The curious incident of the dog in the night
Date: Sun, 11 Dec 2005 14:18:58 GMT
Message-ID: <mzWmf.794$Y2.785_at_trndny04>
Whenever the subject of nulls comes up, I always make reference to the curious incident of the dog in the night.
I'm referring to "Silver Blaze", a Sherlock Holmes mystery by Sir Arthur Conan Doyle. It occurs to me that there are some people who have never read "Silver Blaze" and don't recognize the reference, or think I'm referring to a book whose title refers back to Silver Blaze.
Here's a brief account of the curious incident of the dog in the night, from the point of view of a discussion of missing data.
The dog was set as a watchdog, to guard a race horse.
Something happened to the race horse in the night.
The dog failed to raise the alarm, by barking, and thus failed to alert the humans that were in charge of the race horse.
Later, inspector Lestrade asks Holmes if there is anything Holmes wishes to call Lestrade's attention to.
Holmes replies, "to the curious incident of the dog in the night". Lestrade responds, "the dog did nothing in the night". And Holmes says, "that was the curious incident".
The observation here is that the dog did not bark in the night. The inference that Holmes drew is that the intruder was someone known to the dog.
BTW, I'm no more of an acute observer than Lestrade. I completely missed the fact that the dog DIDN'T bark in the night. It takes an acute observer to see what isn't there. Holmes is an acute observer.
NULLS are different. They are conspicuous.
But the example of the dog in the night is still a useful illustration. A careless reader, reading Silver Blaze, might gloss over the difference between ovserving the curious incident of the dog in the night, and interpreting it. A careless analyst might gloss over the difference between observing a NULL and interpreting its significance. Received on Sun Dec 11 2005 - 15:18:58 CET