This author calls others who use the term "software failure" unsophisticated. I call this author unsophisticated. They are making a pedantic argument. Any seasoned engineer knows exactly how software works and this essay does nothing to enlighten anyone with new information or insight. It only proposes a new narrow definition for "failure" in a sideways fashion and assumes everyone should have been using it all along.
The term 'bitrot' exists for a reason. Most software doesn't exist in isolation, the things it interfaces to directly (eg, libs, the OS, bits in cpu words) or remotely (crypto certs, algs, protocols) will change over time and break it. Even if it truly was not broken before.
I think the author would say that it's "not working" rather than "failing", but I would also say that this is a pedantic distinction. One could also arbitrarily say hardware never truly fails, e.g. if you account for MTBF as part of the design.
3 comments
[ 1.8 ms ] story [ 21.7 ms ] thread>> Software does not decay.
The term 'bitrot' exists for a reason. Most software doesn't exist in isolation, the things it interfaces to directly (eg, libs, the OS, bits in cpu words) or remotely (crypto certs, algs, protocols) will change over time and break it. Even if it truly was not broken before.