On 2/5/08, Yuvaraj Athur Raghuvir <[EMAIL PROTECTED]> wrote:
> Explicit statement for assigning names makes me wonder if the foreign
> conjunctions are limited in their use due to this restriction.

Well... you could use #@(3!:1), but this measures something different
(in essence, the space it would take to serialize the data).

   (#@(3!:1) % [EMAIL PROTECTED]) a
27.3333

You could also define an explicit verb to perform the operation you
seem to be trying to get at:

   sz=:3 :'7!:5<''y'''
   (sz % [EMAIL PROTECTED]) a
10.6667

Perhaps 7!:5 should have instead been design to have the
interface that I used for sz, but then I do not know how you
would deal with something like:
   mean=: +/ % #
   7!:5 <'mean'
512

> On the other hand, given that I have a named noun available with me (say the
> a below), is it possible get the boxed representation of the name?

Sure -- supply the name then box it (which is what I
gave you in my first reply).

If you are asking whether the anonymous data can backtrack to find
the name it came from, for the cases where it had a name... I suspect
the answer will always be 'not in the general case'.  You probably can
do something by enabling debugging, forcing an error and then examining
your stack -- but I do not think this effort would be worthwhile.

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

Reply via email to