pmat abbreviates matrix entriew without notice

[Simon King]

The following happens both with the CVS-version (February 16, 2009) and with Singular-Versions before 3-0-4:

>    ring r=0,(x,y,z),ls;
>    ideal i= x,z+3y,x+y,z;
>    matrix m[3][3]=i^2;
>    pmat(m);
x2,     xz+3xy,     xy+x2,
xz,     z2+6yz+9y2, yz+3y2+xz+3xy,
z2+3yz, y2+2xy+x2,  yz+xz
>    pmat(m,5);
x2    xz+3x xy+x2
xz    z2+6y yz+3y
z2+3y y2+2x yz+xz

So, in pmat(m,5), several matrix entries are abbreviated, but it is impossible to guess whether yz+3y is a valid polynomial or just an abbreviation of yz+3y2+xz+3xy.

Moreover, I find it not nice that pmat(m) uses ", " as column separator, whereas pmat(m,5) just uses " " (single blank space, no comma) as separator.

Therefore, after my suggestion, pmat was changed in Singular-3-0-4, so that the output was:

   pmat(m);
x2,     xz+3xy,     xy+x2,
xz,     z2+6yz+9y2, yz+3y2+xz+3xy,
z2+3yz, y2+2xy+x2,  yz+xz
   pmat(m,5);
x2,    xz+.., xy+x2,
xz,    z2+.., yz+..,
z2+.., y2+.., yz+xz

Hence, in Singular-3-0-4, the pmat-output clearly shows whether a matrix entry was abbreviated or not.

Is there a good reason why Singular-3-1-0 should return to the old confusing output of pmat? Otherwise I suggest to keep pmat as it is in Singular-3-0-4.

[hannes]

CVS: back to 1.30, with id->@@id change from cvs log

[greuel@…]

The changes of Simon in 3-0-4 should be included in 3-1-0. Gert-Martin