In the following sentences where c is a conjunction v
a verb, and s a sentence fragment.

sx (m c n) v y
s x(m c n) v y
if x is a noun, then (m c n) has to be a dyad,and
sx (m c n) v y is (s x) (m c n) v y
s x ... is s (x(m c n) v y) ie. regular parsing

if x is a verb, then even if (m c n) is unknown...
s x(m c n) v y is the only case that would imply
special bracketing. 
sx (m c n) v y parses the same whether (m n c) is noun
or verb. Same with s x (m c n) v y.
Probably the best thing to do is to treat x(m c n) v y
as intending (x(m c n)) v y.  The issue though is that
if (m c n) is a verb then a train is formed.  The
reason for bracketing is assuming that the programmer
was intentionally assuming either a noun and/or the
automagical bracketing by sticking x to the
conjunction without space.

Also, if somehow the above turns out not to work,
presuming that conjunctions and adverbs will produce
verbs is workable. 1. Its the usual reason for
creating an adverb or conjunction, and 2. If the
programmer wants special bracketing accross noun
producing modifiers, then he can do it the old
fashioned way().

--- "Miller, Raul D" <[EMAIL PROTECTED]> wrote:

> p j wrote:
> > I'd also need to know whether specific
> > conjunctions/adverbs evaluate to a verb or noun
> (after
> > their arguments processed).  I'm guessing there's
> only
> > 2 ways to find out: 1. execute the conjunction
> (not
> > safe obviously), or 2. Would need J support for
> > classifying user-defined conjunctions by their
> > possible return types.  
> 
> In general, the only completely valid way to
> determine
> adverb/conjunction result type is to evaluate the
> adverb
> or conjunction.  J itself does not know except
> through
> evaluation.
> 
> However, you don't have to be completely general
> when
> transforming programs.  It's perfectly alright to
> punt
> at some point -- you don't have to solve the halting
> problem, you just need to document where you give
> up.
> 
> A classic example of a conjunction which can return
> any
> of the four major syntactic types is :
>  0 :'', 1 :'', 2 :'' and 3 :'' all have different
> syntactic
> types (by the way: all these functions have empty
> domains
> because none produce a result -- but it's the :
> conjunction
> I'm using as an example here, not the result of :).
> 
> A slightly more abstract case of this would be 2 :'m
> :n'
> 
> -- 
> Raul
> 
> 
> 
>
----------------------------------------------------------------------
> For information about J forums see
> http://www.jsoftware.com/forums.htm
> 


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to