Recent changes to 561: matrix(...taylor...)^^-1 wronghttps://sourceforge.net/p/maxima/bugs/561/2004-04-21T02:37:57Zmatrix(...taylor...)^^-1 wrong2004-04-21T02:37:57Z2004-04-21T02:37:57ZStavros Macrakishttps://sourceforge.net/u/macrakis/https://sourceforge.net10cf800cac6e461a27d232c2b43a80d47cfdb6c9Consider
m: matrix\(\[taylor\(1+a\*x,x,0,1\),0\],
\[0,taylor\(1+d\*x,x,0,1\)\]\);
Now,
m^^-1
=> matrix\(\[1/\(d\*x+1\),0\],\[0,1/\(d\*x+1\)\]\)
There are two problems with this. First, the answer is
incorrect. Compare:
matrixmap\(ratdisrep,m\)^^-1
=> matrix\(\[1/\(a\*x+1\),0\],\[0,1/\(d\*x+1\)\]\)
Second, the answer is not in terms of taylor series. A
silent ratdisrep was done in the middle, losing truncation
information.