2009/6/11 Martin Sch?necker <ms_usenet at gmx.de>:
> (%i1) depends(t, x);
> (%o1) [t(x)]
> (%i2) diff(t, z);
> diff(a, z);
> (%o2) 0
> (%o3) 0
> (%i4) diff(a(t), z);
> ev(%, diff);
> (%o4) 'diff(a(t),z,1)
> (%o5) 'diff(a(t),z,1)
> (%i6) dependencies;
> (%o6) [t(x)]
Looks like a bug to me. In the absence of any dependence of t on z,
seems like diff(a(t), z) should be 0. Or have I misremembered the
chain rule? That's entirely possible.
Robert Dodier