Notes from software customization talk

| ibm, research

I caught the subway and the bus for the first time on my way to the
IBM Center for Advanced Studies. Mark and I arrived too early for the
lecture, so Pierre Duez showed me around IBM. (I owe him a thank-you
note.) It's a great building, with beautiful themed spaces (Asian,
English cottage, etc.) and game rooms (billiards, air hockey, table
tennis, computer games, gaming consoles). Neato.

The talks started at 11:00. Both presentations came out of term
projects in the requirements engineering course under U of T professor
Steve Easterbrook.

The first presentation was about cognitive anchoring bias in project
estimation, which is when our estimates are affected by the starting
number. High starting numbers result in high estimates, low numbers
result in low estimates.

The second presentation, though, raised goosebumps.
This-is-what-I-want-to-do goosebumps. Sotirios Liaskos talked about
goal-oriented software customization, using Mozilla Thunderbird as an
example of an option-laden program that's hard to customize. I was
blown away by graphs of people's goals and how those goals are
affected by the different options.

I found another name for what I want to do! =) Soft goal analysis.

I want to do that for tasks. I want to analyze the different
strategies people use and _why_ they use them. For example, why people
keep track of contexts, why people use dated or undated tasks… I
want to write something that will profile a person's task-management
preferences and suggest software support. When people want to modify
their task management strategy, I want to suggest step-by-step ways to
achieve their goals.

So basically, I'm looking at:

  1. software support recommendations for task management strategies
  2. support for changing task strategies
  3. customization of task management software using soft goals (harder)

I need to learn how to do things like cognitive work analysis and soft
goal analysis while I'm here. I need to find other people who are
analyzing similar domains so that I can bounce ideas off them.

Here's what I should read next. I'll grab the URLs when I connect back
to the Net.

  • http://www.cs.toronto.edu/~sme/ – readings in requirements engineering
  • Triggers and barriers to customizing software
  • User customization of a word processor
  • An evaluation of a multiple interface design solution for bloated software
  • Reasoning with goal models
  • Simple and minimum-cost satisfiability for goal models

People met today:

  • Ryan from the lab
  • Sotirios Liaskos, who gave the talk on software customization
  • Pierre Duez, person who showed me around
  • Leah, working on software customization of Rational software
  • Steve Easterbrook, professor who taught requirements engg and is now doing experimental methods for software engg
  • Jen, research manager for Pierre(?)

People heard about:

  • Eric Yu, in charge of the mailing lists
  • Alexei, who's looking at goal models for business processes
You can comment with Disqus or you can e-mail me at sacha@sachachua.com.