projects in which the mix of languages is well known, relative percentages can be applied to the overall gearing factor to yield estimated language gearing factors for each constituent language. It is obviously preferable, however, to use single language projects to calculate language-specific gearing factors.
ESTIMATED GEARING FACTORS
If gearing factors are not available from your completed projects, they can be derived during the estimation process by breaking the work to be accomplished into discrete steps or components and dividing each step or component into complexity bins (each with an associated gearing factor that ties it back to the basic unit of work). For a COTS package implementation, this process might begin with a high level architecture view that flows into a more detailed breakdown of how requirements are implemented in each tier of the finished application.
In the presentation tier of the example pictured above, the presentation layer items are identified (a style sheet that standardizes the look and feel and update screens that allow users to log in, display and update their user profile, and manipulate complex data arrays). These “steps” or size components are then logically decomposed into complexity bins and mapped to their associated technologies. In the first example (a simple login screen implemented in JSP) the basic work units represent the work required to configure fields on a login screen.
Example 1: Simple Login screen implemented in JSP
Fields (2) *
# User Actions (1 per field)+
# Checks (Low 2, Average 3, High 5)
The second example (a style sheet design element), looks more like what we might expect for traditional lines of code, though it may in fact have been produced in a GUI environment:
© Copyright 2011 Quantitative Software Management. Inc. All Rights Reserved.