@@ -169,7 +169,7 @@ by checking the version:
169
169
$ magma
170
170
> AttachSpec("IHecke.spec");
171
171
> IHeckeVersion();
172
- IHecke version 2021-09-10
172
+ IHecke version 2021-11-01
173
173
174
174
Now, create a ` GrpFPCox ` of your favourite type:
175
175
@@ -353,6 +353,17 @@ In order to extract a particular coefficient, use the `Coefficient` function.
353
353
> Coefficient(H.[2,1] * H.1, W.1);
354
354
0
355
355
356
+ The ` Coefficient ` function can be given an extra argument to extract the v^d coefficient from the Laurent polynomial.
357
+
358
+ > Coefficient(H.[2,1] * H.1, W![2,1]);
359
+ v^-1 - v
360
+ > Coefficient(H.[2,1] * H.1, W![2,1], 1);
361
+ -1
362
+ > Coefficient(H.[2,1] * H.1, W![2,1], 0);
363
+ 0
364
+ > Coefficient(H.[2,1] * H.1, W![2,1], -1);
365
+ 1
366
+
356
367
The Kazhdan-Lusztig bar involution can be calculated using ` Bar ` .
357
368
358
369
> Bar(H.1);
@@ -762,6 +773,8 @@ We aim to keep these to a minimum once the package is in use.
762
773
- Development version
763
774
- Added a faster Standard x Canonical -> Canonical multiplication.
764
775
- The cell order relations are precomputed, making cell order testing much faster.
776
+ - Fixed a crash when creating a Hecke algebra for a group not of affine or finite type.
777
+ - Added a third argument to ` Coefficient() ` for extracting the v^d term.
765
778
- Version 2021-11-01
766
779
- Added an experimental "literal" basis type (a basis specified by a partial table). I will wait
767
780
to see how it plays out in other projects before making it a feature.
@@ -789,6 +802,9 @@ We aim to keep these to a minimum once the package is in use.
789
802
790
803
# TODO
791
804
805
+ - (High priority) Enhance the Literal basis
806
+ - Allow it to be a partial basis (eg for affine groups).
807
+ - Make some standard way to save and restore it, probably through MAGMA object files.
792
808
- (High priority) Allow the multiplication ` C(w) * C.s ` to be driven by a table of mu-coefficients
793
809
(in other words, a W-graph) so that this special case is extremely fast. By induction (and the
794
810
standard multiplication formula), this means that multiplication within the canonical basis can
0 commit comments