Indexes start from zero because they’re memory offsets, but array[0] is still the first element because it’s an ordinal number, not an offset. It’s literally counting each element of the array. It lines up with the cardinality—you wouldn’t say ['A', 'B', 'C'] has two elements, despite array[2] being the last element.
Zero-based indexing redefines the meaning of the labels “first”, “second”, “third”, and so on. It adds a new label, “zeroth”, which has the same ordinal value as “first” in one-based indexing. The word “first” does not mean “the element with the lowest index” in zero-based indexing.
If you are using a zero-based numbering system, you would absolutely say that array[2] is the final element in the array, that element having the ordinal label “second”, and yet the length of the array is 3 (cardinal). There is no fundamental connection between the ordinal labels “zeroth”, “first”, “second”, and “third” and the cardinal numbers 0, 1, 2, and 3. The similarities are purely an artefact of human language, which is arbitrary anyway. You can make an equally mathematically valid ordinal numbering system that assigns “third” to the element with the smallest index, “fourth” to the next-smallest, and so on. That ordinal numbering system is mathematically coherent and valid, but you’re just causing trouble for yourself when it comes time to convert those ordinals (such as array indexes) into cardinals (such as memory locations or lengths of fencing to buy).
You can make an argument for why one-based numbering is more convenient and easier to use, but you cannot use the notion that zero-based numbering doesn’t make sense given the assumed context of one-based numbering as an argument for why zero-based numbering is invalid.
I encourage you read up what is meant by “zero based numbering” because you and everyone else who has replied to me has tried to use “but that’s not how it works in one-based numbering” as an explanation for why I’m wrong. This is as nonsensical of an argument as trying to say i (the imaginary unit) is not a number because it’s not on the number line. It’s only not a number in the domain of the real numbers. Similarly, zero-based numbering is only nonsensical in the context of one-based indexing.
It does not matter why indexes start from zero. The memory offset argument is only salient if you are using it as an argument for why computers should use zero-based numbering.
Yeah, fair enough. To my mind I guess I don’t think of array indexes as an example of actual zero based numbering, simply a quirk of how pointers work. I don’t see why one starting from zero has anything to do with the other starting from zero. They’re separate things in my head. Interestingly, the article you linked does mention this argument:
Referencing memory by an address and an offset is represented directly in computer hardware on virtually all computer architectures, so this design detail in C makes compilation easier, at the cost of some human factors. In this context using “zeroth” as an ordinal is not strictly correct, but a widespread habit in this profession.
That said, I suppose I still use normal one-based numbering because that’s how I’m used to everything else working.
Indexes start from zero because they’re memory offsets, but
array[0]is still the first element because it’s an ordinal number, not an offset. It’s literally counting each element of the array. It lines up with the cardinality—you wouldn’t say['A', 'B', 'C']has two elements, despitearray[2]being the last element.Zero-based indexing redefines the meaning of the labels “first”, “second”, “third”, and so on. It adds a new label, “zeroth”, which has the same ordinal value as “first” in one-based indexing. The word “first” does not mean “the element with the lowest index” in zero-based indexing.
If you are using a zero-based numbering system, you would absolutely say that
array[2]is the final element in the array, that element having the ordinal label “second”, and yet the length of the array is 3 (cardinal). There is no fundamental connection between the ordinal labels “zeroth”, “first”, “second”, and “third” and the cardinal numbers 0, 1, 2, and 3. The similarities are purely an artefact of human language, which is arbitrary anyway. You can make an equally mathematically valid ordinal numbering system that assigns “third” to the element with the smallest index, “fourth” to the next-smallest, and so on. That ordinal numbering system is mathematically coherent and valid, but you’re just causing trouble for yourself when it comes time to convert those ordinals (such as array indexes) into cardinals (such as memory locations or lengths of fencing to buy).You can make an argument for why one-based numbering is more convenient and easier to use, but you cannot use the notion that zero-based numbering doesn’t make sense given the assumed context of one-based numbering as an argument for why zero-based numbering is invalid.
I encourage you read up what is meant by “zero based numbering” because you and everyone else who has replied to me has tried to use “but that’s not how it works in one-based numbering” as an explanation for why I’m wrong. This is as nonsensical of an argument as trying to say i (the imaginary unit) is not a number because it’s not on the number line. It’s only not a number in the domain of the real numbers. Similarly, zero-based numbering is only nonsensical in the context of one-based indexing.
It does not matter why indexes start from zero. The memory offset argument is only salient if you are using it as an argument for why computers should use zero-based numbering.
Yeah, fair enough. To my mind I guess I don’t think of array indexes as an example of actual zero based numbering, simply a quirk of how pointers work. I don’t see why one starting from zero has anything to do with the other starting from zero. They’re separate things in my head. Interestingly, the article you linked does mention this argument:
That said, I suppose I still use normal one-based numbering because that’s how I’m used to everything else working.