Spreadsheet Model Refinement

The implementation s vocabulary should be translated to the specification s by using retrieve functions as in Section 6.4, Spreadsheet Model Refinement. 

Whether you believe all that about my spreadsheet is not important here. Certainly, there are many times when performance, decoupling, or a partial spec leads to differences between model and code. So the reviewer is faced with trying to determine whether there is some correspondence between them whether the code conforms to the spec (or refines it). 

We can take the action refinement even further. How is a setAddition performed By typing = and clicking in a cell. For that we need an addition to the model to record that after the =, we re now in the identify C effor adding state and to map mouse coordinates to C etl Yes, finally, we ve gotten down to something that uses the graphical layout of the spreadsheet. 

Focusing on the system we re interested in refining (the spreadsheet rather than the user), we may decide, as we did in this example, to use a different set of objects in the actual design. We have already seen how to use abstraction functions to relate the design model to the specification. 

Semi-formal representations can be used as an intermediate step for the generation of a model with the intended level of detail and formalization First, the operations are specified by chemists or process experts on a conceptual, semi-formal level, and subsequently formalized and refined. Since such semi-formal representations are currently missing, inadequate workarounds like spreadsheets or informal documents are often used in industrial practice. 

Aspen HYSYS Petroleum Refining includes a method to convert limited feed information (distillation curve, density, viscosity, refractive index, etc.) into kinetic lumps for use in the unit-level FCC model. In this section, we present an alternative method based on data and methods available in the public literature. We extend the work by Bollas et al. [52] to infer the kinetic lump composition from limited process data. This method uses techniques to normalize the distillation curve, cut the distillation curve into boiling-point lumps, and infer the composition of the each of these boiling-point lumps. We have developed all of these techniques into spreadsheets using Microsoft Excel. These spreadsheets are available in the DVD accompanying this text... 

---