Project

General

Profile

Bug #1100

PrimaryDecomposition0: says not 0-dim but IsZeroDim says ideal is zero-dim!

Added by John Abbott over 6 years ago. Updated over 6 years ago.

Status:
Closed
Priority:
High
Category:
enhancing/improving
Target version:
Start date:
11 Sep 2017
Due date:
% Done:

100%

Estimated time:
1.01 h
Spent time:

Description

Bug summarized in this transcript

/**/ L:=[3*x^3*y +3*y*z^3 +1,  2*x*y*z^2 +3*x,  2*y^3*z +1];
/**/ I := ideal(L);
/**/ IsZeroDim(I);
true
/**/ PrimaryDecomposition0(I);
--> ERROR: IsPrimary: implemented only for 0-dimensional ideals
--> WHERE: at line 20 (column 5) of maximal.cpkg5
-->     error("IsPrimary: implemented only for 0-dimensional ideals");
-->     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
CONTEXT: function IsPrimary at line 20 of maximal.cpkg5
 CALLED BY: function PrimaryDecomposition0 at line 38 of PrimaryDecomposition0.cpkg5
  called at top-level


Related issues

Related to CoCoALib - Bug #1101: Bug in MinPolyModular (insufficient rational reconstruction)Closed2017-09-11

History

#1 Updated by Anna Maria Bigatti over 6 years ago

  • Assignee set to Anna Maria Bigatti
  • % Done changed from 0 to 10

With SetVerbosity(90) it appears that down in the recursion it get an ideal=(1).
Investigating (should never happen!)

#2 Updated by Anna Maria Bigatti over 6 years ago

Found it: we knew it could happen, and this is the first time I see it!

The modular reconstruction gives the wrong minimal polynomial (may happen, of course).
That fact can be easily detected when one of the ideals I+ideal(factor) = 1.

In fact we never checked that the evaluation of the minpoly is in I!
I'll add this check to the code of MinPolyModular (this is in CoCoALib).

#3 Updated by Anna Maria Bigatti over 6 years ago

  • Related to Bug #1101: Bug in MinPolyModular (insufficient rational reconstruction) added

#4 Updated by Anna Maria Bigatti over 6 years ago

  • Status changed from New to Feedback
  • % Done changed from 10 to 90
  • Estimated time set to 1.01 h

Fixed: was a problem in MinPolyModular.

#5 Updated by John Abbott over 6 years ago

Anna: have you added the example(s) as a tests cases in the CoCoA test suite?

#6 Updated by John Abbott over 6 years ago

Here is a zero-dim ideal with primdecomp having 32 components:

 I := ideal(x^4 +x^2 +1,  y^4 +y,  z^4 +z);

Might be useful sooner or later?

#7 Updated by Anna Maria Bigatti over 6 years ago

  • Status changed from Feedback to Closed
  • % Done changed from 90 to 100

added tests
closing

Also available in: Atom PDF