On Wed, Jun 18, 2014 at 9:11 AM, 'Pascal Jasmin' via Programming
<[email protected]> wrote:
> ... it is extremely unlikely that you want to use !. with a "random"
> verb, and so a sentence would be tied closely to !., and its likely
> that the sentence was not intended.

Yes.

Imagine a discussion of the behavior of this code if !. was defined to
work on the contents of named verbs:

   rot=:|.
   shi=: rot!._1
   _3 shi 7#11
   shi=: rot = p.
   _1 shi 2 3 5
...

Ultimately, at some point, there's a judgement call involved where the
language designer and implementer has to ask what the use of a feature
is.

Also, if there does turn out to be a use, there tends to be less code
you have to worry about breaking, when you turn an error case into a
working case than when you change the behavior of something that used
to work.

Thanks,

-- 
Raul
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to