On Sun, May 11, 2008 at 7:49 PM, Richard Hennessy <rvh2007 at comcast.net> wrote:
> You claim I said this, "2*inf is not equal to inf", but I didn't
So you want it to be the case that a=b does not imply that
integrate(a)=integrate(b)+C?
> I did say "There is no information loss here which is the overriding concern to me.
Well, your convention doesn't help with diff( if equal(x,0) then 1
else 0 ) or diff( if x<0 then 0 elseif x>0 then 1 else k ).
And what do you propose that diff(5,x) be? The unsimplified form 5*0?
You could of course set things up so that every differentiation
carries around its original argument, e.g. diff(5,x) => 0 <<from:
diff(5,x)>> but then what do you do with diff(5,x)-diff(3,x)? Does
this become 0 <<from: (0 <<from: diff(5,x)>> + 0 <<from:
diff(3,x)>>)>> etc.? Or how about 5-5? Should that be represented as
0 <<from: 5-5>>?
-s