I'm not following your thought process on the special case for
complementary indexing. Can you show me where this helps?

Thanks,

-- 
Raul


On Thu, Dec 14, 2017 at 2:29 PM, Henry Rich <[email protected]> wrote:
> !!  That's a GREAT idea!  Having to box the damned indexes has been a
> recurring trouble-spot for me, and when it matters for performance
> workarounds are hard.  I have never had a case where I was tempted to have a
> multidimensional array of atomic indexes for amend, and of course if one
> came up I could just enfile it.
>
> I would change the proposal slightly:
>
> In x m} y, each raveled item of m specifies a selector for y [this selector
> can be an atom, a numeric list, or a list of boxes.]  If the selector
> contains more than 1 atom, an additional level of boxing is added.  The
> resulting
> selector may not specify complementary indexing.
>
> Anybody see anything wrong with that?
>
> Henry Rich
>
>
> On 12/14/2017 2:11 PM, Raul Miller wrote:
>>
>> That said, if this had been done in the original implementation, then
>> a sentence like:
>>
>> "If 2>#$y we ravel each item in the indices"
>
>
>
> ---
> This email has been checked for viruses by AVG.
> http://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