> I basically agree with this point.  On these grounds, ! is the wrong
> way round: you are more likely to calculate x ! y for fixed y than
> fixed x. 

What are some examples of fixing an argument of ! ?  The only non-trivial one I 
am aware of is  
2&! = +/@i.



----- Original Message -----
From: John Randall <[EMAIL PROTECTED]>
Date: Tuesday, April 17, 2007 1:01 pm
Subject: Re: [Jgeneral] "J for C Programmers" - asymmetry /     control 
/datadiscussed without definition?

> Dan Bron wrote:
> > John,
> >
> > I wanted you to know that it was just for you that I wrote:
> >
> > JR>  decades (years? centuries? millennia?)
> >
> > instead my original "centuries".  We've had more than one talk 
> about the
> > modernity of mathematical notation.  And now I'm feeling pretty 
> smug tha I
> > hit it on the money with "decades" :)
> >
> 
> Well thanks, Dan: I rose to the bait.  Cajori's book is the reference
> on such questions.
> 
> > Seriously though, I bet, even before Peano, people would be 
> inclined to
> > say "is x is a member of y?" more often than "does y have member 
> x?".>
> 
> I don't think people really thought about sets the way we do now.  You
> could not make up your own: they were more God-given.  Bear in mind
> the latter half of the 19th century was the time mathematicians were
> formalizing the notions of real numbers and integers that we use now.
> It was not until the early 20th century that people realized all of
> mathematics could be built on sets.
> 
> My point with bonding is that x&P should be more common than P&y.  I
> agree that x e. y with fixed y is more common than with fixed x.  In
> particular you are more likely to ask "which elements are in this
> set?"  than "in which sets is this element?"  Mathematicians do 
> use a
> backwards epsilon, corresponding to e.~ .
> 
> > Because J is right-to-left, a primitive  P  should be designed 
> to make:
> >
> >     x P   y =. thing I'm likely to calculate
> >
> > more likely than
> >
> >     y P ~ x =. thing I'm likely to calculate
> >
> > or
> >    (x =. thing I'm likely to calculate) P y
> >
> > Dyadic  i.  follows this design principle.  But that doesn't 
> mean it
> > wasn't a tradeoff.
> 
> I basically agree with this point.  On these grounds, ! is the wrong
> way round: you are more likely to calculate x ! y for fixed y than
> fixed x.  On the other hand, |  is the correct way round.
> 
> I still think it is a matter of convenience and convention.
> Naturality arguments and self-evident truths often don't last.
> 
> Best wishes,
> 
> John
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to