That would be a potential improvement; but there are details - like what
happens if there is an overflow during the scan.
In any case, I was looking for code that's fast right now.
Henry Rich
On 2/22/2016 5:30 PM, Raul Miller wrote:
The problem with the existing implementation is that there's no
efficient way to use it in a scan.
That said, Henry's point about constant time overhead might be relevant.
That said, I don't think the implementation would need to do the full
verb overhead for primitive scalar verbs (where we've already got
tables of their underlying mechanisms for each numeric type).
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
