Needs for semantic support

Usage Scenarios that Illustrate Needs for Semantic Support... 

Needs for Semantic Support Derived from the Scenarios... 

The collection of needs is the foundation to investigate how well Semantic Web capabilities can address these needs for semantic support. 

Although, the assumptions of K-RAMP deviate from the situation as it is described in usage scenario 3, the highlighted needs for semantic support are similar on a more abstract level. Product functions, product design, and process capabilities have to be represented (Nl), heterogeneous views with respect to the product, the process and process control have to be represented (N2), or analytics of process data is essential in order to classify capabilities of process segments (N3). 

Due to the similarities with the aforementioned usage scenarios, it can be shown that the same needs for semantic support are valid for the ramp-up of new products. 

As a summary, K-RAMP can be seen as another usage scenario with a broad variety of needs for semantic support. All requirements (addressed in Chap. 2 and further extended in this chapter) are shown in Table 9.1. 

The results described in this section also need to be seen in relation to the new tool functionalities of the CRC, as presented in the previous chapter, especially with respect to the PRIME approach in Sect. 3.1. The PDW cooperates closely with the PRIME environment in two primary directions. The models, their realizations, and the extended semantic functionalities, as described in this section, offer a set of services to the fine-grained process support described there on the other hand, the traceability mechanisms needed for process guidance provide part of the functionality required for the process data warehouse approach. 

Common SE approaches are based on standardized and static product data integration, whereas RPD requires dynamic data management in semantic networks in order to enable short control cycles. Short paths and multidisciplinary teams for quick decisions are essential for both approaches. Moreover, RPD requires team-oriented communication systems, which open up new ways of cooperation. They need to offer support not only for management decisions, but also for decision mtiking during the generation of product data. 

It is relatively straightforward to modify the semantics to support such datatypes (see the techniques in Kroger (1987)) however this would require a base first order logic, thus complicating its use in verification. Of course, the need for such datatypes is determined by the types of circuits being modeled and the level of the model. 

Third generation RPC— Java RMI is conceptually similar to the second generation but supports the semantics of object invocation in different address spaces that are built for Java only [15]. Java RMI fits cleanly into the language with no need for standardized data representation, external interface definition language, and with behavioural transfer that allows remote objects to perform operations that are determined at mntime ... 

Although it inherits much of the syntactic idiosyncrasy of C++, Java is a somewhat safer language. Unfortunately, its lack of support for user-defined types with value semantics increases the need for dynamic memory allocation. The Microsoft language C is in many ways similar to Java but supports value types. The provision of a garbage collector in both languages is a boon to commercial software developers but is likely to be unacceptable in real-time systems. This may be less of a problem in future as generational and concurrent garbage collectors (Jones Lins 1996) become mainstream. We note that a real-time subset of Java has been defined (RTJ 2003). 

Assessment of fitness-for-purpose is achieved through systematic examination of the behaviour of a stem. Testing, inspection, and analysis techniques may all be required to demonstrate adequate levels of fitness-for-purpose. For conventional software the prevailing opinion is that use of recognised design practices, control and information flow analysis, and dynamic testing needs to be satisfactorily undertaken to show fitness-for-purpose. There is also support in some industries for the use of semantic analysis against a (de-facto) formal specification. 

---