You can download the book from its website – The Software Conspiracy.

Glenn Alleman of Herding Cats points our attention to the Department of Defense version of the Guide to Project Management Body of Knowledge (PMBOK® Guide). It is a better source of knowledge he says and more than this – it is free. In fact it is called “DoD Extension to PMBOK® Guide” and as they say in the preface:

The primary purpose of this document is to identify and describe defense applications of the core project management knowledge areas contained in the PMBOK® Guide, as well as those defense-intensive knowledge areas not contained in the Guide. It is important to understand that this is an extension to the PMBOK® Guide, and is not intended to be a stand-alone document.

Anyway, the document is free to download and I believe it could be useful source of knowledge to the practicing project managers.

And, at the end, a free e-book from one of the greatest software gurus – Ed Yourdon. His book Death March is still in my personal Top 10 list of all time influencing books on software development.

Craig Brown of Better Projects brought the link to the free e-book Just Enough Structured Analysis to my attention. He says:

[Yourdon] over time has migrated from a view that highly structured processes will improve project results to one where he believes the success factors are quality people and in keeping bureaucracy out of the way.

Ed Yourdon is a world class expert on software development and the book is definitely worth reading. He says in the Introduction:

This book is intended for two audiences: first, the person who is new to the field of systems analysis, and, second, the experienced systems analyst who needs to acquaint himself with systems modeling tools and techniques that have evolved over the past decade.

Happy reading!

You may also find these posts interesting:

• CIO Top 100 Companies For 2008
• How To Motivate Your Team?
• The Two Types of Programmers
• My day-to-day work as a Business Analyst
• The 15 Commandments of the True Leader

Here they are, the 10 mistakes we should avoid in software development:

1. Never committing to project success. This is too obvious, I don’t see any wisdom in it.
2. Freezing the schedule and budget before a project is sufficiently understood. Unfortunately, this happens everywhere, every time. In my opinion, this is the main problem of the software project management: What to do when you are asked to give estimates about the cost and the time, and even to sign a contract before you know what this project is about?
3. Overscoping a solution. Yes, this is a mistake but I believe it happens more and more rarely.
4. Circumventing the application development organization altogether. I don’t think it’s a mistake if you know what you are doing. Sometimes it is the necessary if you’ve been assigned to a “Death March” project. The real question is: Can you still manage to succeed?
5. Underestimating the complexity of a problem. Usually, this is the reason for signing contracts with low budgets and shortened schedules. Unfortunately, this mistake is mostly made by the upper management mostly because they don’t consult with their team. I believe, that every company should invest some time upfront to investigate the customer’s problem ti gain a better understanding of its complexity.
6. Being stingy with subject-matter experts, in which their participation is not sufficient. I think this is also a result of the previous one.
7. Choosing the wrong project leadership. Or the wrong leaders. Or the wrong managers. Or the wrong customers (Can there be wrong customers?)
8. Distrusting managers who have had tasks delegated to them. I would say that not trusting anyone in your team or company is a mistake.
9. Jumping into development without enough research. Again, it happens as a result of underestimating the problem.
10. Suppressing bad news, in which dialogue is insufficient. Do you know a manager who likes to hear bad news?

Well, some would say that I am too skeptic or that I am a naysayer. No, am not. I think there are a lot of problems in our industry but this is my profession and I love it. I really want to discover the biggest mistakes we make, to analyze the reasons for them and to make a plan how to avoid them. Which is really hard. And when we look at our mistakes we just need to be more specific, especially when looking for the causes of those mistakes.

You may also find these posts interesting:

• Krishna Kumar Did An Interview With Me On Software Development
• Classic Mistakes 2008
• A "Classic" Story of Classic Mistakes by Steve McConnell
• Classic Mistakes – GigaLease Case Study, Part 2
• Classic Mistakes – GigaLease Case Study, Part 1

Building up the team

I was so eager to succeed with this project so I and the PM pressed the management very hard to recruit more people to build up our team. Unfortunately, they weren’t looking for the best talents but hired whoever was available.

The first new member of the team was Jannice. She was hired by the HR department without asking for our opinion. Anyway, she claimed to be a very skilled and experienced developer (she was my age) but I gave her a very simple task – to create a form with two simple controls and to bind it to the corresponding record in the database. I knew she needed some time to get acquainted to the development style I established so I gave her a couple of days although I could do it in a couple of hours. Three days later she still had no progress and she told me she still doesn’t understand how to do the task. After a thousand explanations and examples, on the fifth day I understood that she has no idea of software development. She was absolutely useless and from that moment on I assigned her the simplest and the least important tasks. I was still alone and I continued my pressure upon the management to assign more people to the project.

Several months later, when we started to write the user documentation and the help system, Jannice told me that she is really skilled to do that and gave her one more chance. Luckily, she was really good and I was very happy that we finally found her place but we still wasted all those months trying to make a developer out of her.

Next hire was Zoe. This time we insisted to participate at the interview so we were able to ask her some technical questions and we understood that although she was not perfect, she was a developer with some experience and with the great desire to learn and to improve.

Zoe was really fast learning and soon she took responsibility on some of the program’s modules. I was very proud of her for she understood the business logic very quickly and soon she was able to discuss some of the requirements with the customer.

Although Zoe was a hit, the work still remained too much and I kept asking for more developers. Soon the management assigned Frank. He was already an employee but he was famous as the company’s “black sheep”. He had no respect for the company’s rules nor the working hours. I could never be sure when he will come to work and how long he was going to stay. He used to play games during the nights and usually used to come to work at noon. Then he worked for some time and about 5 p.m. left to meet his girlfriend.

Frank was very unproductive and was another peer who reported hours but had no real contribution to the project. We worked on VB6 and it was common situation Frank to commit code that wasn’t compilable (VB6 had that feature “Compile on demand”) and to break off the work of the other team members.

I am still wondering why they kept him in the company. He couldn’t contribute not only to my project, but to all the others he worked on before.

Anyway, we managed to work somehow with this weird team. Unfortunately, we couldn’t hope at all to finish on time.

The first deployments

At some point we were ready with the first release of our product. We decided to make incremental deliveries to show our progress to the customer and to get their feedback as soon as possible in order to understand better the requirements because we didn’t have clear requirements during the whole life of the project.

What was my surprise when I realized that nobody in the customer’s office was using our product! The all boycotted it. Then we started asking questions and a whole new picture arose before us.

The ugly truth

Some time ago, in GigaLease there was a guy, about fifty, who was the chief accountant of the company and who was the only person who kept track of the client’s dues using some magic formulas in Excel spreadsheets. He used to present the reports to his management with no explanation where those figures came from and happily no one asked.

It happened that they hired a young lady on the position of Financial Director and she became a superior manager to the chief accountant. She wanted to know how the reports were generated but he refused. He was insulted of the fact that so young and inexperienced girl happened to be his boss and he boycotted her. Then she decided that their company needs a software product to do the calculations. The formulas will be publicly known and everybody will use that product. And finally, she will show her subordinates who is in command.

Then she came to our company with her 3-page request…

It turned out that there were political intrigues inside that company and they decided the most stupid thing – to solve them using a software program! And our management involved us into somebody else’s war not understanding that nobody wanted or cared about our product. The project was defined as “not important” in our camp and in customer’s. The project was just doomed.

Soon the project was canceled by the customer after nine months of hard work and a lot of problems.

Classic mistakes made

This is a summary of the classic mistakes (according to Steve McConnell) that I see in this part of the case study.

1. Weak personnel (#2). Jannice’s assignment as a developer was a big mistake. We wasted too much time to understand that the development is not her strength.
2. Uncontrolled problem employees (#3). Frank was really uncontrollable. The management had many conversations with him but he couldn’t change. I believe he just should have been fired.
3. Friction between developers and customers (#7). This friction occurred because the customers couldn’t define their requirements during the whole life of the project. And the requirements were just undefinable since the company had very different problems and those problems couldn’t be solved with software.
4. Lack of stakeholder buy-in (#10), Lack of user input (#11), Politics placed over substance (#12). All these mistakes relate to the customer and their political problems. I understand that this is difficult but our management should have understood earlier all those issues and should not have gotten us involved in this farce.
5. Feature creep (#29). The requirements were so unclear and unstable so they were changing all the time.

In my opinion, the biggest mistake was the lack of ability to understand the customer’s problems and needs. We should have understood that their problems were organizational and we couldn’t help with a technical solution. There was no vision about what the product should do and this doomed it to fail.

Do you have such stories? I would highly appreciate your feedback.

You may also find these posts interesting:

• Classic Mistakes 2008
• A "Classic" Story of Classic Mistakes by Steve McConnell
• Classic Mistakes – GigaLease Case Study, Part 1
• Classic Mistakes Forever!
• Top 10 Mistakes in Software Development, According to InfoWorld Tech Watch

How it started

Some time ago I was working at a company called Mega Solutions. It was known as the biggest and the best software company in my country. They claimed that the company hired only the top experts in the area of the software development, used the newest technologies and charged the highest rates.

GigaLease was a small company which offered leasing solutions for industrial equipment and thus had serious customer base and pretty good income.

One day they decided that they needed a custom software product to track all the contracts they had signed, all the credits they had given and all the installments and interests due. So they turned to Mega Solutions to develop such system. They provided a three-page request and they wanted to have the software in four months. The Mega management agreed immediately and they signed a contract before any of the team members was available.

The first clash

I was hired a month later when only the PM was on board. He was a Business System Analyst also and I was the Team Lead and everything else since there was nobody else in the team. Being the only technical person in the team and a person with significant experience in building similar financial systems, the PM asked me to give him an estimate of all the project tasks to help him produce a reasonable project plan. My estimates were more than shocking – I predicted the project to take approximately 12 months if we had a team of 4 or 5 members, depending on their skills. In other words – about 50 man-months. Instead, we had a signed contract for 4 months and a promise from the management that I could have one junior developer added later to the team. It meant that their estimate was about 8 man-months. What a difference, ah!

Ed Yourdon in his famous book Death March defines a project for which one of the constraints – time, scope, or resources – is shortened in half as a “Death March”. What should I say when my project’s effort estimate was shortened at least 6 times?!

Looking for a solution

I and the PM (who agreed completely with my estimates) decided to ask the project sponsor for more staff and for new negotiations with the customer for more time and money since the actual contract conditions were impossible to meet. The answer from the customer about the money was very simple – GigaLease, although very profitable company, had a very small budget for IT solutions and we contracted all of their available money. They were more tolerant on the subject of the total duration and agreed to negotiate it if we reach some significant obstacles during the execution of the project. So, in fact we started the project without a clear idea how many people we would have and without a detailed project plan.

The “Death March”

My concerns about the project’s duration were based on the fact that the customer wasn’t sure what exactly they wanted. The time showed that I was right (as I will explain later). I had some experience with people who wanted to make a revolution in their business with a “small software program” (in their minds) ending with a product, which was a constantly changing and buggy mess of code.

So we shared our concerns with the project sponsor presenting him some simple calculations showing that the project would be a great loss for our company. He said: “Don’t worry! This project is not for profit. It is to make a strategic relationship with the customer and to use their services in the future at a lower cost.” Not only we were put in a “Ultra Death March” and we had to work day-and-night but we had no spark of motivation – there would be no reward, no recognition, no bonus.

This was my first project in the company. I was very flattered when they decided to hire me – it meant that I was one of the best developers in the country since they don’t hire average people. But now I realized that couldn’t prove my abilities – I was entitled “Team Lead” but I had no team and I had no chance to show my leadership skills; the product was decided to be a client-server system build with Visual Basic 6, which was considered already a fading technology, so I didn’t have the chance to improve my technical skills, too.

The classic mistakes made

In this part of the case study I can see the following classic mistakes (according to Steve McConnell):

1. Undermined motivation (#1). Putting the team in a Death March project is a huge risk. Leaving them no straw to catch destroys the motivation totally and gives the project almost no chance for success.
2. Lack of effective project sponsorship (#9). The project sponsor didn’t take our opinion in consideration. He never discussed the great possibility of failure with the customer on a management level. He didn’t approve (at the beginning) our requests for more people in the team so we could finish the project in shorter terms.
3. Overly optimistic schedules (#14). The fact that contract was signed and the schedule was agreed upon without asking the team and without having some estimate was the major reason for the project’s failure.
4. Feature creep (#29). Since the initial request was too general and there was no detailed specification before the contract, it was obvious that a lot of requirements will be revealed during the implementation phase and the feature creep was inevitable.
5. Push me, pull me negotiation (#31). It happened several times when we negotiated more time with the customer. They agreed to give us more time and to delay a milestone but they insisted for some more features in exchange. As a result this made the schedule heavier and more difficult to fulfill.

The biggest mistake in my opinion was putting the team into a Death March project with no motivation. Our company created the “Marine Corps” mentality – if you cannot succeed then you are not suitable for our company. This kind of thinking could never bring a Death March project to success.

What’s next?

The next part of the case study shows how the team was built and how the customer reacted to our product. More classic mistakes!

Meanwhile, if you have some similar projects or similar classic mistakes, tell me about them in a comment, in an email, or in your blog. I believe that telling our stories we can help each other avoiding such mistakes and achieving our projects’ success.

You may also find these posts interesting:

• Classic Mistakes 2008
• A "Classic" Story of Classic Mistakes by Steve McConnell
• Classic Mistakes – GigaLease Case Study, Part 2
• Classic Mistakes Forever!
• Top 10 Mistakes in Software Development, According to InfoWorld Tech Watch