redTail option ignored?

[kroeker@…]

Simple failing example:

ring rng = 0,(x,y),dp;
short = 0 ;
option() ;
ideal J = 3*x-2*y, 4*x-y;
ideal gJ =  std(J);
gJ;

output

> ring rng = 0,(x,y),dp;
> short = 0 ;
> option() ;
//options: redTail redThrough intStrategy redefine loadLib usage prompt
> ideal J = 3*x-2*y, 4*x-y;
> ideal gJ =  std(J);
> gJ;
gJ[1]=y
gJ[2]=4*x-y  // y is not reduced away!

the example seems ok with enabled "redSB":

> ring rng = 0,(x,y),dp;
> short = 0 ;
> option(redSB) ;
> option(r) ;
//options: redSB redTail redThrough intStrategy redefine loadLib usage prompt
> ideal J = 3*x-2*y, 4*x-y;
> ideal gJ =  std(J);
> gJ;
gJ[1]=y
gJ[2]=x
>   std(J);
_[1]=y
_[2]=x
> option(noredTail);
> option();
//options: redSB redThrough intStrategy redefine loadLib usage prompt
>   std(J);
_[1]=y
_[2]=x

[hannes]

The option says (from doc): reduction of the tails of polynomials during standard basis computations

i.e. IF there is a reduction concerning this polynomial, the tail will be reduced. It says nothing about other polynomials (as in this case). The reason for this option is to move a part of the reduction into the middle the GB computation as oppose to redSB which does that at the end.

[kroeker@…]

The option says (from doc): reduction of the tails of polynomials during standard basis computations i.e. IF there is a reduction concerning this polynomial, the tail will be reduced.

Thanks for the clarification!

Am I alone with the misinterpretation of doc? If not, maybe it should be considered to update the documentation, despite of the fact that the documentation is correct as is.

[gorzel]

I have to second, the meaning of redTail needs some more explanation. It was not clear to what Hannes describes.

"IF there is a reduction concerning this polynomial, the tail will be reduced."

is only about the newly chosen polynomial !?

Is there some experience about the effect of the settings for

"redTail" combined with / or opposite to "length"

on the speed of the std comutations in the global inhomoigenous case?

[ren]

Then change

reduction of the tails of polynomials during standard basis computations

to

reductions during standard basis computations also reduce tails of polynomials

or something similar.

If Hans is okay with that, he can merge it when he is back and this issue is solved.