On 12/30/19 4:23 AM, Alex wrote:
> On page 12, for 1d unit interval problem why dim(A)=1/h? rather than 1/h+1?
Fair question -- I should have said
dim(A) = (1/h)+1 \approx 1/h
In most of numerical analysis, we really are only interested in asymptotic
statements "if h is very small" or "if N is large". In those cases,
1/h
and
1/h + 1
are the same.
Best
W.
--
------------------------------------------------------------------------
Wolfgang Bangerth email:
bang...@colostate.edu
www:
http://www.math.colostate.edu/~bangerth/