documents/applications/tutorial.tex

40 \begin{center}\fbox{\large TECHNICAL REPORT IC-PARC-#1}\end{center}
45 \large #3
69 In this tutorial we provide a methodology for developing large scale
86 This tutorial is an introduction to the design, development, test and maintenance of large scale applications with the ECLiPSe system. It follows a top-down methodology of converting an initial high-level specification into executable code of high quality. We assume that fundamental decisions on selecting the right tool and the right problem solver have already been taken so that we are commited to using ECLiPSe and one of its problem solving libraries. We are basically interested in the engineering aspects of the program development, not its research and development content.
211 This structure is typical for large scale combinatorial optimization\index{large scale combinatorial optimization} (LSCO)\index{LSCO} problems. The flow analysis part of RiskWise follows this structure. It consists of five parts, where each performs one particular task.
367 "Sizes":"the sizes of min, small, large, max packets",
1846 In a large program, it may be difficult to leap directly to the interesting part of the program. But we may have to repeat this operation several times, if we repeatedly leap/skip over an interesting statement. We can use the invocation number of a statement to jump\index{jump} to this exact place in the execution again. The invocation number\index{invocation number} is printed in parentheses at the beginning of each trace line. By re-starting the debugger, copying this number into the text field to the right of the button {\it To Invoc:}\index{To Invoc} and then pressing this button we can directly jump to this location.
1848 Unfortunately, jumping to an invocation number can be quite slow in a large program. The debugger has to scan each statement to check its invocation number. It can be faster (but more tedious) to use skip and leap commands to reach a program point\footnote{Experiment! Your mileage may vary!}.
1867 \item A more complex test can be used to exercise all (or nearly all) of the program code. The test data must contain enough variation to force all alternative parts in the program. This can be achieved either by accumulating large number of tests or by designing particular tests to handle special cases.