DB = me, TB = Terrence Brannon
DB> 'abc' anything_I_like"_ 0 'defg'
DB> This will result in
DB> ('abc' anything_I_like 'd'), ...
>Actually the piecing apart of the right noun you do here depends on
>the verb...
Wrong.
It depends on the RANK of the verb, which I declared explicitly, outside the
definition of anything_I_like . I promise you that INDEPENDENT OF THE
DEFINITION of anything_I_like , those are the invocations to that verb that
will be made.
In the message you quote, I specifically distinguished between the definition
of the verb (how that verb acts on its "natural cells") and the intrinsic rank
of the verb.
TB> What you show above depends entirely on the dictionary def of the
TB> verb
No, it depends on the RANK of the verb. This was the entire point of my
message: the verb anything_I_like does not have an definition in the
dictionary! And yet we know precisely what cells will be passed to it.
Here's a thought experiment: What if anything_I_like were a DLL call to a
piece of code in another language? That language is entirely ignorant of
frames, ranks, and cells. Yet it would be passed its arguments with the right
pairings, correctly sized (A).
Here's another:
anything_I_like"_ _"_ 0
would have the same results. As would
anything_I_like"_ _"_ _"_ 0
ad infinitum. There are a finite number of definitions in the Dictionary.
Therefore, there is some general rule being applied here, independent of "the
dictionary def of the verb". The verb isn't chopping up the arguments; rank
is.
If it helps, you may think of every primitive verb P as being defined thus:
kernel =: dictionary def of P as it applies to its natural cells
P =: kernel"natural_rank_of_P
This makes it clear that " is doing the work; partitioning frame and cells,
pairing left and right cells, and replicating singletons. The dictionary
definition of the verb has nothing to do with it.
For the sake of clarity in follow ups, let's standardize on terminology:
P: A primitive verb
Specification of P: The Vocabulary page for P
Intrinsic rank of P: The rank of P, declared in the specification, which
tells you what
the natural cells of P are.
Kernel of P: The English of the specification, describing how P
acts on
its natural cells
Dictionary def of P: Synonym of "Kernel of P" (for backwards compatibility
with previous
msgs)
Using these terms, please make your argument that
"long-frame-rationing/short-frame-replication" for P depends upon the kernel of
P.
-Dan
(A) You could argue that then the verb would depend upon the "dictionary
def" of
the foreign 15!:0 . But guess what ranks that has?
15!:0 b. 0
0 1 1
Which has the same dyadic effective ranks as anything_I_like anyway.
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
