Mobile device testing

If you've ever wondered about some of the problems with testing mobile applications development, then you might want to check out this article. It contains eight lessons for developers of mobile applications.Some may seem self evident (like "Expect Users to make mistakes") but its amazing how easy it is to overlook them in the heat of development.

My favourite is No 1 - Focus on User Experience. Its clear that smartphones have become mobile computing devices and are increasingly used like this. What is often overlooked is that the smaller screen size especially the small screens of some smartphones like the Blackberry mean that the applications can't be laid out as they are for web applications.

Testing the user experience (often called the customer experience for m-commerce applications) is key to ensuring that the mobile application is a success. And as the different mobile devices are running different operating platforms this means testing on a range of mobile devices.

Test performance

Amongst all the positive noise about virtualization, I noticed what appeared to be a negative article here. It advised that you approach virtualization with caution. The underlying concern was that performance was not considered appropriately. And what did it recommend to address this? Why, performance testing of course.

"Performance Testing and resource consumption needs to be conducted on a few servers or workstations with a proper backup for redundancy and failure. Setup tools need to be built in order to monitor and correct performance issues."

The article goes on to fragmentation issues in the virtual environment and bottlenecks. The latter is of course another reason for performance testing. So on closer inspection, it is not so much an anti virtualization article but rather a pro performance testing article. Positive after all.

Scalability testing

Reading over some of the old entries on this blog, I found an interesting article from way back in 2005 on scalability testing. It looks at 7 steps to running effective scalability tests. What surprised me most was not that the steps are still valid all this time later. What is odd is that it talks about SOA in performance testing. This shows how long it has been around. And the issue of performance dependability on external services remains as problematic today as in 2005.

Spike test

Spike testing is an interesting form of performance testing. According to WIkipedia:

"Spike testing, as the name suggests is done by spiking the number of users and understanding the behaviour of the application; whether performance will suffer, the application will fail, or it will be able to handle dramatic changes in load."

I know you shouldn't use the term you are defining as part of the defintion (in this case spiking in the defintion of spike testing) but if Wikipedia is doing it, who would dare suggest this! The reason that spike testing is interesting is not to do with this defintion though. It is now commonplace for a huge number of people to visit a website when it receives media coverage or is part of a viral exercise on the web. And these huge volumes can take out a website because nobody expected such volumes and therefore didn't plan or test what would happen if they came.

Given the amount invested in web applications it is surprising that the website owners don't conduct what can be a relatively straight-forward type of load test. Maybe this is an area where a little more education on the costs of conducting high volume spike tests against the risks of not doing these performance tests would be really beneficial.