Tuesday, January 11, 2022

How Do You Define a Computer Scientist

The last few days have seem some social media discussion about requirements for a computer scientist. Is Calculus a requirement? Is Assembly/machine language a requirement? Among the “yes” and “no” answers there are requests for a definition of a computer scientist. I doubt we could get agreement on a single definition. The mental image people have of a computer scientist is closely related to what knowledge they think a computer scientist should have.

In K-12 education I think we should be very careful and drawing too narrow a definition of a computer scientist. If truth there are many kinds of computer scientist and each kind has its own requirements.

Someone exploring the science of compiler design and code optimization probably does need a solid understanding of machine language. Someone whose focus is on machine learning probably doesn’t as their tools will be at a higher level than the machine. Computer cryptography could probably use a good dose of Calculus. As could some other focuses. Does someone studying user interface? Maybe not.

Should students focus on a special area while in (or before) high school? I don’t think so. They'll have more opportunity to find a focus later. Is Calculus useful? Sure, and for more than just computer science. But there isn’t much that students will do in your average HS CS course that requires it.

The same is true of assembly language. Assembly language clearly gives students greater understanding of the machine below. Some exposure to it is good for programmers and computer scientists. Should it be required? Again, probably not at any depth unless there is a specific need to it.

Back when I was learning CS we had a computer without the memory for an operating system. Programming in assembly language was required. I had to learn how to program subtraction, division, and multiplication on a machine that lacked built-in instructions for those operations. A great learning experience but not one I found much use for later in my (almost 50 year) career. I did work on some projects that required a firm knowledge of the execution speed of different instructions. And knowledge of assembly language helped there. But we don’t see that sort of need except in rare cases these days.

What was required knowledge 50 years ago may not be the same as required knowledge today.  We have to avoid the blind assumption that what we learned and the way we learned “back in the day” is the way things have to be today.

1 comment:

Mark Francis Jaeger said...

As someone who started programming in 1968, I could not have stated this case any better or more succinctly myself.