Round One

The Wednesday meeting started badly and went quickly downhill. To begin with, Tim was late. Being late had been normal for Tim in the past, but Dan had been working with him on it and was proud of the progress Tim had made. When Tim didn't show up at 8:00 A.M., Dan delayed starting for a few minutes, thinking Tim would walk in any moment. Finally, when it became obvious Tim wasn't going to show, Dan sent Jane out to phone him. She returned in a few minutes and said, "I got him at home. He just overslept, plain and simple. Said he'd be here as soon as he could."

"Let's get started, then. He'll just have to catch up," Dan was disappointed, and just a little angry, but tried not to let it show. At least Marilou had remembered the doughnuts. Of course, Tim wasn't there to enjoy them, so Marilou seemed somewhat miffed, too.

click to view at full size

Dan looked at his agenda. "OK, folks, as usual the first item is Agenda Building. Does anyone have any items to add or remove from the agenda, or any adjustments to make?" Marta raised her hand. "Yes, Marta?"

She stood up, and Dan noticed a slight tremble in her hand as she placed it on the conference table to support herself. "Uh-oh," he thought. She had seemed upset at the end of the meeting on Monday, and he'd hoped she would come by his office to talk about it, but she hadn't. He wondered if she would now share whatever was bothering her.

"I've been reading the material you've given us, Dan, and listening to you explain it. I liked the Team Model; it made me excited to be working on this project. When we talked about the background of RMS, I realized that not only was this going to be an enjoyable project, it was also going to be an important one.

"But last meeting, when we worked through the Process Model, I started having my doubts about the work. The concept of iterative planning just doesn't make any sense to me. How can you build something when your plans aren't complete? You put up the Waterfall Model and talked about the MSF method being better, and all I could think about was all the bridges and buildings that have been built over the centuries using the Waterfall method. I'm an engineer by training, and I'm used to doing high-level plans, mid-level plans, detailed drawings, all the way down to subassembly drawings. Only when all the drawings are done and approved do you begin building the bridge or whatever you are building. How can software engineering be different?

"Frankly, the whole process looks horribly flawed to me. I read the documents from your last project, and I think you were lucky that the project came together as well as it did." Marta paused, then took a breath as if gathering herself for a final sprint in a race before continuing. "Dan, I'm just starting out in my career, and I can't trust luck. I think the MSF process is flawed, and I think this project will fail. I can't afford to be associated with such a large failure so early in my career. You need to find someone else to fill the Testing role on the team."

And with that, she gathered up her things and walked out of the conference room.

No one spoke for what seemed like a long time. Everyone looked at Dan to see what he would do, but he simply stood there, looking at the doorway through which Marta had made her exit. Finally, Bill cleared his throat and said, "Do you want me to go after her?"

Dan seemed to shake himself as if coming out of a trance, then looked down at the table for a moment. Finally, without looking at Bill, he said, "No, Chief, it's not your place and it's not your problem. I'll deal with it." To himself he thought, "Not that I have any clue at this moment how I'm going to do that." Again he seemed to shake himself, but this time he looked up at the rest of the RMS team. "Alright, Marta has raised some interesting points. Even though she's not with us to hear the discussion, let's deal with her objections before moving on. Does anyone have a response, either agreeing with her or answering her?"

The rest of the group appreciated the opportunity to talk after the shock of Marta's departure, and a lively discussion ensued. Most of the group was sympathetic to Marta's objections, but wondered if there wasn't something she was missing in her analysis. Jane mentioned the space program. "They surely used the Waterfall Model to do the moon shot, didn't they? Perhaps it's the size of the project that makes it necessary to use that method."

"That's part of it, Jane," replied Marilou. "I think there's another factor, though—the clarity of the goal. In the case of a bridge, or a shot at the moon, you know exactly what your final goal is. You can do all that planning and know it's not a waste of time, because you are all aiming at the same thing, and the target isn't going to change. The bigger and more certain the goal, the more you should use the Waterfall method. Our RMS project isn't like that at all: it's not big, and we're not real sure what we are going to build."

Dan began to answer Marilou when suddenly Bill interrupted. "No, you're both wrong." He jumped up and moved to the whiteboard. Grabbing a marker, he turned to the group to say something and noticed the looks of irritation and hurt on Jane's and Marilou's faces. Realizing what he had done, he stopped. "Sorry, ladies—I didn't mean to be so abrupt. I've just got some background here that may shed some light on this."

He turned to Dan. "You probably didn't know I worked on the space program while I was in the Navy. Nothing big; just part of the data processing team. But I've been thinking back on that, and about what Marta said, and I think I see something I didn't see before."

Bill moved to the whiteboard and drew a rough diagram of the Waterfall Model. "Marta was saying that all engineers know big projects like the space program use this model. We just went along with her on that, but you know what? She's wrong." At the end of the diagram, he wrote moon landing. "If you think about just the moon shot, it looks like a classic Waterfall Model." He drew three MSF Development Process Model circles above the diagram, then said excitedly, "The problem is, we're forgetting the other space shots. We did Mercury, then Gemini, and then Apollo. Each was a 'versioned release,' if you will. And you know what? Inside each of those versions, we did prototypes, and proof-of-concepts, and models for wind tunnels. And every time we did one, we went back and redid the project plan, and the schedule, and the subassemblies, and whatever else needed redoing." Bill was drawing all over the board now, talking while he drew. "All we had when we started was Kennedy's vision statement. We had no idea what 'putting a man on the moon' would look like when we got done. We took what we learned along the way and incorporated it into the next version. It was absolutely an iterative process, and we used that process to carry out one of the most complex, massive development projects ever attempted."

He put down the marker, turned, and looked right at Dan. "You know, I've been something of a skeptic about this MSF stuff. After our meeting on Monday, I went back to my office, shut the door, and went through the documentation from the MSF project at your old law firm. I struggled with it, I admit. When I read that you put out that first version with half the features cut and some of the buttons actually inoperative, it made me cringe. But then I read your user evals."

Bill moved to his chair and sat down. "I was amazed. Your users loved you! I kept reading, trying to figure out why, and then it hit me. Because you locked in the ship date, and then got everyone to buy into the idea of versions. You delivered a product that was both on time and on function. That first win showed the users they could trust you, so they were willing to wait on the second and third versions for the rest of the features.

"The clincher for me, though, was lunch yesterday. You had asked me to line up a team for the prototype, so I took Sam and Beth to lunch. When I explained the approach we were going to use for this project, they couldn't wait to get started. I asked them why, and they said how cool it would be to get a product to the users before the users forgot why they wanted it." Bill smiled wryly and said, "My folks sure know how to tell it straight."

"They learned from the master, Bill," said Jane, laughing. Dan and Marilou couldn't help laughing too, and Bill just shook his head, smiling. Finally he continued, "So I guess you could say I'm coming around to your way of thinking, Dan. I want to see it happen, of course, but I'm actually looking forward to the process of trying to figure out how to make it happen."

"Well," thought Dan to himself, "you win some, you lose some."

Just then Tim walked in. Looking around the room, he asked, "Where's Marta? And where are the doughnuts?"

"And some are rained out," thought Dan, grinning.

"Here, O Sleepy One," Dan shoved the doughnut box across the conference table. "Throw some sugar into your system so you can catch up with the rest of us."

After that, the meeting went fairly smoothly. The team listed their goals for the five iterations Dan had discussed at the last meeting by filling in a chart on an easel. Dan had to help out a few times, but overall it was obvious that everyone had read the material and understood iterations within versions.

click to view at full size

Next he went over the MSF concept of a "tradeoff triangle" and showed how to use it to balance resources, schedule, and features. The team grasped this immediately and was able to give examples of constraining, optimizing, and accepting. He showed them the project matrix, asking them to indicate where the checks should go for RMS. After a moment, Tim had figured out all three checks—optimize schedule, constrain resources, and accept features.

Finally, Dan put a transparency on the overhead projector showing the interim milestones for the Envisioning Phase. The first one, Core Team Formed, drew a chuckle when Dan said, "Well, up to about 30 minutes ago I thought we at least had this one covered." Dan left the milestone checked, though, and moved to the other two milestones.

"We'll do the first draft of the risk assessment document on Friday. Be sure to read the material on risk management that's in your packet." He paused. "I had planned on doing the first draft of the vision statement today, but I think I want to postpone that until Friday." Everyone nodded, and Dan realized that they were all hoping he could persuade Marta to rejoin the team. "Looks like you've got your own interim deliverable," he thought as the meeting broke up.

When he returned to his office, his secretary informed him that the Director of Engineering had already called twice and that he "seemed somewhat upset." Dan returned the call.

"What have you done to Marta, Dan?" the man almost shouted. "She came in and told me she quit the RMS team—something about SMF, or FMS, and iterations. Now she's in her office, and it sounds like she's cleaning out her desk! You want to tell me what's going on?"

Dan hurried to the engineering department and explained briefly to the Director what Marta was upset about. They went to her office to persuade her to stay with both the company and the RMS project. The company was winning, but the project didn't have a chance until Dan brought up Bill's space shot example. Marta thought for a long time, and then finally asked her boss, "What do you think?"

"Marta," he said, wiping his glasses as he talked, "I don't know a lot about the designing end of software. I do know something about the user end, though, and if rapid iterations will get us users a product we need in a shorter amount of time, then I'm all for it. And, I'll tell you one other thing—"

He leaned forward as if about to share a very special confidence. "Frankly, I envy you being on this RMS project. Over here in Engineering, we design the way we do because we have to. Once we put all those rivets in that steel, it takes a lot of work to change it. But with Dan, it's different. You can make changes so much faster, and for so much less money. In a few hours, one programmer can change an entire screen, or even create a new one." He sighed. "If I was younger, I'd jump at a chance to be a part of that, if only for the experience." He put his glasses on and stared at Marta over the top of them. "Marta, you are young. You should jump at this chance."

In the end, Marta changed her mind and agreed to stay with the project. Dan knew beyond a shadow of a doubt that it was the director's comments that had convinced her.

Microsoft Corporation - Analyzing Requirements and Defining Solutions Architecture. MCSD Training Kit
Microsoft Corporation - Analyzing Requirements and Defining Solutions Architecture. MCSD Training Kit
Year: 1999
Pages: 182 © 2008-2017.
If you may any questions please contact us: