ICER 2012 Day 1: Discussion Papers Session 1

ICER contains a variety of sessions: research papers, discussion papers, lightning talks and elevator pitches. The discussion papers allow people to present ideas and early work in order to get the feedback of the community. This is a very vocal community so opening yourself up to discussion is going to be a bit like drinking from the firehouse: sometimes you quench your thirst for knowledge and sometimes you’re being water-cannonned.

Web-scale Data Gathering with BlueJ
Ian Utting, Neil Brown, Michael Kölling, Davin McCall and Philip Stevens

BlueJ is a very long-lived and widely used Java programming environment with a development environment designed to assist with the learning and teaching of object-oriented programming, as well as Java. The BlueJ project is now adding automated instrumentation to every single BlueJ installation and students can opt-in to a data reporting mechanism that will allow the collection and formation of a giant data repository: Project Blackbox. (As a note, that’s a bit of a super villain name, guys.)

BlueJ has 1-2M New users per year, typically using it for ~90 days and all of these users will be able to opt-in, can opt-out later, although this can be disabled in config. To protect user identity, locally generated (anon) UUID will be generated and linked to user+installation (So home and lab won’t correlate). On the technical side, the stored data will includes time-stamps, tool invocations, source code snapshots, and course-grained location. You can also connect (locally available) personal data about students and link it to UUID data. Groups can be tagged and queries restricted to that tag (and that includes taxonomic data if you’re looking into the murky world of assessment taxonomy).
In terms of making this work, ethical approval has been obtained from the hosting organisation, for verified academic researchers, initially via SQL queries on multi-terabyte repository but the data will not be fully public (this will be one of largest repositories of assignment solutions in the world).
Timescale: private beta by end of 2012, with a full-scale roll out next Spring, AY 2013. Very usefully, you can still get access to the data even if you don’t contribute.
There was a lot of discussion on this: we’re all hungry for the data. One question that struck me was from Sally Fincher: Given that we will have web-scale data gathering, do we have web scale questions? We can all think of things to do but this level of data is now open to entirely new analyses. How will we use this? What else do we need to do?

Evaluating an Early Software Engineering Course with Projects and Tools from Open Source Software
Robert McCartney, Swapna Gokhale and Therese Smith

We tend to give Software Engineering students a project that requires them to undertake design and then, as a group, produce a large software artefact from scratch. In this talk, Robert discussed using existing projects that use a range of skills that are directly relevant to one of the most common activities our students will carray out in industry: maintenance and evolution.

Under a model of developing new features in an open-source system, the instructors provide a pre-selected set of projects and then the 2 person team:

  1. picks a project
  2. learns to comprehend code
  3. proposes enhancements
  4. describes and documents
  5. implements and presents
The evaluation seeks to understand how the students’ understanding of issues has changed especially regarding the importance of maintenance and evolution, the value of documentation, the importance of tools and how reverse engineering can aid comprehension. This approach has been trialled and early student response is positive but the students thought that 10,000 Lines of Code (LOC) projects were too small, hence the project size has increased to 100,000 LOC.

A Case Study of Environmental Factors Influencing Teaching Assistant Job Satisfaction
Elizabeth Patitsas

Elizabeth presented some interesting work on the impact of lecture theatres on what our TAs do. If the layout is hard to work with then, unsurprisingly, the TAs are less inclined to walk around and more inclined to disengage, sitting down the front checking e-mail. When we say ‘less inclined’, we mean that in closed lab layouts TAs spend 40% of the their time interacting with students, versus 76% in an open layout. However, these effects are also seen in windowless spaces: make a space unpleasant and you reduce the time that people spend answering questions and engaging.

The value of a pair of TAs was stressed: a pair gives you a backup but doesn’t lead to decision problems when coming to consensus. However, the importance of training was also stressed, as already clearly identified in the literature.

Education and Research: Evidence of a Dual Life
Joe Mirõ Julia, David López and Ricardo Alberich

Joe provided a fascinating coloration network analysis of the paper writing groups in ICER and generally. In CS education,  we tend to work in smaller groups than other CS research areas and newcomers tend to come alone to conferences. The ICER colouration network graph has a very well-defined giant component that centres around Robert (see above) but, across the board, roughly 50% of conference authors are newcomer. One of the most common ways for people to enter the traditional CS research community is through what can be described as a mentoring process, we extend the group through an existing connection and then these people join the giant component. There is, however, no significant evidence of mentoring in the edu community.
Unsurprisingly, different countries and borders hinder the growth of the giant component.
There was a lot of discussion on this as well, as we tried to understand what was going on and, outside of the talk, I raised my suggestion with Joe that hemispherical separation was a factor worth considering because of the different timetables that we worked to. Right now, I am at a conference in the middle of teaching, while the Northern Hemisphere has only just gone back to school.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s