Eclipse Testing Day 2012 Round-Up
10.09.2012 12:05 by Alexandra Schladebeck
The third Eclipse Testing Day on 5th September in Darmstadt was probably my favourite so far. It was great to be back in Darmstadt again, and the format of the program this year was, in my eyes, more adapted to the full-day event than in previous years. By having talks of 35 minutes plus questions, and with the replacement of two final talks with an interactive panel discussion, I think the day had a lovely flow.
Testing of Eclipse RCP based Products
Our first speaker was Manuel Bork from Yatta Solutions. His talk on testing Eclipse RCP products highlighted the challenges involved with multiple versions (old versions and bleeding edge), and various configurations (OS and plugin combinations for example) and presented the solutions that Yatta have implemented to make their quality better and reduce their manual testing effort. His comparison of Eclipse with a candy store with so much to choose from is well fitting, and their strategies to perform configuration testing, migration testing and functional testing were well explained. Also, from the Jubula perspective, it’s great that they are using Jubula and extending it
Cutting the right corners
Second on the list was my talk about cutting the right corners. Since the talk was about being pragmatic when faced with challenges in projects, I was really pleased to find that many of the other talks and discussions throughout the day also had a good dose of pragmatism to them. I also stick to what I said during the talk – if I ruled the world, continuous integration would be in place at the start of every new project!
Design by contract
After the break, we moved onto a more developer perspective on testing by looking at contracts in the C4J framework as an Eclipse plugin. As a non-developer, I was happy that the demo was easy to follow, and found the approach interesting. The idea of specifying generic pre- and post-conditions for methods and classes is something that harmonized well with my experience in UI testing, where checking that pre- and post-conditions are fulfilled (GUI state, database state …) is also hugely important.
Iterative model-based generation of Test Cases
Just before lunch, we had our first talk on modeling. It was refreshing to find a modeling talk that didn’t assume that everyone always wants to model – much focus was put on the ability to work with Test Cases or with the model. Using models to gain an overview of complex tests was another interesting point that Raimar and Johannes introduced.
Delta-Model based testing
After lunch, we continued in the modeling vein with the second and final modeling talk. This talk involved a lot more theory and dealt with the difficulty of testing multiple variants or product lines. In order to remove redundancy and make testing possible, only the differences between variants are tested – and the necessary test cases to execute can be generated based on models of the whole product line. Even faced with so many different test objects, we have to remain pragmatic – and find the minimum amount of test cases that will give us the information we need.
Testing for Tool Qualification
The theme for the third Testing Day was “Testing and Beyond”, so our final talks were planned to be not just about testing, but about testing as a part of the bigger picture. Oscar Slotosch was our first speaker, and talked about the work the Eclipse Foundation is doing with Validas to work on Tool Qualification so that companies working in safety critical areas can use the Eclipse tool chain. This was a totally new area for me, and I found the talk interesting and relevant – especially the statement that although Tool Qualification is important, we should find way to reduce or avoid it!
Mind meld with Mylyn
The final talk was an excellent way to zoom right out of our testing corner and look at testing as a part of the whole application lifecycle process. Benny demonstrated how Mylyn can be used to link the tools used by testers, developers and managers so that everyone can talk about the “fluffy clouds of relevance” associated with a project, and make developing, testing, deployment and production more successful.
The newest part of the testing day program this year was the panel discussion. The topic chosen proved to be sufficiently tricky to get people thinking, and varied enough that the discussion seemed to fly by. There were some excellent points made about the value of community and user feedback to teams, how this can be used, how it can be encouraged, and how it is becoming more relevant as a part of a test strategy now that there are so many different versions of web browsers and mobile hardware and software on the market.
After the panel discussion, we had some time to carry on our conversations over a beer and to chat to attendees, speakers and sponsors. Many of the thank-yous were said on the day, but it is definitely worth thanking our sponsors for their support, our speakers for their input, and the attendees for their presence! Over the next few days, we’ll be putting slides (and possibly videos thanks to Linus and Josha!) online, so keep an eye on the Eclipse Wiki page.