Skip to main content
10 events
when toggle format what by license comment
Oct 1, 2011 at 17:27 comment added fedja a) You can approximate $x$ by linear combinations of $x^2,x^3,\dots$ uniformly on arbitrarily large intervals, so, no, pointwise convergence does not imply the coefficientwise one. b) It is a nice problem, indeed, but I made it a rule not to answer to unknown(yahoo)'s. If the OP creates a unique username, I'll think of the question and vote it up.
Jul 25, 2011 at 21:24 comment added Pietro Majer This question seems to me really interesting. I don't understand why it got one vote only. Would it meet with more success if the OP re-edited it and add some more details and motivation?
Jul 21, 2011 at 3:04 comment added user16456 Coefficients are linear combinations of derivatives, derivatives are linear combinations of differences, differences are linear combinations of function values - hence point-wise convergence does imply coefficient-wise convergence.
Jul 20, 2011 at 21:13 comment added fedja Erm, the pointwise convergence doesn't imply the coefficient-wise one either.
Jul 19, 2011 at 3:27 comment added user16456 (Pietro Majer) Yeah, sorry.
Jul 18, 2011 at 12:04 comment added Pietro Majer (precisely, point-wise convergence implies coefficient-wise convergence, but not the opposite, as the example $x^n$ shows).
Jul 16, 2011 at 12:43 comment added fedja The first claim is true only if the degrees are uniformly bounded.
Jul 16, 2011 at 6:37 comment added user16456 A sequence of polynomials converges point-wise to a polynomial if and only if it converges coefficient-wise to it. A linear space of polynomials is translation invariant if and only if it is invariant under partial differentiation.
Jul 14, 2011 at 20:21 history edited user16456 CC BY-SA 3.0
added 28 characters in body
Jul 14, 2011 at 17:12 history asked user16456 CC BY-SA 3.0