The Economist welcomes your views. Please stay on topic and be respectful of other readers. Review our comments policy.
You must be logged in to post a comment. Log in to your account.Don't have an account? Register
This is what happens when a computer program
takes the job of people who once administered
the vacation schedules.
Wait until those computer-driven vehicles
encounters black ice on the road.
American Airlines sponsors hundreds of foreigners on H1-B visas to program its computer systems. These "glitches" are to be expected when H1-B people are doing the programming.
In the mid 1980s, I ran across a not-quite joke about software engineering :
When civil engineering was this new, they hadn't invented the right angle yet.
This is still true.
Despite the many, many attempts to write standards and management procedures which would result in robust and fault-free software, we as an industry are still producing cobbled together software products that are amazing only because they work most of the time - despite the chaotic processes used to develop them.
It is true that we are improving, but nowhere near quickly enough to keep up with the demands of the marketplace. And the industry will continue to be mired in this state unless and until computer science is devoted to discovering first principles, rather than just training more technicians to fulfill the job market demand and corporations are content to pay cut-rate programmers to write cut-rate code.
I wish people wouldn't refer to 'an IT glitch'. If a bridge or office building falls down you don't say there was a 'construction glitch' but comment - at least in TE - that 'some floor supports were under-specified' or ' a support beam buckled due to being incorrectly fitted'. Software is engineered - or at least it should be - and we won't get to real software engineering until corporations and governments realise that 'software failures' - for that is what they are - arise from inadequate engineering and the causes can be identified just as forensically as would be done in civil engineering. So let's have some more curiosity about the structure, provenance and production process for large software systems and have reports such as 'the core scheduling algorithm was developed by sub-contractor x but was not subjected to independent verification and validation as would be best industrial practice'.