Thread: Calculate the sum of all integers in an array using a function and return the sum?

Originally Posted by MutantJohn

So accumulate just takes the beginning and end of the range, sums it all up and then returns it to you. std::vector handles all the things like allocation of memory and whatnot. You can also use std::array with std::accumulate as well! This is where you can really start to see the fundamental power in C++ of iterators and algorithms implemented in terms of iterators.

You probably meant to say this:

This is where you can really start to see the fundamental power of iterators in C++ and the algorithms implemented in terms of iterators.

But yeah, that is pretty useful and cool stuff! I know I can learn it, but I want to make sure to take it in baby steps. I don't want to jump in the deep end of the pool before learning to swim. I understand some parts of it but not all of it. Thanks for showing this to me though! It's always interesting to see how other people solve the same problem.

However, I don't mean to always use consecutive numbers like 1, 2, 3, 4, 5 in my array, or list, or whatever you want to call it. This was just a practice example for me, I may go a different direction later on, and maybe use a difference sequence of numbers. Would accumulate work for something like 1, 2, 4, 6, 8, and so on? I was unable to test it because I got an error on line 11 (expected nested-name-specifier before 'integral').

I don't mean to ask too many questions about this and divert too much from the topic. We can leave some of that juicy stuff for later!

Aren't all programs just algorithms?

We're getting a bit philosophical here, but if a program is just one big algorithm solving a bigger (if well stated) goal, then that program can be composed of smaller algorithms with smaller goals. Bubble sort will put an arrays elements in order by a particular series of steps. It is also true that bubble sort is slower on average than many sorts. That isn't to say that it is bad, but bubble sort has steps and qualities unique to it. So it can be more productive to think of an algorithm, especially a named algorithm, as a specific way to achieve a general, small goal, like searching and sorting. You'll learn a lot of algorithms and invent your own in the course of your study.

So the kind of algortihms that you're referring to here are the ones that are already implemented in C++ standard library, that are ready and waiting to be used?

Yes. The great thing about it is that you don't have to write your own if you use the STL versions. It is practically guaranteed to work as long as it is written correctly, all other things being equal.

Besides, that sounds a lot like the way command line tools work where some of them will do stuff even if you don't specify any argments (options or flags). They just assume some defaults. Is that how this works also?

I do enjoy that you think of it that way, but no. It all comes down to an array of strings that main() receives. Inside of the program, the array is parsed, converted, bleeped, blorped or whatever in order to make the magic happen. Defaults are generally defaults only because main() was programmed that way.

If you are interested in learning command line arguments now, read something like this: A Little C Primer/C Command Line Arguments - Wikibooks, open books for an open world

Now, some of the code on that page might look like a clusterfluck. It can be helpful to write functions to work with argv and argc. As a source of inspiration, there are cool functions or classes out there to help people code what they want out of main(), like getopt(). Unfortunately, getopt() is platform dependent though... as cool of a function it is, Windows does not support the POSIX standard. Otherwise it would have it.

Indeed, by "algorithms" I meant the standard <algorithm> header and its goodies, and by extension the rest of the more modern parts of C++ standard library you'd in the past address as STL (standard template library). Sorry for the confusion, I assumed it was obvious, but I can well understand why it wasn't. My bad.

An example of a function with default arguments:

Code:

#include <iostream>

int sum5(int v1 = 0, int v2 = 0, int v3 = 0, int v4 = 0, int v5 = 0) {
    return v1 + v2 + v3 + v4 + v5; // possible overflow... oh who cares
}

int main() {
    using std::cout;
    cout << sum() << "\n"; // 0
    cout << sum(1, 1, 1, 1); // 4
    cout << sum(1, 2, 3, 4, 5, 6); // error! too many arguments!
}

It works, but the usefulness of that particular piece of code is... Debatable. ; ) Default parameter values are useful from time to time, but as with anything fun - be careful not to overuse it.