Processing Data Objects

All data in a DSPOOM model is contained in Processing Data objects, a concept that is related to the Payloads pattern explained in section 1.5.3.

DSPOOM Processing objects can only process Processing Data objects. To be fully usable in such a context, a Processing Data class must offer a number of services, namely:

  1. Introspection: A Processing Data object must be able to respond when queried about the number of attributes it holds and their type. It must also be able to know its own type or class name.
  2. Homogeneous interface: All Processing Data classes must have a homogeneous interface so as to be queried transparently without any knowledge of the concrete subclass.
  3. Encapsulation: Attributes must be protected and access to them only given through appropriate setters and getters.
  4. Persistence: A Processing Data must have automatically built-in persistence into an appropriate format.
  5. Display facilities: Any Processing Data must be accompanied by display facilities that allow for debugging and visualizing its content at any time.
  6. Composition: An attribute of a Processing Data class may be another processing data.
Note that many of these requirements are related to the services offered by MOF meta-objects (see section 1.4.3).

Figure 4.6: Processing Data Class
\includegraphics[%
width=0.20\paperwidth,
keepaspectratio]{images/ch4-DSPOOM/ps/DSPOOM_ProcessingData.ps}



Subsections
2004-10-18