Though I’m reluctant to put too much faith in surveys – and am astonished when numbers are quoted to precisions of a tenth of a percent – they do serve an important function in painting an impressionistic view of the industry. The data might be fuzzier than related, but surveys can give one a general picture of industry trends and practices.
This year’s 2010 Market Intelligence Service Research data from VDC is out. I have a copy of the highlights, which are revealing.
For instance, offshoring is expected to continue to grow. A third of the engineers responding think their company will increase this activity, with only 10% anticipating a decrease.
Industrial automation continues to be the biggest sector in this industry: 14% work there, with only 12% in consumer electronics. That’s followed by military/aerospace and automotive with similar scores. 6% work in the energy and power sector. I wonder how many of those jobs are in alternative energy?
Firmware is the biggest part of projects. If one excludes test/verification/validation engineers, the software folks make up about half the team (I exclude those as it’s not clear to me if their role is mostly biased in a software or hardware direction ).
I found it quite surprising that 60% of us chose the CPU first and the OS or tools second. Only 12% consider tools before any other component; 22% select the OS first. The latter number is almost exactly the same as the percentage of people using a FOSS OS (in which Linux probably dominates ), so it’s natural to wonder if there’s a correlation between the two. The results are silent on that.
46% of us use a TCP/IP stack in our projects. I guess even electric toothbrushes now need an Internet connection.
Tool use figures are scary: 43% don’t use any form of VCS, a number practically unchanged from a decade ago. Less than half use bug tracking tools. Requirements management tools – and let’s face it, project success is all about managing requirements – score a pitiful 17%.
Only 30% do code peer review. It’s not clear if that’s review of all new code or perhaps a subset. But a solid 27% admit to cowboy coding, coupled with 15% who have no idea what coding methodology is used.
The data shows that, though we do know better ways to build software, a lot of us chose to ignore those approaches.
You can learn more about the survey by going to the VDC Research Website .
Jack G. Ganssle is a lecturer and consultant on embedded development issues. He conducts seminars on embedded systems and helps companies with their embedded challenges. Contact him at . His website is .