This is a parsing issue, not a rank problem.
(i.&.>("0) 4 2 3) -: (i.&.(>"0) 4 2 3)
0
It's been helpful to me to use the parts of speech section of the
dictionary rather than the vocabulary page of the dictionary.
Using the parsing table found in
jforc parsing_and_execution_ii
i.&.>("0) 4 2 3
Break sentence into words. $ marks beginning of line.
Words are separated by 2 spaces in this representation:
$ i. &. > ( " 0 ) 4 2 3
I find it much easier to follow the table by
rewriting as $VCV(CN)N
Now parse and execute simultaneously.
Words stack match action
$VCV(CN)N empty no push
$VCV(CN) N no push
$VCV(CN ) N no push
$VCV(C N )N no push
$VCV( C N)N no push
$VCV ( CN) N (CN Rule 6
hmm, " is a dyad.
load'trace'
trace'i.&.>("0) 4 2 3'
which matches as bident. After some searching I
find in
../help/learning/91.htm A1.3.7 Bident Rule
that Conjunction Noun produces Adverb.
Words stack match action
$VCV ( CN) N (CN Rule 6, Bident
$VCV ( A)N (A) Rule 8, Paren
$VCV A N no push
$VC V AN no push
$V C VAN no push
$ V CVA N no push
empty $ VCV AN VCV Rule 4, Conj
VCV produces the anonymous verb
"integers under open".
empty $ VAN VA Rule 3, Adverb
leaving us to evaluate
"integers under open applied to atoms of the noun".
Consider next, i.&.(>"0) 4 2 3
Words stack match action
$VC ( VCN )N VCN Rule 4, Conj
anonymous verb "open atoms"
$VC ( V)N (V) Rule 8, Paren
$VC V N no push
empty $ VCV N VCV Rule 4, Conj
"integers under open atoms applied to the noun"
This works despite the number of atoms changing,
and the fills are empty boxes. In my opinion this
is j working perfectly. No bug here.
On Sat, 2010-03-20 at 22:37 +0800, [email protected]
wrote:
> > 2010/3/18 Frederick W Mellem <[email protected]>:
> > > ? i.&.>("0) 4 2 3
> > >
> > > --------T---T-----?
> > >
> > > ?0 1 2 3 ? ? ? ?0 1 ? ?0 1 2 ? ? ?
> > >
> > > L-------+---+------
> > >
> > > ? i.&.(>"0) 4 2 3
> > >
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
