Quality Money Management : Process Engineering and Best Practices for Systematic Trading and Investment

106 CHAPTER ◆ 1 0 Prototype in Modeling Software

experiment to a final, consolidated version of the trading/investment system. Whatever

the case, prototypes allow product teams to:

● Clarify calculations and expose inconsistencies in specifications.

● Rapidly evaluate alternative methods.

● Deliver intermediate, working versions to end users and other interest parties for feedback.

Furthermore, prototyping leads to:

● Improved morale because progress is visible.

● Early feedback on whether the final trading system will be acceptable.

● Decreased overall code length due to better designs.

● Lower defect rates because of better requirements definition.

● Smoother effort curves, reducing the deadline effect.

10.1. Rapid Development

The term “ rapid development language ” (RDL) refers to any programming language that offers

speedier implementation than do traditional third-generation languages such as C. RDLs

produce their savings by reducing the amount of construction needed to build a product. Shorter

cycles make incremental development methods, such as evolutionary prototyping, practical.

Studies have shown that early introduction of software products is the dominant factor in long-term prof-

itability for many, though not all, products. This holds true for trading/investment systems as well. Early

and often delivery of potentially tradable prototypes is imperative.

Software developed in RDLs is generally better suited to the development of in-house

software, with more limited distribution than systems software. Widely used RDLs in

finance, however, such as Excel, Resolver, and MATLAB, require many fewer lines per

function point relative to other languages. Function points, a language-independent mea sure

of program size based on a weighted sum of the number of inputs, outputs, inquiries, and

files, allow developers to think about program size in a language-independent way, easing

comparison between languages. A low-level language, like Assembler, requires many more

lines of code to implement a function point than does a higher-level language such as C.

A language ’ s level is the number of Assembler statements needed to replace one statement

in the higher-level language. Here are the levels of some different languages:

TABLE 10-1

Language Level Statements per function point

Assembler 1 320

C 2.5 125

C 6.5 50

Excel c. 50 6

C# 12 27

SAS, SPSS 10 30

Java 6.5 50