Brian Schott wrote:
>      I cannot parse ident and I cannot get trace to parse it either.
>      Is there a monadic part a:1 and a dyadic part (u b.1)?
>      Or is there an adverb 1 :(u b.1) within an adverb?
>      Neither of these seem plausible because the a: is not compatible.

Here's ident:
   ident=:1 :'a:1 :(u b.1)'

The core of this statement is b.

According to the dictionary: 'u b. y gives ... the linear representation
of the identity function if y is 1'.  For example:
   + b. 1
0 $~ [EMAIL PROTECTED]

But to be useful I need to convert this linear representation to
a verb.  In ident, I'm using 1 :() to perform this conversion.

A simpler expression than ident might be
   1 :(+ b. 1)
1 : '0 $~ [EMAIL PROTECTED]'

Except, I don't want an adverb, I want the verb, so I need to
throw an argument at it (which will be ignored) to get the
value the adverb produces
   ''1 :(+ b. 1)
0 $~ [EMAIL PROTECTED]

Now I have a verb.  But I need to use 'u' in place of + so
that I can pass in the original verb I'm deriving from.  Thus,
the original 1 :'a:1 :()' formulation.

Does that make sense?

Thanks,

-- 
Raul


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

Reply via email to