CS vs CT vs Coding

# COMMENTS

There's been a lot of buzz recently concernting Computational Thinking (CT) vs Computer Science (CS) vs Coding / Programming on the interwebs.

    Some of the questions and concerns that I've seen recently include:
  • What is CT??
  • Will rich schools get CS and poor only CT?
  • Will rich schools get CS and poor on coding?

The first question is a big one and as a community we haven't answered it yet. I think about it as including things like algorithmic thinking, decomposing problems, iterating, pattern recoginition and the like but I'm hardpressed to come up with a succinct formal definition. This is somewhat related to the bigger problem of "what is CS?" A problem the CS community is still dealing with. There seems to be an editorial or letter in the Communications of the ACM every few months talking about this. "We're an enginering discipline. No, we're a science. No we're an artistic endeavor" and on and on. It's no surprise that the rest of the world doesn't understand what we are when we haven't figured it out ourselves. It's also no surprise that this problem also extends to CT.

One thing seemed clear on the twitter threads and posts I read is that people seem to believe that CT and Coding are potentially overlapping subsets of CS but you can do CT and Coding without CS.

While there's probably a lot to say on the CT without CS or rather CT without programming but today let's look at the programming without CS side of things.

This is probably of particular concern to me as a high school and now college guy. There's no doubt to me that there's a big difference between a kid learning coding/programming and computer science.

Coding can be a useful skill but without at least some computer science it's of limited value. Look at code schools as examples. Many code schools work to prepare their customers for entry level tech jobs. Usually, this includes the fundamentals of programming on top of a typical business toolset. Examples would be a node.js backend, Mongodb, and React.js. Another, albeit older one so probably not taught as much now would be Ruby on Rails. Upon completing the program, one can create a formulaic full stack applicationand knowing that formula can make you employable in our current tech sector. The downside though, as I've heard from countless employers, is that code school graduates, by training, can't break out of that formula. They just don't have the context, nor the problem solving ability - they lack the CS. I'm sure many go on to learn the CS but that's another story. I was talking to a code school founder a few years ago about this and asked "what happens when the formula chages." The answer I got was "that's ok, we'll just bring them back for a new bootcamp." Of course that will cost the kid another 10K or whatever bootcamps are charging these days.

Now of course some code schools try harder than others and some attempt build some CS into their programs but while there are always outliers who can learn a great deal in a short period of time, and often on their own, real education takes time and effort and a short term bootcamp just doesn't have that time.

Back to schools.

To start coding is easier to teach than CS. It can be formulaic and there are lots of resources out there to assist teachers without a CS background.

It's also easy to test and show off.

I was at the NYTM once and they were celebrating a code school project. The idea was really a good one - automatic sensing of heat in low income apartments. I had a couple of questions about getting their data out to the authorities but thought that if they figured that out it could really do some good. On the other hand there was practically no CS. As the project was described - it sets up it's own mesh network and communicates heat readings etc etc., I said to my neighbor in the audience - sounds like an arduino with an XBee shield and yep, when they showed the case, it was the exact form factor to hold an arduino and xbee. That told me there was no CS. The mesh networking is automatically done by the XBee so all they had to do was take the sample arduino code to read the heat sensor and instead of sending it to the serial port as the sample code does, send it to a particular XBee node. Pretty much no CS and almost no programming. Just a great idea and gluing some components together.

I've seen the same with student projects that have been highlighted at events with funders and politicians and in the media. I won't describe them as they are student projects and wouldn't do that to a student but the projects really had no CS under the hood. Nice ideas or even more importantly nice stories around the kid but it was glue and simple basic programming not CS.

Teaching real CS is harder to do, harder to assess and harder to show off which makes teaching coding so enticing.

This is not to say that teaching coding without CS doesn't have value. I believe it does but it also leads to another problem -- it can be assessed improperly either with showy projects that are merely gluing a theme on a fixed pattern or by tests that allow for memorization and this at times has led to students being given too much too early.

We've seen this with algebra and physics. Students end up taking these subjects before they're ready for them. They can pass the standardized test because they can plug numbers into formulas or can recognize what form to use because there are a limited set of problem patterns on the exams but they don't really "get" the subject. This leads to problems later on but of course that's another teacher's problem.

This also happens in coding classes. While it's usually pretty clear when a kid isn't getting a CS concept, I've seen programming experiences where the student at first or in the short term can appear to pass the assessments but later on it's revealed that they weren't ready and they didn't really learn what they were supposed to.

This is a problem. Just as we might have some students exposed to CT without CS I think a major problem will be kids getting Coding without CS. This will particularly be a danger in poorer schools and communities but possibly across the board.

Just another thing we have to really be on guard for as we try to achieve CS for all.