X hits on this document

PDF document

A FLEXIBLE SIZING APPROACH - page 6 / 7

18 views

0 shares

0 downloads

0 comments

6 / 7

Example 2: Average ‘look and feel’ style sheet design element (HTML):

Average or Expected Sizes (to get Low/High, add +/- 20%)

  • Header = 30 HTML

  • Footer = 8 HTML

  • Navigation = 12 HTML

When sizing the middle tier, a variety of different components requires a flexible approach. A complex report that allows users to drill down to derive information is sized as follows:

Tables:

  • 6-10 Tables * 6 Definition Steps + Calculations & Control Breaks:

  • 12-20 Fields * 6 Definition Steps

Sizing a complex controller yields the following calculation:

  • #

    of Fields (Low 5 Fields, Average 7 Fields, High 10 Fields), *

    • 4 Methods per Field

    • up to 2 Java Beans

The work involved in data persistence and lookup is estimated by breaking this task into its simplest form:

1 Lookup Statement +

  • #

    of Fields to Map in Single Table

  • Low 15, Average 100, High 200

More complex lookups are accommodated by adjusting the number of lookups and multiplying the average number of fields in a table by the number of tables to be queried. The beauty of tying the basic work unit to a rough line of code equivalent is that it allows development teams the freedom to ignore code, use code exclusively, or combine GUI and SLOC estimates, as in this interface example:

  • 13-50 Data Elements +

  • (6-20 Data Translations * 10 ESLOC per Translation) +

  • 150-300 ESLOC (Code to filter the data)

Regardless of the task, the process of deriving the gearing factor is the same. The estimator begins at the highest level and logically walks through the process of creating each component or “step”, asking questions such as, “How do you create a simple login screen?” “How many fields does a simple login screen contain? Do you have to configure each field? If so, how many configurations/properties (on average) must be performed per field?”

Often the answer to these questions will be a range (high, medium, and low) rather than a single number. This is fine, because it allows estimators to determine an expected value and uncertainty range for each size estimate. Once the individual low, most likely, and high estimated gearing factors for each sizing component have been rolled up, they can be loaded into a sizing spreadsheet to speed up future size estimates and encourage standardization across projects. The example below shows the roll up for the data tier of the example shown earlier:

5

© Copyright 2011 Quantitative Software Management. Inc. All Rights Reserved.

Document info
Document views18
Page views18
Page last viewedSun Dec 04 04:49:53 UTC 2016
Pages7
Paragraphs91
Words2542

Comments