Subject: quoting of operators like diff and integrate
From: Robert Dodier
Date: Thu, 11 Dec 2008 22:56:20 -0700
On 12/11/08, Richard Fateman <fateman at cs.berkeley.edu> wrote:
> diff(u,v) is 0
> 'diff(u,v) is 'diff(u,v,1) or some 2-d version of that.
>
> integrate(u,v) is u*v
> 'integrate(u,v) is u*v. By analogy with the above, shouldn't it be
> 'integrate(u,v) ?
>
> This makes it nearly impossible to do something like
> tellsimp ('integrate(u,v), ....), because the first argument becomes
> u*v.
>
> Bug?
I'm not opposed to changing 'integrate(u, v) (and 'sum(u, k, 1, n))
to yield noun expressions. At present those simplify to u*v and n*u
respectively. Seems OK to let verb expressions yield those results.
> If not, how would you produce the noun form of integrate(u,v)?
> '(integrate(u,v)) quotes the verb form.
When I have wanted stuff like that for patterns, I usually do simp:false first.
That's often useful for other kinds of patterns too since helps
ensure that the constructed pattern actually matches what it
appears to match.
Robert Dodier