It’s common knowledge that code bases have continued to grow in size and complexity and that more product end value is reliant upon software. This growth requires organizations to commit more and more of their time and resources to software development.
With OEMs under mounting pressure to find new and more efficient ways to maximize their software content creation without proportionate increases in software engineering man-power, engineering organizations are utilizing more third-party code sources as well as more sophisticated tools that can generate production ready code from higher level designs.
In fact, embedded engineers responding to VDC’s 2012 Software and Systems Development survey indicated only 55% of their in-house developed code was hand-coded for their current project; 17% of their code was auto generated, while 28% was leveraged from other projects. With enterprise engineers, there was even a greater use of these alternate sources; just 52% of their in-house developed code was hand coded for their current project.
Our research confirms demand and use of code generation tools is increasing significantly for both V-model and Agile methods. VDC expects this code source will continue to bear more weight in the years to come as the technology continues to improve. This increasing reliance on third party code is fueling an expansion of the test tool market as companies are held responsible for the overall quality of their end product, including the components received from outside sources.
Given this paradigm, it is no surprise that respondents to VDC’s 2012 survey selected “Impact on Quality” as the most important selection criteria for Static, Dynamic, and Model-based test tools. With the acceleration of software content growth expected to persist, we believe that the evaluation and adoption of automated test tools will be increasingly regarded as critical for quality assurance and as an important instrument for remaining competitive.
The potential costs for software defects in embedded devices, especially those in safety critical applications, are just too severe to leave addressed by manual processes or chance.