I was just wondering if J has something equivalent, not asking for a
language extension.
Key, as suggested by Mike, can be used to achieve a similar effect (just
not quite the same).

As for a[x] +=: a[x], I wonder what implementation would be more useful :
- creating a temporary array with a[x] and then going ahead with the +
- doing the additions in place using the elements of x in catenated
row-order

On Sun, Nov 3, 2019 at 8:49 PM Henry Rich <[email protected]> wrote:

> I'd say that scratches the surface.
>
> First you would have to define something akin to a[x], which is not J
> syntax.
>
> Then you would have to decide what assignment to a[x] means when x has
> repeated indexes.
>
> Then you would have to decide what a[x] +=: y  means when x has repeated
> indexes.  Does it impose an order of operations?  Do you insist that it
> work atom by atom, as if we were running on a 68000?
>
> What would a[x] +=: a[x] give?
>
> What about a[x] +=: a[a[x]] ?  In what order are the updates to a to be
> made?
>
> I think you would end up leaving a large part of the spec undefined.
> That might be OK is the defined bit is very useful.
>
> hhr
>
>
>
>
> On 11/3/2019 8:08 PM, Raul Miller wrote:
> > Two tricks here:
> >
> > (1) Designing the altered parser table to handle this case (without
> > breaking existing code), and
> >
> > (2) implementing it.
> >
>
>
> --
> This email has been checked for viruses by AVG.
> https://www.avg.com
>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
>
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to