On Tue, Mar 19, 2024, 4:36 AM Zhihao Yuan via Std-Proposals <
std-proposals_at_[hidden]> wrote:

> On Monday, March 18th, 2024 at 7:59 AM, Jonathan Wakely via Std-Proposals <
> std-proposals_at_[hidden]> wrote:
>
> >
>
> > I assume the suggestion is that it would be confusing for matrix[x,y]
> > to return an element of the matrix, but str[i,j] to return a slice of
> > the string. It would be using the same syntax for two different
> > things.
>
> I considered this syntax but did not suggest
> it for the same reason. Moreover, unlike
> Python or MATLAB, we don't have a syntax for
> implied bounds (s[i:] and s[:j]), so s[i, j]
> doesn't buy you much.
>
> --
> Zhihao Yuan, ID lichray
> The best way to predict the future is to invent it.
>

If we can actually spell this s[i:j] (and as Zhihao points out also s[:j]
and s[i:]), then I think that's worth pursuing and useful. In this context,
I think "j" is very clearly a position rather than a length.

But just adding a s.slice(i, j) that looks identical to s.susbtr(i, j)?
Doesn't seem worth it. s[i, j] seems actively bad - you can slice
multi-dimensional arrays too, so using identical syntax for slicing and
indexing isn't a good idea.

Barry

>