Tying up loose ends

In my last two posts, I mentioned that there has been a change of plans. After this week, I will no longer be EDN’s technical editor for embedded processing. My goal with this post is to share where I was going with the active threads, and to provide a means so that I can contact you when I find a new place to continue discussing the embedded processing space with you. Please email me to let me know if I should try to continue these discussions and how to notify you when I find a new home. In the short term, I have a new website at www.embeddedinsights.com that lists my new contact information.

There are three active threads that are running on this forum. The first thread centers on the technology inflection points series. The second series focuses on robust design principles. The third series focuses on the trend to smaller processors. There is a lot of material that we can explore in each of these series, but there is an overarching set of concepts that I planned to slowly introduce that ties these three series together.

My goals for the inflection point series is not just to point out successful changes in technology, but also to understand the failures – those changes that industry leaders thought were going to change the way we do business, but they failed to measure up in some material fashion. I plan to introduce and expand on two key concepts through this series.

1)      We are reaching the limits of what I currently refer to as the “omniscient knowledge syndrome” – and I think this describes why robust systems are increasingly harder to develop and build.

2)      Design is largely a process of best guesses – and this is a source of increasing uncertainty in contemporary designs.

My goal for the robust design series is to more fully develop the concept that design is largely a process of best guesses. I plan to explore the tools, mechanisms, and design culture based on lessons-learned for dealing with uncertainty in our designs. This exploration includes discussing different schools of principals for dealing with different types of uncertainty and consequences of unintended behavior.

My goal for the smaller processor series is to understand the material time-to-market advantage that lower cost and lower power processors have over “mainstream” 32-bit architectures. I also plan to introduce some of my thoughts on how different organizations of smaller processors, along with changes in processor architectures, software development tools, and how we abstract command control of systems might help our industry deal with the “omniscient knowledge syndrome.” Most notable of these crazy ideas is multi-directional feedback mesh architectures that I first touched on in a multiprocessing options article.

These three threads taken together might allow us, as a community, to explore and verbalize in a consistent fashion possible changes to processor architectures, software development tools, and system verification tools to encourage a much needed technology inflection point that reverses the complexity trend facing today’s embedded developers.

Please help me to justify a new forum to continue these threads by emailing me at my new location so that I can demonstrate how much interest there is for this type of discussion.