Thought Leadership

Embedded software engineering priorities

By Colin Walls

Apart from writing about embedded software matters, I also read widely and enjoy the perspectives offered by many of the key figures in the field. One writer, who always gets my attention, is Jim Turley. Jim is never afraid of raising contentious issues and tends to take something of a chip-centric view of embedded systems.

He seemed surprised, in a recent article, to find that engineers’ priorities might be different …

Jim did a survey and used the “desert island” model to ask engineers to rate which was the key part of the embedded system development process that engineers would rather not change:

  1. the chip
  2. the OS
  3. the tools

He claims to have expected the answer to be (1). The answer he got was (3). I was unsurprised and, in his analysis, Jim dissected the reasons for this preference very well.

Clearly the selection of a microprocessor, microcontroller or SoC for a design is crucial. I do not think that any engineer would think that choosing the device is a low priority, which can be made after other selections have been made. It is definitely part of the process.

Historically, a lot of emotion has been attached to OS selection. This still needs care. At the highest level, it might be: do you use Linux or an RTOS. But it is more complex than that and numerous technical and commercial factors must be considered. A while ago I ran a Web seminar on just this topic – the recorded version is available and I would welcome any queries by email or comment.

Development tools have always been close to my heart, so I am glad to hear that others recognize their priority. It is clear to me that, in my other interests in life, the tools I use really matter: for cooking I have a number of wonderful Le Creuset pans, which behave perfectly and look beautiful; for photography, my main camera is a Sony NEX-5, which may need upgrading soon, but it is a joy to use. Software development tools have to be just right in terms of functionality and ease of use. Every engineer would agree, but they face the problem that engineering managers do not always appreciate their importance. For that reason, they are always on the lookout for cheap or free options. Open source is very attractive and this is the basis for Mentor Embedded’s Sourcery CodeBench.

So, overall I was not surprised by the result of Jim’s survey and it confirms in my mind that Mentor Embedded is applying the right priorities.

Leave a Reply

This article first appeared on the Siemens Digital Industries Software blog at