I want to talk about indexing. More importantly, I want to talk about how we are explaining indexing to new programmers. But first, I need to lay some background.
Background 1
Right now I'm working as an engineer, but I'm a teacher at heart. I love taking someone through the process of learning something, finding the analogy or way of looking at a concept that really makes it click for them. Especially in the areas of math, mechanics, and programming, where things are very predictable and there's always concrete reasoning behind each concept. I'm a firm believer that students should never just accept what the teacher says on faith, and there should definitely be no magic.
Background 2
I'm taking another JavaScript course to try and fill in the blanks that I have and see if this new teacher explains anything in a new or different way that clicks better with me. Everything was going well until we got to the section on indexing lists. The teacher started off by saying, "In order to get the first item out of the list, you reference it like this: items[0]. You'll just have to remember that computers always start counting at zero." To which I replied:
There's got to be a better way. You shouldn't have to just remember anything. And so, without further background sections, I present to you:
A Better Way
I'm also going to use Python for my explanation because I like it better and I do what I want. Let us consider the string beans. Not string beans, but-- you get it.
The main thing to understand is that indices don't actually fall on a particular item in a list or character in a string (since they are accessed the same way). They surround the letters. They surround the letters. They surround the letters. One more time for the people in the back.
Indices surround the letters.
Thus, when you try this:
>>> fruit = "beans"
>>> fruit[1:3]
"ea"
You get "ea" because those are the letters in between the two indices you specified. When you only specify one index, the program defaults to making the second limit the next index.
>>> fruit[4] # Is the same as fruit[4:4+1]
"s"
The benefit of understanding things this way is that negative indices get a lot less confusing.
>>> fruit[-3:-1]
"an"
In Python, in order to say "all the way to the very end," you just leave the index off. This is mostly convenience and because -0 doesn't make any sense as an index (since -0 technically equals 0).
>>> fruit[1:5]
"eans"
>>> fruit[1:]
"eans"
>>> fruit[:-1]
"bean"
Cool right?
Conclusion
A good number of the 2 people that read this post will have already learned (one way or another) about indexing strings and lists. But do you see now that there is a good reason for starting at zero and ending our for loops at len(string) - 1? It's not magic. It's not because computers are weird and count from zero because that's how their culture does it. There's no little gremlin decreeing new ways to count inside your computer. Now go forth and spread the logical step-by-step teaching approach!
This is a re-post from my blog. Please send feedback! It makes me a better person :)
Top comments (4)
Hi, since I am apparently half of your audience :) I'd like to thank you. I work in a language that is one-based and for the life of me I could not imagine why on earth anyone would like to start counting at zero. I have read religious stories on why one is better than the other, but none of them convinced me.
In the language I work with (OpenEdge 4GL) strings are 1-based. The reasoning behind this is that indices in OpenEdge do not surround the letters, but you merely indicate which letter you want to have. The first letter of a word is letter number ONE, so you specify ONE: SUBSTRING('beans',1,1) gives 'b'. And if you want to have 'an' from 'beans' you would have to use SUBSTRING('beans',3,2).
I still don't think one of the methods is better than the other. They both exist, but your explanation actually makes it easier to understand. It's not that I now find it 'better' than 1-based counting, but at least it is explainable. Thanks!
Interesting! I've done some work in MatLab and I think you can set it up similarly. It always threw me off to go back and forth, but I agree, I can see the benefits of both as long as there is a reasoning behind it.
Hi, I must be the other one of you audience...
Anyway this is a great content for me as I'm teaching programming to my daughters. I'm going to use your post as part of our curriculum.
100% of goal met! π¬ Glad I could help!