Kitayama, Uchida - Explicit Self-Criticism and Implicit Self
Pair Programming Sarah Lee Grace Uchida Masis Nguyen Michael Hart Jeramy Zapotosky.
-
Upload
wilfrid-stevens -
Category
Documents
-
view
219 -
download
0
description
Transcript of Pair Programming Sarah Lee Grace Uchida Masis Nguyen Michael Hart Jeramy Zapotosky.
Pair Programming
Sarah LeeGrace Uchida
Masis NguyenMichael Hart
Jeramy Zapotosky
Problems to Address
• System-as-is o Pairs usually restricted to no repeats o Tough for professors to quickly check pairings, enforce rules,
very repetitious, and tedious manual processo Tedious peer evaluations through the EEE survey tools
Methods to Solve Problems
• Interviewo Gain insight from professors who use pair programming in
their courses • Prototype
o Created based on interviews with professors• Cognitive Walkthrough
o Each team member will walk through the interface and perform all functionalities
o Note any pros, cons, and problems with the interface • Refine Prototype
o Enhance the interface based on feedback from cognitive walkthrough
• Usability Studyo Experiment with selected professors and students to test the
interface
Project Progress
• Created interview questions • Interviewed professors
o Jacobson, Thornton, Pattis• Sketched mockups on how the protoype should look • Created first prototype as standard Windows desktop application• Transferred the desktop application into a web prototype
Interview Questions
• Are students paired for the entire quarter? Or do students change partners for each project?
• How do you currently organize pairs in the class?• How do you handle a class with an odd number of students?• What kind of restrictions do you place on pairings?• How do you handle grading and evaluation for the pairs?• What kind of information or reporting would you be interested in
seeing throughout the quarter?
Interview #1 - Alex Thornton• Students are required to switch partners for each project.• Students pick their own partners and whoever's left gets paired by
the Professor or TA.• If there is one student left, allows the student to work alone or
make a team of 3.• Uses Professor Kay's evaluation method but always assigns the
partners the same score unless a trend of complaints is seen.• Software would primarily take work off T/A's plate.• Software could encourage students to pair up earlier rather than
last minute.• Emphasizes allowing people to self select rather than automatic
pairing.• Emphasizes that existing resources such as EEE should be used to
gather student and class information.
Interview #2 - Norm Jacobson• Students should pick their own partners in person
o In-person contact is better than randomly selecting on a website since they will need to work closely with their partners
o Solution: Students find partners in person and use the online interface to input/select the name of their partners
• Important: automatic email notificationso When to evaluateo If the student does not have a partner
• Reporto Statistics on numerical evaluations
percentage that thought their partners were good/helpfulo How many students still need to be paired
• Evaluationso Short answer questions are more important
Interview #3 - Rich Pattis• Student pairing policies are relatively free
o Repeated pairings are allowedo Students do not have to change partners for every projecto If there is only 1 student left unpaired, allow the student to
work alone or form a group of 3o Policies might change if there is a system to assist management
• Currently does not have peer evaluation enabled. Grades are the same for both partners in a pair
• Desired functionalityo Automatically computed average evaluation scoreso Encouragement for students to perform pair programming
• Worrieso Privacy issues--will the system be too intrusive/paternalistic?
Insights Gained
• Professors favor customization because every professor is different
• Professors tend to allow exceptions in a class with an odd number of students (solo projects or groups of 3)
• Professors prefer students to find partners on their own (in person) instead of pairing students automatically
• Professors might want statistical information from student evaluations
Problems Encountered
• Unable to contact one professor for interviewing• Conflicts in user interests
o Example: Jacobson disagrees on system-to-be's function of having students be paired up through the online interface with an accept/deny request.
Decisions
• How to improve/finalize the prototype for testing • Who will be selected for usability testing
o Professorso Students
• Where the testing will take place (for students)o Testing for professors will take place in their office
• How to overall prepare for the final report/presentation/prototype
Prototype 1.0
Developed in Visual Studio using C#.Early version ran locally New version runs on the web:http://www.doanthangthien.org/inf132/
Problems that CW Will Address
Completeness: Are all the needed features there? Organization of Functionality: Does where each function is placed make the system easy to understand and use?
Cognitive: Would users be familiar with any terminology and easily recognize features?
Balance of Versatile & Simple: Does the system balance ability for customization of surveys while not making it overly complicated?
Cognitive Walkthrough
Team will evaluate the prototype:• Jeramy, Sarah and Grace do the evaluation.• Tasks to be performed:
o Add an existing classo Create a new classo Create a new class with specific pairing.
Task 1
Tester will navigate through the existing class list:1. Add Computer Science: Operating Systems class– Set options to not allow solo projects, prevent repeats and setup
evaluations.– Add a new Project titled "Memory Allocation."– Setup dates from: June 4th, 2010 to June 11th, 2010.– Do not modify partnering deadline. – Have the System automatically pair partners.
Task 2
Tester will create a new class:1. 1. Add CS 113: Computer Game Development– Set options to allow solo projects, prevent repeats and setup
evaluations.– Add a new Project titled "Solo Game Project."– Setup dates from: August 12th, 2010 to Sept 2nd, 2010.– Set the partnering deadline to August 27th, 2010. – Have the System automatically pair partners.
Task 3
Tester will create a new class:1. 1. Add Informatics 113:Requirements Analysis– Set options to not allow solo projects, allow repeats and setup
evaluations.– Add a new Project titled "Req't Doc 1"– Setup dates from: December 4th, 2010 to Dec 8th, 2010.– Set the partnering deadline to December 6th, 2010. – Pair Bobby Abreu and Shawn Abner.
Task 4: Evaluations
Directly following completion of task 3:User will set up a custom evaluation using any variety of scaled and open ended questions.
Task 5: Student's View
Repeat Task 3 change #6:6. Allow student pairing. 1. Sign in as Bobby Abreu– Pair up with Shawn Abner– Evaluate Shawn
Usability Testing
• A week prior to actual test will run a pilot test with a few pilot users just to tweak and modify the test.
• Testing will be performed on professors initially interviewed: Jacobson, Pattis, and Thornton.
• Testing will be lead by student who did interview.• Additionally Jeramy will do notetaking in the background.• Screen capture will be used while tests are being run.• Tasks will be more refined versions of tasks 1-4 from the CW.
Timeline