Friday, February 03, 2017

Its the data, not the program

Computing is simply about 2 things :
  • Data

  • Program

Which came first ? Of course, the data. What do we mean by data ? Well, data is something that is pure, untouched, virgin, pristine or simply not processed. For example milk that we get from a cow is data. When this milk is converted into yogurt or clarified butter or cheese, then it becomes information, which is another word for processed data.
Image result for robots images
Data in its raw state is like mother nature. As long as we see things objectively, we are witnessing the true side of things. We really don’t need a program to see nature, as it is. This viewing without any affliction’s is called Chitta (a part of our mind). Not many of us see with the Chitta of our mind.
Instead, most of us view nature subjectively through our values, beliefs, ideas, opinions etc. Hence what we see may appear colored, whereas it’s not. This is what happens when we use a program. It converts the meaning of pure data into a higher abstraction called information. This further becomes knowledge (actionable information) and Intelligence (patterns). Finally this intelligence is converted into wisdom through our experience. Wisdom is nothing but the discrimination between good and bad. If the wisdom that we acquire by processing data is good, well then the purpose of the data is served. However, if we stop at knowledge or intelligence, then the interpretation of that data is left to the observer. Every single one of us has a subjective universe, existing in our mind. We spin dreams and stories upon stories to protect what we think is right. If you look at nature, it doesn’t have any contradictions. We do. Because we see things through our lens.
Now what does this have to do with computing ? Well our brain is the algorithm that acts on data collected by our five senses (limited) and interprets it to arrive at some conclusion. A program is simply a set of instructions which operates on data. (or should I say a version of the data). In order for the program to do a good job, it must have the right data. If there are 10,000 data points in a problem, all of them should be available to the program. Sampling or extrapolating won’t be as useful as ‘N’ or all data points. Now you see why MDM (Master Data Management) is so crucial.
GIGO – Garbage In Garbage Out.
With the advent of Big Data, the data from various sources are being crunched to give a unified view of the situation. However, note that too many data sources may worsen the output. Only use as many datasets, as is required.The days when the data used to go where the program is located is becoming extinct. Now the program goes to data. (Hadoop)
The algorithm is getting complex. But even the smartest AI algorithm that we have today pales in comparison to what a single human cell does in a second (3 trillion operations)
Miles to go  before we sleep … (Robert Frost)
The moon is shining in the sky, whether we look up or not.
In a similar way, unbiased or unadulterated data is out there ,for our contemplation, We don’t need a program to realize that (Stillness of the mind).
To end this, I quote the Bible ‘Be Still, And know that I’m God’

No comments: