for any sentence fragment a, and noun b
'a* b' would be equivalent to 'a * b' and 'a*b' and
equivalent to the existing parsing rules.
No invalid syntax cases are created by the proposed
space conventions. But I screwed up in the proposal.
There's a mistake in my original 2 simple rules.
Needs 3 rules instead. Hopefully
1. replace " " to right of noun with ")"
2. replace pattern noun-dyad-space with no space
between left noun and dyad with noun)dyad
3. replace " " to left of noun with closing "("
Ommitted implied rules:
- Left edge is equivalent to a space
- bracketing conjunctions and adverbs according to J
rules still takes precedence over verbs.
So, Going back to nouns a,b,c:
' c+a* b ' in a sentence fragment that could have more
to the left and right parses according to the 3 simple
rules as (c+a)* b... equivalent to the wordier (c+a)*
(b). ' c+a *b ' would also parse to the same
expression... but with wordier version: ' ((c+a)*b) '
' c+a * b ' is also the same expression. Wordy
version: ' (c+a) * (b) '
The equivalent human reading/writing rules are
"evaluate what is grouped together (and separated by
spaces) according to language parsing rules before
applying joining dyadic verbs."
Another way of describing the algorithm (but less
machine friendly) is for any dyad, its left argument
is determined by putting closing brackets around the
first noun that is left bounded by a space, and its
right argument bounded by the noun is right bounded by
a space.
Regarding a preprocessor, could it still be practical
for private use, even if no one else understood the
benefits of the syntax, and was ruled out of being
incorporated into the language?
--- Alexander Schmolck <[EMAIL PROTECTED]> wrote:
> it's a reasonable idea, but maybe a bit tricky to
> get right. One problem is
> that it doesn't scale well at all (distingusihing
> between more than ``a*b``
> and ``a * b`` is down the road to madness, IMO; but
> then what do you do about
> assymetrical ws as in ``a* b`` -- outlaw it?).
>
> Having said that, why don't you try to implement a
> preprocessor for J or
> another language of your choice, and see how well it
> works in practice?
__________________________________________________
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
