I don't know about that. If I want 1e20 | 1.07e21, the current result
(7e19) is better than a limit error, even though it might be off by a
few million.
If the arguments are of such a size that the result is truly
meaningless, limit error seems better than producing a meaningless value.
I don't think it's necessary to give limit error for <: >:
Henry Rich
On 10/16/2017 6:06 AM, Raul Miller wrote:
Presumably, 'too big' would be values where (= <:) is true.
Which, of course, would also suggest a limit error for things like <: and >:
Meanwhile, though, the obvious problem becomes what does a developer
do about errors in production code that did not show up during the
development process?
Thanks,
---
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
