Re: Fun with junctions (was Sets vs Junctions)
On Sat, 12 Feb 2005, Damian Conway wrote: @xyz = uniq @xyz; or better still: @xyz.=uniq; Speaking of which, while I think that methods on the implicit topicalizer and the C.= assignement operator are indeed cool, I wonder if any provision will be made for a convenient stand in for whatever is on the left side of an assignment operator, e.g. @xyz=uniq _; # indirect object notation (I guess that _ won't be taken for filhandles any more, I don't know if it's already taken for something else...) Michele -- The trouble with engineers is that given the problem of knocking down a stack of 100 bricks, they will start at the top, and work all day removing them one at a time, while the mathematician will, after a momemt's thought, remove the bottom brick and be done with it. The trouble part, is having to brook the noise of the engineer boasting about how much harder he worked while one is trying to think about the next problem. - Bart Goddard in sci.math
Re: proposal: use \ as none junction delimeter
HaloO Luke,
you wrote:
if $a \ $b == 3 {...}
*If A nor B is 3 ...
What does the * in front of the if mean? Not?
With grammar reason I meant the formal grammar of Perl6
not the one of natural english. Are you aware of such reasons?
In English it's more like:
if \ $a \ $b == 3 {...}
If neither A nor B is three ...
I'm not a native english/american speaker but I think it's
pretty much the same as the German weder ... noch ... for
neither ... nor ... and entweder ... oder ... for the
either ... or ... of xor. Also the plural/singular usage
is the same. So I fully agree with you, that xor and nor have
the same feature/burden as infix ops.
But since xor, ^ and ^^ are in the language already, consistency
would at least allow nor, \ and \\. And nor is a proper english word,
xor isn't ;)
BTW, there also was a thread about ',' and 'then'. There people
complained about the missing 'and' in front of it.
I definitely like the symmetry. But we have to remember that the P in
Perl (can) stand for Practical. The advantages of having syntactic
symmetry are nice, and they aid learning. But they aren't nearly as
important as semantic symmetry/consistency. (Not to say that they
aren't important)
The lack of infix ops to go with the none junction has worried others
before me. Unfortunately I don't find the thread were Larry rambles
about this. The only thing I found was Damian disliking ! and !!.
Of course I wouldn't ask for a nand infix op for two reasons:
1) there's no corresponding junction
2) what should be the high precedence version?
We have a surplus of semantic ideas, and a shortage of characters on the
keyboard. Adding seldom-used (and in this case, confusing) operators in
the face of consistency is a Maxwellian thing, but I don't think it's
hiding any deep truth behind it this time.
Well, not depth but notational convenience for small to medium sized
junctions.
Do you see chances to get that through? Or is the design more or less finished?
Compare:
say nope if $a \\ $b;
say nope unless $a $b;
I find the latter actually reads well in our precious natural languages.
The former just seems like golfing.
I agree fully. The if/unless and while/until pairs are nice lunguistic
concessions of Perl. On the other hand, as a programmer it's not so far
out. Definitly not more than xor---but I repeat myself.
Good night!
--
TSa (Thomas Sandla)
Re: proposal: use \ as none junction delimeter
Thomas Sandla writes:
HaloO Luke,
you wrote:
if $a \ $b == 3 {...}
*If A nor B is 3 ...
What does the * in front of the if mean? Not?
Ungrammatical
With grammar reason I meant the formal grammar of Perl6 not the one
of natural english. Are you aware of such reasons?
Ahh. Using \ as an infix operator doesn't conflict with the grammar of
Perl 6. It's perfectly feasable in that regard.
In English it's more like:
if \ $a \ $b == 3 {...}
If neither A nor B is three ...
I'm not a native english/american speaker but I think it's pretty much
the same as the German weder ... noch ... for neither ... nor ...
and entweder ... oder ... for the either ... or ... of xor. Also
the plural/singular usage is the same. So I fully agree with you, that
xor and nor have the same feature/burden as infix ops.
But since xor, ^ and ^^ are in the language already,
See, I'm trying to change that. The former and probably the latter can
stay, but I'm convinced that the operator form of ^ doesn't huffmanly
belong, even if it does mnemonically.
BTW, there also was a thread about ',' and 'then'. There people
complained about the missing 'and' in front of it.
Yeah, and it got rejected, too. People complain about a lot of things.
That's why I'm here: to comfort people out of their complaints, and to
replace them with my own. :-)
I definitely like the symmetry. But we have to remember that the P in
Perl (can) stand for Practical. The advantages of having syntactic
symmetry are nice, and they aid learning. But they aren't nearly as
important as semantic symmetry/consistency. (Not to say that they
aren't important)
The lack of infix ops to go with the none junction has worried others
before me. Unfortunately I don't find the thread were Larry rambles
about this. The only thing I found was Damian disliking ! and !!.
Of course I wouldn't ask for a nand infix op for two reasons:
1) there's no corresponding junction
2) what should be the high precedence version?
And 3) who the hell would use it?
We have a surplus of semantic ideas, and a shortage of characters on the
keyboard. Adding seldom-used (and in this case, confusing) operators in
the face of consistency is a Maxwellian thing, but I don't think it's
hiding any deep truth behind it this time.
Well, not depth but notational convenience for small to medium sized
junctions. Do you see chances to get that through? Or is the design
more or less finished?
Well, the design is more or less finished in that we have enough
information that the language can be implemented completely. The design
is not finished because it keeps changing.
I never consider the design to be finished (even after it's
implemented). I just assign it an ever-thickening viscosity. If we
wish to make sweeping changes that will change the langauge for the
better, now is the time to do it, because it only gets harder as time
progresses.
Luke
Re: Sets vs Junctions
Jonathan Lang wrote:
Maybe set should be an operator akin to any, all, one, and none,
at least in terms of and |. That is, if junctions are special cases
of sets, why not allow for the creation of generic sets in much the same
way? Then you could have:
# $A and $B are sets,
union($A, $B)=:= ($A | $B);
intersection($A, $B) =:= ($A $B);
not($A) =:= (! $A);
difference($A, $B) =:= ($A - $B);
$e is_element_of $A =:= ($e ~~ $A);
$A is_subset_of $B =:= ($A = $B);
$A is_superset_of $B =:= ($A = $B);
...and so on. Hmm...
I certainly wouldn't have a problem with that. Of course, you could also
easily implement those semantics on top of disjunctions:
class Set is Disjunction {
multi sub set is export(:MANDATORY) (Junction $j) {
return Set.new(values=$j.values)
}
multi sub set is export(:MANDATORY) ([EMAIL PROTECTED]) {
return Set.new(values=@_)
}
multi sub infix:! (Set $a, Set $b){ set($a | $b) }
multi sub infix: (Set $a, Set $b){ set($a $b) }
multi sub infix:- (Set $a, Set $b){ set( $a none($b) ) }
multi sub prefix:- (Set $a) { set(none($b) ) }
multi sub infix:«=» (Set $a, Set $b) { true all($a.values) ~~ $b }
multi sub infix:«=» (Set $a, Set $b) { true $a ~~ all($b.values) }
}
Damian
