Complex Embedded/Mobile/IoT Software Is Common and May Be The Weakest Link

The public now has expectations about the software in modern devices, such as their car or their phone. The have had experiences with Apps and see news stories, such as:
http://www.nytimes.com/2015/09/27/business/complex-car-software-becomes-the-weak-spot-under-the-hood.html?smprod=nytcore-ipad&smid=nytcore-ipad-share&_r=0

Many of us have been writing and preaching about the activities that should take place in developing such software driven devices. Parts of the software industry know some of the things that should be done. There are many software related books (hundreds including mine), societies (e.g., IEEE, ACM, ISO, and AST), conferences (seems like one every week), schools (Agile, traditional, Dev-Ops, context-driven), and standards about systems, software, and testing (e.g. 12207, 15299, and 29119).
However, the industry is adding software to everything with millions or even hundreds of millions of lines of code. The software is adding functionality to devices and (hopefully) making life better, but the industry (not just the automotive makers) still struggle with how to do software “right”.  A Forester Study indicated only 1% are mature and another 14% are maturing with the others not so mature. At the same time, the public is beginning to demand better protection, regulations, and software. They will accept some software qualities, but as software costs rise in terms of money, time (wasted by users), and company reputation, then getting the right level of software quality will grow as more companies try to become mature in Mobile/Embedded/IoT.
Now what is “good enough” software will vary device-to-device, system-to-system, and even user-to-user. The government will set baselines, courts will determine common law, the public will vote with the money, while manufactures struggle to get “good enough” right. Cars are just the tip of a large iceberg. As IoT grows and the amounts of software and data expand, the software-computer industry will continue to be challenged. In some ways the software challenge is not new. I have been reading about it my whole career (35 years). The industry knows many concepts which can help and argues about others.
The help available to industry is contained in the sources mentioned above, but many of the references are underused or missed all together. There is no one “best” right way or reference, but many software people use only one or two ideals as if there were a “best.” Engineering has to be concerned with heuristics. Because engineering is about heuristics, testers, developers, managers, and support people all need to have knowledge of engineering references and have the practiced skills to do trade-offs between the options. However, it seems that many people are stuck in their “silos” and miss references/ideals until it hurts their software system product.
I will have more about how we are all stuck in silos later.

Advertisements