On Jan 25, 12:27 pm, Jason Grout <jason-s...@creativetrax.com> wrote:
> On 1/25/11 11:17 AM, kcrisman wrote:
>
>
>
> >> If f is a function explicitly of one variable (e.g., f(x)=x^3+1), then
> >> it makes sense to use this variable as a default for differentiation or
> >> integration. However, if we are dealing with just a symbolic expression
> >> (e.g., f=x^3+1), then a default makes less sense to me.
>
> > Correct, that was the rationale. I disagree, if there is only one
> > variable in the expression.
>
> (to recap discussion from before for those that weren't on the list at
> that time...)
>
> I agree with Robert on this one. If you have symbolic expressions:
>
> var('x,y')
> a=x+y
> b=-y
>
> then it seems you would have integrate(a+b) behave differently than
> integrate(a). That's why it makes less sense to me; to avoid confusion,
> the user should specify intent somewhere. In the f(x)=a+b case, the
> user is specifying that a+b is only a function of x, so the integral can
> use a default variable without confusion.
Wait, now *I'm* confused. Are you suggesting that
integrate(b)
should or should not have a deprecation/require integrate(b,y)? In
particular, I'm not sure what agreeing with Robert means :)
My view is that if there is only one variable (broadly interpreted,
i.e. x^n does not count), we might as well figure the user isn't
totally stupid and let them omit the variable of integration (though
of course allowing for integrate(b,x)!). Otherwise it is reasonable
to raise an error. Same for differentiation.
> >> IIRC, previous conversations about this centered on the symbolic
> >> expression case, not the explicit one-variable function case. I might
> >> be remembering incorrectly, though.
>
> > No, I think you're right, but the current behavior is the same in both
> > cases. Also, note that the deprecation on integrals ignores this
> > distinction - again, wrongly, in my view.
>
> If the deprecation on integrals ignores the distinction, then maybe a
> ticket could be filed which deprecates part of the deprecation :).
Perhaps. I seem to recall Burcin being part of this discussion as
well, since it was about the time of the Pynac switch... Anyway, I'm
not sure three or four people is a consensus, but I'd be all for
removing the deprecation in this particular case.
- kcrisman
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
