Software Integrity

By K. J. Hamer-Hodges, FIEE

Increasingly, software controls life through Virtual Reality (VR) and Artificial Intelligence (AI). But abstract software is besieged by undetected malware. Furthermore, components cannot be engineered into trusted designs. The deep fakes of programmed abstractions mechanize AI, confusing and corrupting individuals and society. The illusions cover the waterfront of life, starting with hacking and global cybercrime, progressing to industrial catastrophes and, ultimately, Orwellian society.

The Boeing 737 Max is a classic example. The aircraft were all grounded for almost two years, from March 2019 through December 2020. This engineering catastrophe proved how dangerous life-supporting software becomes. Boeing put the total cost of the two deadly crashes at nearly $19 billion as its first annual loss in over two decades was reported by The Guardian in 2020. This cost does not include the losses to Boing's suppliers, clients, staff, and over 300 families suffering personnel loss. 

The aircraft software was redesigned to hide the bigger engine's weight and changed position, deliberately and selfishly taking control away from the pilots. It led to unforeseen aerodynamic conditions resulting in vertical dives, killing all aboard. Other examples exist; worse is to come as the world's dependence on AI and VR using advanced software that lacks sound engineering. Software progress rides on the back of shared, centralized computer hardware that remains unchanged, unimproved, and dangerously outdated. The problem is that computer software lacks the industrial strength of other engineered disciplines in the mold of biotechnologies and semiconductors inherited from the Industrial Age when Charles Babbage and Ada Lovelace pioneered flawless computer science. 

The software cannot be trusted because programs lack clockwork digital integrity. Designers cannot confirm what they claim. Like Boeing's 737 MAX, abstraction cannot be proven true; too many assumptions exist. Some, like the shared hardware, are easily abused. Others, like centralization and superuser privileges, introduce unavoidable scientific mistakes. The centralized power system of first generation computer science is a single point of failure that permanently stains, the future with medieval baronial dictatorship that corrupt and ultimately destroy a balanced, democratic cyber society.

Centralization leads cyber society in the wrong direction, away from democracy back to centralized dictatorship. Citizens, not privileged superuser must take charge of cyberspace if we wish to avoid software domination. Software domination led to the Boeing 737 MAX deaths. It leads to undetected malware and unpunished cyber crimes. From data thefts and stolen intellectual property to well known identity theft, hacking, and Cyberwar, enemies like Russia, China, Iran, and North Korea (RCINK) are at work, intent on America's downfall by speeding centralized software in cyberspace to a dictatorial end game or untrusted AI and VR. 

Trusted, industrial-strength software is vital for a world driven by AI and VR. More specifically computers as digital machines must be engineered to detect every mistake. Fraud, forgery, malware, hacking, and centralized power cause breakdowns, enable crime and prevent software integrity. To defend society as a democracy software must be controlled scientifically, ultimately by distributing power to citizens, the rightful pilots of Cyber-democracy. The privileged skills and the centralized operating superuser make computers ideal partner in cyber-dictatorships, already in place for RCINK. But computer science, as defined by mathematics and logic is uniformly fair, equally for all, no centralization exists, no assumption are needed, no dictatorial power is built in. 

The expression a=b+c is always the same as demonstrated by the abacus for the past 4,000 years. The expression a*b=c is unchanged since the sliderule democratized Napier's logarithmic scale and when the first mechanical computer, Babbage's Thinking Machine democratized mathematical machinery to allow Ada Lovelace to build software as individual error-free, functions, flawlessly programmed as taught at school and university. This scientific solution was defined by the Church-Turing Thesis, in 1936 and it still lies dormant in the centralized, flawed binary computer. 

Indeed, the Abacus, the first industrial-strength computer that dates from Babylon and the birth of Civilization, obeys the same science. Likewise the slide rule and the Thinking Machine. In each case, integrity and industrial strength are achieved by engineering the details as a fail-safe, mathematical framework for abstract computations.