The General Effect of an Integrated Software Product Evaluation
Jozsef Gyorkos , Ivan Rozman, Robert T. Leskovar 1 University of Maribor, Faculty of Electrical Engineering and Computer Science, Smetanova 17, SI-2000 Maribor, Slovenia; e-mail: gyorkos@uni-mb.si
Changes in the labour market have brought even more intensive outsourcing. Recently companies have moved away from traditional long term employment arrangements (insourcing) to relatively short-term market mediated arrangements (outsourcing) [Slaughter, 1996]. This source states (Derived from research in U.S.), that firms in the public sector are more likely to outsource IS employment, than firms in the private sector.
Our paper will use as a basis for the discussion the Government Information Centre of the Republic Slovenia, the CVI.
Outsourcing has brought about dynamic boundary changes, i.e. (distributed application development, non transparent organisation structure of vendors and heterogeneous methodologies). However, the continuos operation of an integrated information system, must be assured at a conceptual level, being in relation to the data and common functional structures. The structure is based on the rules and processes of public administration, but at the same time it is highly dependent on actual decisions, rules and policies of the State and (actual) Government. Outsourcing also circulates the resources (money) received from the tax payer, back to its productive source - the tax payer.
CVI has a very broad aim, that is the introduction and maintenance of information technology in all government institutions, and to lesser degree public institutions.
Slovenia is small country in relation to many other European states, and like others born or 'reborn' in the late 1980s and early 1990s, it has had to develop a whole new IS supported administrative apparatus, which is a requirement for a modern functional state. In the world of computer technology, there is a similar situation, whereby, there are not so many clients, but the functionality of the server is the same as in larger states.
Nam mentions four types of outsourcing when he talks of the relationship between 'Extent of substitution by vendor' and 'Strategic impact of applications', these are: support, reliance, alignment and alliance. Using the CVI as a model we can see the distribution of outsourcing contracts among different types of outsourcing in Table I below.
In non-information technology industries, 'alignment' and 'alliance' are the usual types of outsourcing. In the forthcoming information society, the business process in a given public administration will be under pressure to re-engineer, due to the process automation and the demand for information dissemination, with or without administrative assistance.
The low level of alignment and almost non-alliance in outsourcing explains the strategic position of CVI. Whereby, we have not only a reliable information technology (IT) service provider for public administration but also the added responsibility of having a forum for making long-term, highly influential decisions in IT.
Table I
Types of Outsourcing and
their Approximate Extent in Case of CVI
technical supervision for IS planning and design,
small contracts with strategic influence
based on mutual relationships,
highest commitments from vendors and clients
* Derived out of values of contracts.
The scope of preparation is the requirements definition and selection of the vendor (usually directed by the legislative council). The outsourcing vendor is the organisation which delivers the required product and/or service. In this paper we will focus on outsourced projects where software develops within information technology projects and all IT related services.
Control of the Outsourced project by client consist of project tracing, quality assurance and of change management (technical or contextual – business process - changes).
When an outsourced project is running, threshold must be drawn on the scale, representing the degree of supervision of the vendor by the client. Table II highlights the achievements and drawbacks for a client regarding the different degrees of supervision.
Table Ii
Degree of Client Supervision in outsourced Projects
contract,
integration test results of the final product
project tracing results (schedule, resources)
technical specification,
change management records,
test coverage plans and test results
reliability
asking about worth of outsourcing
It is important to stress, that these processes not only control the vendor but also contribute to a flexible and transparent project management in relation to the client (in our case the CVI). The intensity of project management activities, carried out by a client, depends on the type of supervision, formulated in Table II. Table IV, enumerates the existing practices within CVI in relation to risk prevention.
Table Iii
Means of risk management in the software process.
Review of specification
(PROCESSUS Tool, chapter 5.1)
Project tracing
When consistent project management is used only the ‘products’ of the project management process can be checked. In our case the consistency of project management, ensures the adopted PRINCE methodology [PRINCE, 1996]. This PRINCE methodology is applied in the form of ProjectOffice, a distributed environment with functionalities, i.e. project group co-ordination, process control, exception handling and quality assurance.
Before a release, the customer (CVI) performs one final assessment of the product not only from general view. This assessment can take up to five days, depending on the size of the product and the level of its integration with other (existing or in-development) products. These results are then collated in a tool called PRO+ which is planned to be a consistent part of the ProjectOffice application.
PRO+ is based on quality attributes described in standards; ISO 9126 (IT – Software product evaluation – Quality characteristics and guidelines for their use), and partialy on ISO 9127 (Information processing systems – User documentation and cover information for consumer software packages) and ISO 12119 (IT – Software packages – Quality requirements and testing). In the background of the tool a formal decision support is defined, this enables calibration of parameters and in the process will give reports of the different views (client, developer, user). The calibration is fully transparent and enables multiple-level assessment of the same product.
Table IV
Existing risk prevention practices in CVI
TBD – to be declared (method preparation/selection in progress).
BOR – based on legislative regulation.
Fig. 1. High-level report from PRO+ Tool.
Fig. 2. The logical object model of the PROCESSUS Tool.
The results of the final decision (assessment of the product), are given in the form of a vector, containing multiple parameters, with normalised values (from 0 to 1). Each parameters (functionality, reliability, usability, efficiency, maintainability, portability and completeness of the documentation) is kept on its own ordinary axis, in the form of a Kiviat diagram (Figure 1).
Initially we listed the various types of outsourcing, and declared the degree of client supervision in various projects. A practice based on mutual project management and methodology-approach to the quality assurance. With three years of experience, practicing this approach, we can show an essential progress in the controllability of outsourced projects. This has resulted in a more reliable risk prevention, better product quality, resource optimisation and not least of all, in better communication and co-operation with vendors.
[Györkös, 1996] J. Györkös, I. Rozman, R. Vajde-Horvat, M. Hericko, "Quality Management in Software Development Process: An Empirical Model", Proceedings of IEEE International Engineering Management Conference, IEEE Press, 1996.
[Paulk, 1995] M.C. Paulk, B. Curtis, M.B. Chrissis, and C.V. Weber, Capability Maturity Model for Software, Version 1.1, Software Engineering Institute, CMU/SEI-93-TR-24, February 1995.
[Slaughter, 1996] S. Slaughter, S. Ang, Employment Outsourcing in Information Systems, Communications of ACM, Vol. 39, No. 7, 1996.
[Rozman, 1997] I. Rozman, R. Vajde-Horvat, J. Györkös, M. Hericko, "PROCESSUS - Integration of SEI CMM and ISO quality models", Software Quality Journal, Volume 6 Number 1 March 1997, ISSN 0963-9314, Chapman & Hall, 1997.
[Nam, 1996] K. Nam, S. Rajagopalan, R. Rao, A Chaudhury, "A Two-Level Investigation of Information Systems Outsourcing", Communications of the ACM, Vol. 39, No. 7, 1996.
[Paquin, 1996] J.P. Paquin, J. Coulliard, P. Paquin, D. Godcharles, "Earned Quality: Improving Project Control", Proceedings of IEMC'96, IEEE Press, 1996.
[PRINCE, 1995] Management of Programme Risk, CCTA The UK Central Computer and Telecommunications Agency, London, 1995.
[PRINCE, 1996] PRINCE 2, Project Management for Business, The Stationery Office Publication Centre, London, 1996.