Project

General

Profile

Bug #171

Negate LIST, MAT,...

Added by Anna Maria Bigatti about 12 years ago. Updated about 10 years ago.

Status:
New
Priority:
High
Assignee:
-
Category:
CoCoA-4 function to be added
Target version:
Start date:
31 May 2012
Due date:
% Done:

0%

Estimated time:
5.00 h
Spent time:

Description

Negation ("-") is undefined for X, a list or matrix (and what else???)
However (-1)*X is allowed.

Should we define negation whenever "(-1)* " is allowed?
... I think so...


Related issues

Related to CoCoA-5 - Feature #183: Template for unaryMinus?New2012-06-11

Related to CoCoA-5 - Feature #184: MAT/INT ? MAT/RINGELEM ?Closed2012-06-11

History

#1 Updated by John Abbott about 12 years ago

If N*X is valid (for N integer, and X of some type) then it makes perfect sense that -X should be equivalent to (-1)*X. We should also check that X*N is allowed and is equivalent to N*X.

I wonder if the equivalence can be implemented "automatically"?
[I have my doubts]

In any case, we should make sure that the equivalence is fully tested.

#2 Updated by Anna Maria Bigatti about 12 years ago

  • Category changed from CoCoA-4 function: obsolete to CoCoA-4 function to be added

-M is not available in CoCoALib whereas
matrix operator-(ConstMatrixView Mleft, ConstMatrixView Mright)
is.
Should we add it or not?

#3 Updated by Anna Maria Bigatti about 12 years ago

John Abbott wrote:

If N*X is valid (for N integer, and X of some type) then it makes perfect sense that -X should be equivalent to (-1)*X. We should also check that X*N is allowed and is equivalent to N*X.

I wonder if the equivalence can be implemented "automatically"?
[I have my doubts]

better not. That should be implemented at cocoalib level.

In any case, we should make sure that the equivalence is fully tested.

True: ... again should be done in cocoalib

#4 Updated by John Abbott about 12 years ago

I've just added operator- (scalar multiple by -1) to MatrixArith.

Contrary to my suggestion that multiplication by an integer be commutative, I notice that M*n is not yet offered in CoCoALib. Should it be added?

#5 Updated by Anna Maria Bigatti about 12 years ago

John Abbott wrote:

I've just added operator- (scalar multiple by -1) to MatrixArith.

Contrary to my suggestion that multiplication by an integer be commutative, I notice that M*n is not yet offered in CoCoALib. Should it be added?

yes, in fact

/**/ IdentityMat(R,2)*3;
ERROR: I don't know how to evaluate operator * between MAT and INT
IdentityMat(R,2)*3;
                ^

#6 Updated by John Abbott about 12 years ago

Added scalar multiplication on the right (for matrices) to CoCoALib.

#7 Updated by John Abbott about 12 years ago

Added scalar mult on the right & unary minus for matrices in C5.

#8 Updated by John Abbott about 12 years ago

Scalar multiplication could be useful for:

[LIST, MAT, MATRIXROW, RECORD, VECTOR]

JAA thinks it is better to exclude RECORD, at least initially.
At the moment the only operation on MATRIXROW is indexing (and printing). JAA would prefer not to have scalar multiples for MATRIXROW because he thinks that a MATRIXROW is a (read-only?) reference to a row in a matrix, and it would be unclear whether a scalar multiple of a MATRIXROW should modify the original matrix.

Can we create values of type VECTOR? Are they the same as ModuleElem in CoCoALib?

#9 Updated by Anna Maria Bigatti about 12 years ago

I agree to exclude MATRIXROW and RECORD

Can we create values of type VECTOR? Are they the same as ModuleElem in CoCoALib?

not yet. I think they sould be called ModuleElem in CoCoA-5 as well

#10 Updated by John Abbott about 10 years ago

If L is a list C5 cannot compute -L but it can compute (-1)*L and also L1-L2 if L1 and L2 are lists. This seems incongruous.
I tried implementing but couldn't figure out how: there is a commented out attempt in Interpreter.C:2168--2181.

Additonally, Anna questions how sensible it is to have "list arithmetic".

Is Giovanni willing to help, if we should decide to implement?

#11 Updated by Anna Maria Bigatti about 10 years ago

  • Target version set to CoCoA-5.1.0 Easter14

#12 Updated by John Abbott about 10 years ago

  • Target version changed from CoCoA-5.1.0 Easter14 to CoCoA-5.?.?

Also available in: Atom PDF