Attribute versus Variable data

I was able to inspect this "T" bud (the green shoot) today and trim the host-branch back to just short of the growing bud. That is a task I was not able to do yesterday.

 

It is convenient to separate types of data into two very broad categories:

• Attribute data
• Variable data

Of the two, variable data is much more valuable for problem solving.

Attribute data

Attribute data is binary, yes/no information.

• Pregnant/not pregnant
• Pulse/no pulse
• Edible/not edible
• Minor/adult

Variable data

Variable data is continuous.

• 3.14159/3.14160
• 98.6F/101.5F
• 2678fps/2702fps
• 6574.0/6574.9

Good attitude

People continue to comment on my "good attitude". In part, it is due to my using variable data wherever possible.

So far, every day has given me a gift of things I could do that I was not able to do the day before. Sometimes it is small, maybe just another half-inch in how far I can swing a limb. Other days it is huge, like being able to put weight on my left leg.

Attribute data has its uses. It is fast and decisive. That makes attribute easy-to-use for triggers.

But in general, if you have a problem you are grappling with, tracking variable data is far more robust than using attribute data.

When does attribute data become variable data?

This is not a trivial question. Precision costs money and ever finer increments cost exponentially more. So how much precision is enough?

That is dictated by the problem. If I want to raise cattle for the market a scale that measures in 25 pound increments might be precise enough for me to evaluate different types of cattle feed. If I am going to be competing in long-range target shooting, I might need a scale capable of measuring to the 1/20th of a grain. That is a difference of 3,500,000 in precision.

Mathematicians to the rescue

Mathematicians tell us that data starts behaving more like variable data than attribute data when the precision is sufficient to separate the data into seven or more "bins".

And you thought Microsoft Excel used seven as the default number of bins for histograms because it looked nice. Nope. The 957 Math Ph.Ds who work for Microsoft agreed that seven was the right number for mathy reasons.

Minor/Adult, 6574.0/6574.9 days old

I don't comment on mass-shootings unless I feel I have something unique to add. Since I have no first-hand knowledge of any shootings to-date, I don't comment.

Boys?

 Violent crime is a problem, not just mass casualty events. A disproportionate number of them involve young men who recently turned 18. A VERY large percentage of the time these individuals were already known to local law enforcement officials.

Unfortunately, the current way of looking at "young people" is attribute data. They are either considered minors or adults and there is a fire-wall between the two categories.

Many school districts in the United States separate students, based on age, into Pre-elementary, Elementary, Intermediate, Middle and High School. They do this based on expert's advice that these age groups have different cognitive abilities, different needs and present different educational challenges.

Throw in folks over age 80 and you have the seven categories that turn it into variable data.

But how would it be useful?

This is just a stake in the ground, a starting point for discussion, but what if any minor, Elementary or above who committed a homicide had that stay on his permanent record and followed him into adulthood?

What if any minor who was Intermediate or above who committed armed robbery or aggravated assault had that stay on his permanent record.

Or any minor who is Middle or above who Breaks-and-Enters has that stay with his record.

I am not married to this-crime-but-not-that sticking. I am married to a sliding scale that selectively softens the impenetrable firewall that allows people who are documented as being violent to walk unfettered among us.