Indices should start at one. Its the joining of the meanings of offset and index that leaves us with arrays that start at zero. Offset is a measure of how far to travel from the start and there it makes sense that travelling nowhere from the start leaves you at the start. Calling the first element the zeroth element is just sloppy.

Maybe if this were C or some sort of assembly then I'd think that using offsets makes more sense than indices.

it seems to me that having indicies starting at 0 makes programming much easier, because this concept removes many playings with adding/substracting 1 when calculating them.

Imagine you do two-dimension array simulation
With 1-based arrays you'll need to do something like this:

$a[3*($i-1)+$j-1]
[download]

and this is equivalent to

$a[3*$i-2+$j]
[download]

which is dumb

Wow, I made SO many errors in code examples,
but those prove my point.

Well, that would be:

    $a [$J * ($i - 1) + $j]
[download]

as opposed to

    $a [($J + 1) * $i + $j]
[download]

where $J is the highest second dimensional index. (Which if indices start with 1 happens to be identical to the number of indices).

If indices start at 1, taking a slice till the end of an array can be written as:

    $a [5 .. @a]
[download]

instead of having to write

    $a [5 .. @a - 1]
[download]

or to have to use the $#array form:

    $a [5 .. $#a]
[download]