Comment1
Comment2
Reference
Lowering the barriers to website testing with CoTester
Jalal Mahmud, Tessa Lau
IUI'10, February 7-10, 2010, Hong Kong, China
Summary
In this paper, Mahmud and Lau introduce a system called CoTester which primary goal is to decrease the difficulty of web applications testing. The authors express that the current tools for web testing require programming knowledge and maintenance which tends to be time-consuming. Their system CoTester is built on CoScripter platfrom and its main goal is web applications testing. CoScripter is a scripting language which supports assertions and does not require advanced programming skills to use. CoTester uses subroutines within a testing unit to improve test management and also tries to provide the tester a visual structure of the changes being made. With the use of these subroutines a change can be automatically made to similar instances of a certain process, for example changing the log in process of an application. In order to make testing easier, this system relies on a machine algorithm that identifies subroutines in testing scripts. Mahmud and Lau claim that their algorithm is capable of recognizing subroutines with 91% accuracy in a subset of seven subroutines. The recordings in testing scripts are recorded by actions in a web application. The recording can include objects present or not present, links, presence of text, and buttons clicked.
 |
| Figure 3 from paper |
Discussion
The last two papers I read have dealt with website research. I found this paper to be extremely boring and although there are some interesting aspect such as the script subroutines, I do not agree with the author's ideas. First of all is someone is going to develop website applications, I would expect that person to have an advanced knowledge of computer programming. This system is trying to implement a scripting programming language that can be used by people who do not necessarily have programming knowledge. CoTester is trying to aim towards programming by display which I do not disagree with. However I still think someone with programming skills should be the one doing this. The accuracy of their algorithm is pretty high and might be useful in another application, but the main purpose of this paper to me is irrelevant to future web technology.
There are a lot of people with no skill at all that try to create website applications by copying and pasting bits of code together. For those people, this would be a helpful system. But I do agree that someone with programming skills should be doing this, since they are more likely to apply good design techniques and such.
ReplyDelete