J. Software Engineering & Applications, 2010, 3 Published Online January 2010 in SciRes(www.SciRP.org/journal/jsea) CONTENTS Volume 3 Number 1 January 2010 Models for Improving Software System Size Estimates during Development W. W. AGRESTI, W. M. EVANCO & W. M. THOMAS………………………………………………1 A Massively Parallel Re-Configurable Mesh Computer Emulator: Design, Modeling and Realization M. YOUSSFI, O. BOUATTANE & M. O. BENSALAH………………………………………………11 Properties of Nash Equilibrium Retail Prices in Contract Model with a Supplier, Multiple Retailers and Price-Dependent Demand K. NAKADE, S. TSUBOUCHI & I. SEDIRI……………………………………………………………27 A Polar Coordinate System Based Grid Algorithm for Star Identification H. ZHANG, H. S. SANG & X. B. SHEN………………………………………………………………34 Evaluation of Feature Subset Selection, Feature Weighting, and Prototype Selection for Biomedical Applications S. LITTLE, S. COLANTONIO, O. SALVETTI & P. PERNER…………………………………………39 Cryptanalysis of TEA Using Quantum-Inspired Genetic Algorithms W. HU……………………………………………………………………………………………………50 Application of Design Patterns in Process of Large-Scale Software Evolving W. WANG, H. ZHAO, H. LI, P. Li, D. YAO, Z. LIU, B. LI, S. YU, H. LIU & K. Z. YANG…………58 Element Retrieval Using Namespace Based on Keyword Search over XML Documents Y. WANG, Z. K. CHEN & X. D. HUANG………………………………………………………………65 Integrated Web Architecture Based on Web3D, Flex and SSH W. J. ZHANG……………………………………………………………………………………………73 Analysis and Comparison of Five Kinds of Typical Device-Level Embedded Operating Systems J. L. WANG, H. ZHAO, P. LI, H. LI & B. LI……………………………………………………………81 Makespan Algorithms and Heuristic for Internet-Based Collaborative Manufacturing Process Using Bottleneck Approach S. A. BAREDUAN & S. HASAN………………………………………………………………………91 Copyright © 2010 SciRes JSEA Journal of Software Engineering and Applications (JSEA) Journal Information SUBSCRIPTIONS The Journal of Software Engineering and Applications (Online at Scientific Research Publishing, www.SciRP.org) is published monthly by Scientific Research Publishing, Inc., USA. E-mail: [email protected] Subscription rates: Volume 3 2010 Print: $50 per copy. Electronic: free, available on www.SciRP.org. To subscribe, please contact Journals Subscriptions Department, E-mail: [email protected] Sample copies: If you are interested in subscribing, you may obtain a free sample copy by contacting Scientific Research Publishing, Inc. at the above address. SERVICES Advertisements Advertisement Sales Department, E-mail: [email protected] Reprints (minimum quantity 100 copies) Reprints Co-ordinator, Scientific Research Publishing, Inc., USA. E-mail: [email protected] COPYRIGHT Copyright© 2010 Scientific Research Publishing, Inc. All Rights Reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise, except as described below, without the permission in writing of the Publisher. Copying of articles is not permitted except for personal and internal use, to the extent permitted by national copyright law, or under the terms of a license issued by the national Reproduction Rights Organization. Requests for permission for other kinds of copying, such as copying for general distribution, for advertising or promotional purposes, for creating new collective works or for resale, and other enquiries should be addressed to the Publisher. Statements and opinions expressed in the articles and communications are those of the individual contributors and not the statements and opinion of Scientific Research Publishing, Inc. We assume no responsibility or liability for any damage or injury to persons or property arising out of the use of any materials, instructions, methods or ideas contained herein. We expressly disclaim any implied warranties of merchantability or fitness for a particular purpose. If expert assistance is required, the services of a competent professional person should be sought. PRODUCTION INFORMATION For manuscripts that have been accepted for publication, please contact: E-mail: [email protected] J. Software Engineering & Applications, 2010, 3: 1-10 1 doi:10.4236/jsea.2010.31001 Published Online January 2010 (http://www.SciRP.org/journal/jsea) Models for Improving Software System Size Estimates during Development William W. AGRESTI1, William M. EVANCO2, William M. THOMAS3 1Carey Business School, Johns Hopkins University, Baltimore, USA; 2Statistical Solutions, Philadelphia, USA; 3The MITRE Corpo- ration, McLean, VA, USA. Email: [email protected], [email protected], [email protected] Received August 28th, 2009; revised September 16th, 2009; accepted September 29th, 2009. ABSTRACT This paper addresses the challenge of estimating eventual software system size during a development project. The ap- proach is to build a family of estimation models that use information about architectural design characteristics of the evolving software product as leading indicators of system size. Four models were developed to provide an increasingly accurate size estimate throughout the design process. Multivariate regression analyses were conducted using 21 Ada subsystems, totaling 183,000 lines of code. The models explain from 47% of the variation in delivered software size early in the design phase, to 89% late in the design phase. Keywords: Software Size, Estimation, Ada, Regression, Re-Estimation, Metrics 1. Introduction the need for an accurate cost estimate often translates into a need for an accurate size estimate. Before software development projects start, customers and This paper addresses size estimation, but distinguishes managers want to know the eventual project cost with as between estimation before development and during de- much accuracy as possible. Cost estimation is extremely velopment. Size estimation prior to the start of a software important to provide an early indicator of what lies ahead: project typically draws on analogy and professional will the budget be sufficient for this job? This need has judgment, with comparisons of the proposed system to motivated the development of software cost estimation previously developed systems of related functionality. models, and the growth of a commercial market for such Function point counts may also be used to provide an models, associated automated tools, and consulting sup- estimate of size, but, of course, the accuracy of the size port services. estimate depends on accurate knowledge of the entities When someone who is new to the use of cost estima- being counted. tion models looks at the estimation equation, it can be Once the project begins, managers would like to keep quite disconcerting. The only recognizable variable on improving their size estimate. However, we are moti- the right-hand side is size, surrounded by a few modify- vated by our observations of current practice in which ing factors and shape parameters. So, if the project is just managers, during development, revert to predevelopment beginning, how do you know the size of the system? estimates of size (and effort and cost) because of a lack More experienced staff members explain that you must of effective ways to incorporate current development first estimate the size so that you can supply that value in information to refine their initial estimates [3]. Ideally the equation, thus enabling you to estimate cost. How there would be straightforward methods to improve the good can the cost estimate be if it depends so strongly on predevelopment estimates based on early project experi- a quantity that won’t be known until the end of the pro- ences and the evolving product. In this paper we focus on ject? The underlying logic prompting this question is the evolving product as a source of information for up- irrefutable. This situation is well known and recognized dating the size estimate. (e.g., [1, 2]). Users of such cost estimation models are The research reported in this paper addresses the ques- dependent, first and foremost, on accurate size estimates. tion of how to improve our capability for estimating the Of course, there are other ways to estimate costs (e.g., by size of software systems during a development project. analogy and experience), but the more analytically satis- More specifically, it reports on building a family of fying models estimate cost as a function of size. Thus, models for successively refining the size estimate during Copyright © 2010 SciRes JSEA 2 Models for Improving Software System Size Estimates during Development the development process. The notion of a family of mod- relationship diagrams. It was found that a model using els is intended to address the challenge of successively the number of data elements crossing an elementary pro- refining the initial estimate as the project unfolds. The cess in the data flow diagram as the sole explanatory research has three motivations: the widely known poor variable performed fairly well as a predictive model. record of large software projects to be delivered on time Much of the research using early system artifacts for and on budget (due in part to poor estimation capability), estimation is directed at estimating effort and cost. Pau- the persistent illogic of needing to know size to estimate lish, et al., discussed the use of the software architecture cost, and the challenge of successive size reestimation document as the primary input to project planning [8]. during a project. Mohagheghi et al., described an effort estimation model The remainder of the