Effort estimation Of web based applications

Abstract

Countless organizations around the world have developed commercial and educational applications for the World Wide Web, the best known example of a hypermedia system. But developing good Web applications is expensive, mostly in terms of time and degree of difficulty for the authors. Our study tries to predict the effort needed for the development of web pages of a particular category, here we have restricted our self to the domain of news web sites. We try to forecast the average effort required to code a page of a new site belonging to the same category based on the analysis of the data available from the existing sites. Here we have considered the data from Top Ten News Sites. The number of pages in the site, Average Number of Lines per Page, Average Number of Scripts per Page, Link Density, and Media Density are taken into account while predicting the effort required to code a page of the site. We consider the effort required to be directly proportional to the number of lines of code. It can be expressed to be in man hours only when we have information about the actual effort in man-hours required to code the site. We use Multiple Linear Regression, Stepwise Regression and Polynomial Regression to analyze the data and obtain the graphs which a be used to predict the approximate effort required to code a web page of a news site. Finally we devised a method to estimate the effort required or the number of lines of code required for a webpage of a news site. These results can be used to devise a standard for the coding of newer news web sites. If they are incorporated in to a web authoring software it would help the author to stick to the guidelines and the standards automatically.