Spiralling cost of error fixing story
Ok, now this tale comes from a lecture I attended at Swansea University in about 1975. Truth is, I don't know if it was made up by the lecturer, whose name I forget, or whether it's a fact.
Googling comes up with nothing.
Anyhow, here it is.
The subject was "How much does an error cost to fix?". The lecturer wrote a single line of FORTRAN code on the board:
DO 10 I = 1,3
"Now when the programmer took this down to be punch-carded onto Hollerith cards the typist made a mistake. They entered:
DO 10 I = 1.3
How much to fix the change from a full-stop to a comma?"
A long discussion followed - nothing if the programmer had spotted the mistake by checking the card immediately, more if it was found during testing and finally a lot if the project had already been delivered.
He gave us the answer:
"$50,000,000. It was in the aerial alignment code of NASA's Pioneer spacecraft. It should loop 3 times to line "10 CONTINUE" in an attempt to acquire a lost uplink signal. However, this error proceeded to set set a variable "DO10I" and set it to 1.3. So the link was lost with subsequent loss of a $50m spacecraft."
I can't remember, was it Pioneer? Was it $50m?
I don't know. If someone remembers this, or actually were in that lecture, then please make a comment.
Googling comes up with nothing.
Anyhow, here it is.
The subject was "How much does an error cost to fix?". The lecturer wrote a single line of FORTRAN code on the board:
DO 10 I = 1,3
"Now when the programmer took this down to be punch-carded onto Hollerith cards the typist made a mistake. They entered:
DO 10 I = 1.3
How much to fix the change from a full-stop to a comma?"
A long discussion followed - nothing if the programmer had spotted the mistake by checking the card immediately, more if it was found during testing and finally a lot if the project had already been delivered.
He gave us the answer:
"$50,000,000. It was in the aerial alignment code of NASA's Pioneer spacecraft. It should loop 3 times to line "10 CONTINUE" in an attempt to acquire a lost uplink signal. However, this error proceeded to set set a variable "DO10I" and set it to 1.3. So the link was lost with subsequent loss of a $50m spacecraft."
I can't remember, was it Pioneer? Was it $50m?
I don't know. If someone remembers this, or actually were in that lecture, then please make a comment.
Comments