|
|
|
On the Usability of |
|
Software Engineering Research |
|
|
|
|
|
|
|
|
|
|
|
Empirical research engineering is VERY hard…. |
|
Understanding human behaviour (humans are more
complicated and less predictable than computers) |
|
study designs |
|
data analysis |
|
ethics |
|
generalizing results |
|
finding the “right question” to ask…. |
|
False results can be very damaging… a big responsibility |
|
But empiricists also need to advertise the
inherent complexity in this work as well as advertise the potential
benefits |
|
|
|
|
|
Some successes: |
|
Software inspection techniques |
|
Usability of formal methods |
|
Requirements engineering |
|
User studies of software tools |
|
Separation of concerns |
|
Surveys |
|
Case studies |
|
….. |
|
But |
|
Many of these evaluations are done very badly,
increasing the risk we will abandon good ideas, or worse that we spend more
time pursuing bad ideas |
|
|
|
|
|
|
|
For more information: |
|
Collaborative tool demonstration site:
http://www.csr.uvic.ca/chisel/collab |
|
Beg, borrow and steal workshop: http://www.csr.uvic.ca/icse2000/ |
|
My website:
http://www.csr.uvic.ca/~mstorey |
|
Notes