to what is called operational programming in [MS94, Sol87] or megaprogramming in [WWC92] The abstraction allows to implement meta algorithms on objects of function classes that are independent of the specific type of the functions. For example, the current version of Ypsilon comes with the realization of various Eigenmodels [BMR89, MMR94] i.e. special models that represent the ....

G. Wiederhold, P. Wegner, and S. Ceri. Toward megaprogramming. Communications of the ACM, 35(11):89--99, 1992.

....model has fields input and output. The typical usage of a function model in a C program is illustrated below. Here, F is a function model representing a given software module, and f is an instance of F. F f; f.input( a; f.evaluate( b = f. output( As suggested by Wiederhold et al. WWC92] the function call is decomposed into three steps: i) retrieval of the input data; ii) execution of the method by means of evaluate( and (iii) assignment of the output data. 4 THE IMPLEMENTATION OF MMM 16 Function models can also encapsulate stateful services. As such, they are implemented ....

G. Wiederhold, P. Wegner, and S. Ceri. Toward megaprogramming. Communications of the ACM, 35(11):89--99, 1992.

.... interoperation issues intended for the specification and design of interoperable information systems [Kalin94j] The goal of an information system design in SYNTHESIS is a creation of a composition of pre existing components (information resources) This composition treated as a megaprogram [WWC92] should serve as a concretization of a specification of an information system. The resources in a megaprogram interoperate through HIRE built around an architecture providing technical level of interoperability [CORBA92] In SYNTHESIS semantic interoperation framework an information system design ....

G. Wiederhold, P. Wegner, and S. Ceri. Toward megaprogramming. Communications of the ASM, 35(11):89--99, November, 1992.

No context found.

G. Wiederhold, P. Wegner, et al., "Towards Megaprogramming", Comm. ACM, vol. 35(11), Nov 1992, pp. 89-99.

No context found.

G. Wiederhold, P. Wegner, and S. Ceri, "Towards Megaprogramming", Comm. ACM, vol. 35(11), Nov 1992, pp. 89-99.

No context found.

G. Wiederhold, P. Wegner, et al., "Towards Megaprogramming", Comm. ACM, vol. 35(11), Nov 1992, pp. 89-99.

No context found.

G. Wiederhold, P. Wegner, and S. Ceri, "Towards Megaprogramming", Comm. ACM, vol. 35(11), Nov 1992, pp. 89-99.

....utilize functionalities provided by existing service components. Commercial off theshelf (COTS) software applications and other information services are the building blocks that provide pieces of functionalities. This vision of software composition [15] is echoed in the megaprogramming framework [5, 25], which builds on software components called megamodules [18] that capture the functionality of autonomous services provided by large organizational units. Autonomous services are linked together according to composition specifications [19, 23] to form megaservices. We use the term autonomous ....

G. Wiederhold, P. Wegner, and S. Ceri, "Towards Megaprogramming", Comm. ACM, vol. 35(11), Nov 1992, pp. 89-99.

....as they would expect from a program running solely on local resources. Systems such as CHAIMS (Compiling High level Access Interfaces for Multi site Software) allow clients to abstract away heterogeneity and service autonomy while simultaneously compensating for pitfalls associated with both [15]. We focus on scheduling with CHAIMS because its preferred development language (CLAM Composition Language for Autonomous Megamodules) provides key language primitives that enable dynamic scheduling with the possibility of recovery from hazards. CLAM contains a primitive to get estimates of the ....

G. Wiederhold, P. Wegner, S. Ceri, "Towards Megaprogramming", CACM, Nov.1992.

....in a homogeneous language, so that the occurrence of a constant, say 4, in two distinct methods, has the same meaning in both of them, while the same constant cannot appear in two different forms in two different methods. However, in real life, this is clearly not a very realistic assumption (see [24] for discussion on this subject) an essential part of our future research will be devoted to tackle this problem. Acknowledgements We thank Gerard Comyn and Alexander Herold for their encouragement and support. We are indebted to Mike Reeve for discussions on the importance of property driven ....

G. Wiederhold, P. Wegner, and S. Ceri. Towards mega-programming. Technical report, University of Stanford, Dept. of Computing, Standford, Ca, U.S.A., 1990.

.... up module interfaces prior to executions, request performance estimates from modules prior to their invocation, schedule module execution in parallel, monitor execution of invoked modules, interrupt inadequately performing modules, and provide data and metainformation to customer interface modules [13]. This work has been supported by DARPA EDCS D884 and Deutsche Forschungsgemeinschaft, SFB 373. CHAIMS supports the paradigm shift which is already occurring in building systems: a move from the focus of programming to a focus on composition. This shift is occurring invisibly to many ....

....that module performance, although not addressed directly by CHAIMS, continues to critically affect system performance. In a multi site system composed via a CHAIMS megaprogram, individual module performance will be improved by allowing a choice of platforms and module programming languages. [13] discuss in detail further megaprogramming concepts like the possibility for testing of new and updated software without committing to specific configurations, adaptation to change of scale of modules, or network configuration. 3 Related work Not surprisingly, there are many other initiatives ....

[Article contains additional citation context not shown here]

Wiederhold, G., P. Wegner, and S. Ceri, Towards Megaprogramming. Comm. ACM, 1992(11): p. 89-99. <http://www-db.stanford.edu/CHAIMS/>.

No context found.

Wiederhold, G., Wegner, P., and Ceri, S. Towards Megaprogramming. Communications of the ACM 33, 11 (Nov. 1992), pp. 89 -- 99.

No context found.

G. Wiederhold, P. Wegner, and S. Ceri. Towards Megaprogramming. Communications of the ACM, 35(11):89--99, November 1992.

First 50 documents  Next 50

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