Architecture-Centric Software Project Management: A Practical Guide [Paperback]

As computer hardware provides more functionality at a lower cost, the need for new applications software is exploding. The world-wide-web is providing more information to more people at an ever-faster rate. Software products must be developed quicker, with increased functionality, performance, and quality. The pressure on the software engineers who are developing new products and maintaining existing products is increasing.

This book provides some support to the software project managers who are attempting to juggle the demands of meeting their schedule while delivering features with good quality. Our experience with observing and participating in many software development projects indicates that good design and project management skills go a long way in achieving successful projects. What is very clear is that it is unlikely that projects will be successful when the software architecture is not well designed or project management skills are missing. We have observed the connections between good software architecture and good project management on many projects, and we hope that some of the tips provided will result in better products.

As an industry, we have not been very successful in managing successful software projects. Successful projects are those that meet their planned development schedule, provide the functionality promised, and deliver good quality software. From the 1995 Standish Group CHAOS report, their research on software projects reported that 16% of projects were completed successfully, 31% were cancelled outright, and 53% were substantially over budget and schedule and delivered less functionality than specified. By 1998, more projects were successful, with 26% completed successfully, 28% cancelled, and 46% over budget and schedule with less functionality Johnson 1999. Thus, things are improving, but we still have a terrible track record in our industry for successfully completing software development projects.

The experience for writing this book was gained while managing software design and development projects in Siemens. As part of the Siemens Software Architecture R&D Program, a large number of Siemens projects have been investigated in order to capture how Siemens software architects design software systems. The knowledge gained from this research has been embodied within the four views architecture design approach described in the Applied Software Architecture book written by Christine Hofmeister, Rod Nord, and Dilip Soni Hofmeister 2000. As the four views approach was being developed, we had opportunities to participate as architecture design team members for new products being designed in various Siemens businesses. In some cases, we were also asked to plan and manage these new product developments and implement subsystems of components of the architecture. Thus, our project planning and management methods were developed in parallel with the four views design approach.

A concrete example of this correlation between architecture design methods and project planning methods is our architecture centered software project planning (ACSPP) approach described in chapter 2. We used this approach to develop cost and schedule estimates for the development projects, based on the software architecture. Since we were heavily involved with participating in software architecture design teams, we began to believe in the advantages to be gained when project planning was done in parallel with design. We were also called into Siemens companies as reviewers, from time to time, and we consistently observed warning flags for projects that either were not planned well or were missing a software architecture that could be easily communicated to the reviewers or the development team.

Over this same time period, the importance of software products and good development practices increased. We began to observe that our project planning and management methods were having a significant business impact in that they helped get software products to the market quicker and more predictably. Thus, our initial research interest was focussed more towards design methods and tools, but also we began to see the importance of good project management practices for meeting project goals.

As we became involved with real design and development projects, we realized that the key to effectively applying our methods rested with the working relationship between the project manager and the chief architect. These roles are described respectively within chapter 8 of this book and chapter 12 of Hofmeister 2000. When the chief architect and project manager work together as a decision making team, it's much easier to introduce and tailor the methods we describe herein within a specific development project. We also got to observe the problems that can arise when one individual tries to fulfill both of these roles for development teams that are bigger than a few people.

Not all the project management tips we present in this book are directly related to architecture design. But, enough of them are related, and we feel that a good software architect needs to understand some simple project management techniques in order to be successful within the chief architect/project manager leadership team. Good architects are primarily involved in making technical decisions, but they also appreciate the soft factors involved with leading projects. They often serve as sounding boards or critics of the project manager concerning any decisions that may affect the morale of the developers. Furthermore, we believe that project managers should provide a supportive environment to the chief architect and the entire development team, such that good design practices are consistently applied.

The biggest benefit of working within an industrial laboratory is that you have opportunities to do both research and development. In our case, we've had the opportunity to research methods such as the ones that are described in this book. But, we've also had the chance to apply the methods to real projects. As a result, the methods have been tweaked to be practical such that they are relatively easy to apply. Along the way, we've had the opportunity to design, plan, and develop Siemens software products that have been successfully sold in the market. We have attempted to conceal the real identity of these products in our examples and case studies, since we often describe real world problems that we have had to overcome. But, for the most part, the products that we have applied our design and project management methods to have been successfully developed and sold in the market.

The projects that we have worked on tend to be mid-sized projects. Typically, they have been in the order of ten to twenty software engineers. We cannot claim that our project management methods will work for very large or very small projects, since we have limited experience. Furthermore, since software development projects can last a year or more in duration, we have limited personal experience even for the types of projects that we have worked on. We will conveniently ignore these facts, and we will explain the various tips and methods as if they are great things that every project manager should do. However, there are no silver bullets within what we describe, although it may sound that way at times. A "leap of faith" will be required on the part of the reader to experiment with applying the methods we describe. Unfortunately, this is a persistent problem within most of software engineering since real experimentation is often very limited. Thus, our tips are mostly anecdotal in nature.

The primary audience for this book is software project managers. The secondary audience is software architects who must work closely with project managers. The third audience is software developers who are looking for insights on how to work within a project team, and who may be considering career changes to the first two groups. The book is mainly a collection of tips that could be used for software development projects. Each tip must be tailored to the unique circumstances of the project being worked on.

The tips are structured roughly in the sequence of planning, organizing, implementing, and measuring for which a project manager will be involved. However, there is no strict sequence of steps implied, since these management tasks will be iterated up until the product is released.

It's probably best to begin reading the planning Part II of the book, particularly the description of architecture centered software project planning (ACSPP) in Chapter 2. Once you understand the steps involved with estimating the schedule for your project, you can read the other chapters. The case studies in Part VI of the book should probably be read last. These may or may not be interesting to you, depending on the type of software that you develop. Examples from the case study projects are also included throughout the book.