By: Ron Jeffries; Ann Anderson; Chet Hendrickson
Publisher: Addison-Wesley Professional
Pub. Date: October 16, 2000
Chapter 7 : Small Releases
This chapter emphasizes the importance of small releases and releasing small
bits that add value. The XP programming methodology claims that these small
releases will keep the programming team confident and productive and gives the
customer feedback and the ability to evaluate.
After explaining their reasons for emphasizing small releases, the book goes
into a few examples of how to split up "all or nothing" kinds of software projects.
Examples included:
1) Payroll System - replace parts of the legacy system one by one, do GUI first...
2) Personnel System - give users better way to input, web interfaces, make new database compatible with the legacy system
3) Tax Package - Start with GUI, replace individual form calculation incrementally with new engine.
4) Distributed Manufacturing Control System - control individual machines in the first, eventually make way to the centralized control.
5) Air Traffic Control System - Distributed Communication and Control (plane to plane)
Chapter 8 : Customer Defines Release

In this chapter, the book explains that the Customer defines what the product is and what is released in an individual iteration. Primarily, the customer writes the stories in a session that constitutes the exploration part of the XP methodology. As the customer writes the stories, the programmers ask questions and the customer answers or asks for clarification. Ambiguity should be avoided at any cost. The programmers will then estimate the time to implement each feature of each story and figure out a velocity for the iterations. Once the iterations are defined a tentative schedule can be made.
The basic steps of the release plan
1. Write enough stories to define a successful product.
2. Do any necessary exploration.
3. Estimate the difficulty of implementing each story. (Story Estimation)
4. Estimate the speed of story implementation.
5. Choose stories for the first release based on business value and difficulty.
Chapter 9 : Iteration Planning
Since the XP methodology prescribes working on small iterations, correct estimation and planning for the iterations is critical and any estimations need to be realistic rather than based on "best intentions."
The three main steps of Iteration Planning:
1. The customer presents user stories - The exploration phase of the iteration planning. The customer explains each story. Understanding of what's needed is important.
2. The team brainstorms engineering tasks - East feature of each story corresponds to different engineering challenges that should be accounted for. If the programmers can't accurately estimate a story based on past experience a spike solution should be made within a day or two.
3. The programmer signs up for and estimates the work - Individual programmers volunteers for different tasks on the story based on what they think they can accomplish. It's preferable that a programmer should be responsible for a story but splitting the story to different programmers is acceptable if expertise on certain engineering issues forces the assignment. Programmers are obligated to help others if they ask for it.
The Design of Future Things
Design of Future Things
By: Donald Norman
Publisher: New York Basic Book
Pub. Date: November 2007
Chapter 3 : Natural Interaction
In this chapter, Norman discusses how to get Machines to communicate naturally without signaling obtrusively or annoyingly. First, Norman explores the idea that audio communication should reveal the nature of the material or action that the machine made or doing, respectively. Norman provides an example of the common teapot where the boiling of the water is used to produce a signal that conveys the degree of readiness of the water inside. Norman also points out that a wide range of communication can be implicitly provided by affordances built into the design of the machine itself. By providing affordances in design, designers can inform the user the range of actions and states the machine has without explicitly saying so. He also mentions of the danger of removing naturally occurring "noise" because it reduces the level of implicit feedback to the user.
Norman also discusses a means of introducing implicit feedback by design by visiting the "tight-rein / loose-reign" manipulation schemes originally found in horseback riding. In other words, he proposes that machines should switch allow for smooth transitions between machine autonomy versus user control.
Finally, Norman discusses the phenomenon and implications of Risk Homeostasis where it has been found that the safety provided by safer designs and technologies is counteracted by increased risk taking activity in human actions. He makes the implication that we should design to make inherent dangers apparent to those exposed to them to promote safer behavior.
Discussion:
Extreme Programming:
I really felt that we were re-hashing ideas already introduced, so I don't have much to say about those. I will say, however, that the rule of always helping programmers that ask for help is appealing to me. Programmers should always be part of a team. The field of computer science and the array of technologies is both deep and wide and the need the supplement our learning experiences and expertise to our peers is necessary to produce great software that we can all be proud of. Furthermore the requirement to always help team members would naturally increase teamwork and thus overall workplace harmony, which is important seeing as how we spend most of our life working.
Design of Future Things:
I feel that implicit and natural feedback is a great topic for exploration in design is great. Decreasing the annoyances while improving feedback is certainly a topic that deserves exploration given the array of digital devices in common use today.
By providing "loose-reign / tight-reign" transition methodologies by design, the interaction scheme could vary by situation where user control is inserted where needed while the machine can in other situations with minimal user input. I think this would increase the degree of symbiosis described in previous chapters and it sounds like a great idea, provided the mode switching is seemless and loss of control is non-obtrusive.
As far as implementing to make dangers more apparent is concerned, I stand almost entirely opposed. Increasing perceived risk would cause undue mental fatigue and stress on potentially (most likely) non-consenting populations. Even if the overall design of the system results in increased safety, we increase psychological harm, overall frustration, and a perceived loss in societal progress that drives future innovation in design. It's just not worth it.
No comments:
Post a Comment