Wednesday, August 15, 2007

5Qs on Agile

A new interview with me appeared this week on PM Boulevard where I discuss agile project management. It’s really interesting to compare and contrast my answers with those of Steve McConnel. Steve is best known for his book Code Complete and his firm Construx is known for training classes in traditional Software Engineering Institute and academic style software engineering. In my opinion his answers reflect much more of a mechanical iterative approach to agile. My answers really look more at the people aspects - the psychology and sociology involved - and the economic benefits of agile in terms of transaction costs and elimination of waste. Technorati tag: Agile, David+Anderson, Steve+McConnell, Project+Management

Tuesday, August 07, 2007

10th Anniversary of “Singapore Project”

This week marks the 10th anniversary of what has become known as the Singapore Project. It’s real name was Commercial Lending System (II) or CLSII. The II came from the fact that the 3 year old CLS project had just been cancelled without delivering anything. Jeff De Luca was the only manager retained from the first project and he started again and built a new team, mainly from internal staff supplemented with a number of external contractors and consultants. Soon afterward the project became known as PowerLender - a name given to it by Lim Bak Wee the Executive VP of IT (CIO) for United Overseas Bank.

At the beginning we didn’t have any idea of the influence the project would have. What we knew was that the project was big (maybe $20M) and strategic for the bank and that we each had a personal professional stake in delivering it. Jeff had identified several risks from the earlier failed project and chose to mitigate them in several ways. One risk was the architecture and object modeling and to mitigate that risk he brought Peter Coad in to the project. Another risk was the user interface design (we didn’t have fancy terms like user experience in those days) and he brought me in to mitigate that risk.

Jeff always set out to build a great team capable of delivering the project and to identify and mitigate risks. He brought a wealth of experience in managing projects and his own unique approach based on his law that “It’s 80% psychology, only 20% technology.” When we mixed that with Peter Coad’s techniques and his Coad Method approach, what emerged was a very fun way of working that enabled us to work quickly, and maintain high levels of professionalism and quality, to communicate easily and clearly, and to report progress in a fashion that was useful at the team level and right up through the executive ranks to the Chairman of the bank. It took months of gradual refinement and growing team maturity to put in place the methodology that we now know as Feature Driven  Development. Neither Jeff nor anyone else on the team set out to create a methodology. We set out deliver a great piece of working software that would enable UOB to be more competitive. It wasn’t enough for us that the system got delivered. It had to be great! It had to be well received, easy to adopt, intuitive, and most of all capable of delivering on the vision of enabling faster and better lending decisions at all levels in the bank.

By spring of 1999, we had achieved that. A first deliverable version of the system was rolled out. Many of the foreign contractors began to roll off the project. I was one of the first to leave - as the UI designer, I wasn’t required any more. In the latter end of the project I was in charge of the user documentation and manuals. Jeff followed me out the door 1 month later. Stephen Palmer stayed a few months longer. Others, mostly developers stayed up to 1 year longer. Eventually, the project transitioned completely to UOB internal IT staff and remained under maintenance for years and years to come. Perhaps they are still upgrading the system today. [If anyone reading this knows, please comment.]

Along the way, we had pushed the state-of-the-art in Java and OO methods. We’d built one of the biggest enterprise Java systems at the time - about 1.5 million lines of code. We’d evolved the Coad Method in to what we now call Color Modeling. We’d created Feature Driven Development. We’d provided Cliff Berg with lots of material for his book on enterprise Java development. And both Stephen Palmer and I went on to some notoriety in the industry publishing material that we’d development and matured during our stay in Singapore.

For me, the project restored my faith in software engineering and leadership and management in technology. Both Jeff and Lim Bak Wee were great people to work for - great managers, great leaders! Had I not done that project I would almost certainly have quit the profession in my early thirties and gone off to run a completely different type of business.

So, ten years on… Thanks Jeff, Pete, Steve, Paul, John, James, Terry, Lim Bak Wee and all the team from UOB. It was great working with all of you and one of the best professional experiences of my career. Technorati tag: Agile, David+Anderson, Feature+Driven+Development, FDD, Jeff+DeLuca, Stephen+Palmer, Peter+Coad

Monday, August 06, 2007

New Books on Trust

A couple of new books, I’d like to bring to your attention, How: Why How we do Anything mean Everything… in Business (and in Life) by Dov Seidman. I’m half-way through this one already and its a great read. And  Stephen Covey getting in on the trust act with The Speed of Trust: The One Thing That Changes Everything. My copy is lying on my bookcase at the top of my reading list. Technorati tag: Agile, David+Anderson, Stephen+Covery, Dov+Seidman, Trust, Management

Balance Demand Against Throughput

Recently, I’ve changed how I articulate my Recipe For Success. The third point is badly worded because capacity implies a volume and is often used to describe the amount of work-in-progress. This was never really what I intended. I always intended that the input is balanced against the output. The term productivity is commonly used to describe output. However, I like the TOC terminology of throughput as it is unambiguous. Hence, I’ve modified point 3 to…

Balance Demand Against Throughput Technorati tag: Agile, David+Anderson, Software+Engineering, Management

Thursday, August 02, 2007

Challenges Repairing Trust with a Client

Like a fine wine, Clarke Ching seems to improve with age. Here is one of his most recent articles - he’s been so prolific recently - where he retells a tale of facilitating the rebuilding of trust between a client software firm and their customer.

In recent years, I’ve been talking a lot about trust in my presentations, and on this blog. Expect me to write a lot more about trust over the coming years. Meanwhile, here are my two cents to add to Clarke’s story: trust is event driven, little and often is better than huge (but risky) promises offered seldom to never; and collaboration is as Clarke observes a large piece of the solution to lost trust. If you want to get people to collaborate you have to position the collaboration as a game of mutual advantage -what economists and mathematicians describe as a collaboration game. Since, this notion of games drive collaboration hit me recently, I’ve been paying a lot more attention to games in agile development - The Planning Game in XP and Luke Hohmann’s Innovation Games. Technorati tag: Clarke+Ching, Agile, Trust, XP, Planning+Game, Luke+Hohmann, Collaboration