Article Preview
TopIntroduction
Packaged software products (also referred to as software package or fully packed product) are software tools that provide pre-built functionality to develop the solution application for a particular domain. Specific purpose-built Packaged software products serve as development accelerators reducing overall program cost.
Usage of packaged products such as portal products, CMS products, ecommerce products, digital marketing products is very common in software projects. Packaged products offer pre-built functionality to accelerate the solution development and attain higher productivity by reusing the out of box capabilities. The out of box product functionality can also be extended through configurations and extensions to reuse the product feature in wide variety of scenarios.
In this paper we propose a novel estimation method, “Software package product estimation framework” for software projects that employ packaged products for solution development.
The state of the art estimation models do not address the product specific scenarios such as build and unit effort, project lifecycle effort.
The Software package product estimation framework provides the estimation guidelines, complexity definition scale and effort calculation formulae for packaged product based software solutions.
Generic software estimation models include COCOMO (Boehm, 1981), Function Point/FP (Matson et al.,1994), COCOMO II (Boehm et al., 2000), SEER-SEM (Jenson 1984), SLIM (Putnam & Myers, 1992), PRICE-S (Freiman, 1975), Delphi (Boehm, 1984), Rule-based/Rule of thumb, Use case point (Ochodek, 2011), Work breakdown structure (Jorgensen 2004), Planning poker, Story point estimation, learning (Goldberg 1989), Case based reasoning (Aamodt & Plaza 1994), Analogy based estimation, Select Estimator, top-down estimation, bottom-up estimation, price-to-win, Stepwise ANNOVA (Basha, 2010), Ordinary Least squares (Griffiths et.al.1993). Other software estimation method are machine learning-based estimation (Mair et al., 2000), fuzzy logic based estimation (Gray, 1997), genetic programming (Burgess et al. 2001) and expert based estimating method (Jørgensen, 2004).
For software packaged product estimation, Boehm et al. proposed Constructive Product Line Investment Model (COPLIMO) based on COCOMO 2 model to estimate the effort for broader product lines. In the absence of a specific estimation model, most of the projects use function point or use case point estimation method for packaged product based solutions.
Gaps With State of the Art Techniques
The main gap with state of the art estimations models are as follows:
- •
State of the art estimation methods do not provide estimation framework for end to end software development lifecycle stages for software projects using packaged products.
- •
State of the art methods do not provide estimation framework that can be used for wide variety of software products.
Method
In this paper we have proposed an estimation framework “Software packaged product estimation framework” (SPPEF) that can be used for applications developed using packaged products.
High level steps used in the “Software packaged product estimation framework” are as follows:
- 1.
Obtain the historical data for various project lifecycle activities and support activities for software projects using packaged products.
- 2.
Calculate the build effort using the complexity scale factors and effort calculation formula described in this section. We can then calculate the effort estimate for project lifecycle activities and support activities using the guidance values elaborated.
- 3.
The overall effort is the sum total of build effort, lifecycle effort and support activity effort.
In the coming sections we will elaborate the calculation for each of these categories.
Pre-Requisites for Software Testing Estimation Framework
For the software packaged product estimation framework, we need the historical effort data for baselining. We need to get the historical data for previous projects that used the packaged products.