Mini-editorial (PGN) 2020 was a crazy year, with all kinds of risks on display. As usual, many of the lessons noted in past issues of SEN and RISKS have been largely ignored, and failures continue to mirror events from the past that have long been discussed here. Issues such as safety, security, and reliability always seem to need more foresight than they receive. Y2K con- tinues to hit somewhere each New Year's Day, when short- term remediations that demanded periodic upgrading have been forgotten. (I suppose old COBOL code will still ex- ist in year 2100, when there may be ambiguities relating to dates that could be 21xx or 20xx (although 19xx is unlikely), and the narrow windowing xes will fail even more dramati- cally.) Election integrity continues to be a real concern, where we are caught in the crosshairs between computer systems and networks that are not meaningfully trustworthy or au- ditable, and the nontechnological risks are still pervasive from unbalanced redistricting, creative dysinformation, poli- tics, Citzens United, and foreign interference. We need non- partisan scrutiny and defense against would-be subverters to overcome potential attacks and inadvertent mistakes. In pres- ence of potential risks in every part of the process, a strong sense of risk-awareness is required by voters, election officials, and the media (both proactively and remedially, as needed).
An improvement of both security and reliability for AES implementations