Center of skew rings and kernel of RingMap to skew rings by joel-dodge · Pull Request #4485 · Macaulay2/M2

joel-dodge · 2026-07-26T18:18:17Z

Add centerRing method to compute the center of a ring. Currently supports skew polynomial rings (and commutative rings 😂 ).

Then adds a new strategy for computing the kernel of a RingMap whose target is SkewCommutative by reducing to a map to the center of the target.

This requires some awkward packaging since kernel is in the core and the strategy for computing the center of a skew commutative ring depends on the PushForward package. We rely here on the "hack" of calling needsPackage inside the hook only when needed although @d-torrance says "I'm not sure if this is a very elegant solution" and I am inclined to agree.

Note while editing ringmap.m2 I took the opportunity to clean up ws formatting in the block declaring different kernel strategies for RingMap. Consider toggling the "hide whitespace" option while reviewing this PR.

…te kernels

bugs/dan/1-jacobian through 1-loadPackage. Six fixed, two duplicates, two wontfix, two filed. Filed: * Macaulay2#4589 1-join-hash -- the guess in the file is right and the saving is real. Every List construction folds the whole list (basic.d:114-117 does r.hash = hash(r); the fold is at :82-85), so join of two 200k lists costs 11.2ms of which 7.5-8.5ms is hashing -- measurable because a Sequence's hash is not cached and refolds on demand. The fold is combinable: h(L++M) = 1299833^#M * (hash L - c) + hash M mod 2^64, c = hash class L + 23407, verified on seven cases. My first attempt used the Sequence constants and failed all five, which is how the two folds turned out to differ. * Macaulay2#4590 1-kernel-ringmap -- Mike's 2008 recipe, unevenly landed. The Hilbert hint is implemented for degree-preserving maps (ringmap.m2:320-332, verified live: gb records rawGBSetHilbertFunction log), the inhomogeneous case it was written for is skipped because canUseHilbertHint requires isHomogeneous (gb.m2:271), and the recipe's forceGB step is Macaulay2#686. PR Macaulay2#4485 is adding another strategy to the same table. Duplicates, both with comments posted: 1-kernel-ringmap-ZZ -> Macaulay2#1518 (the message moved to methods.m2:695, and ringmap.m2:336's ZZ strategy only fires when the source is ZZ itself), 1-line-wrapping-with-negative-numbers -> Macaulay2#4149 (the same splitcolumn gap one character class over; three lines still end in a bare hyphen). Fixed: 1-leadTerm-crash.m2 (the ignored engine error is propagated, so enginering.m2:496 errors instead of smoking), 1-lift-RR-CC (both spellings return precision 53), 1-lift-modules (lift((QQ[x])^3, QQ) is QQ^3), 1-lifting (.3 * id_(QQ^3) is over RR_53 now, and promote(.3, QQ) errors -- the direction PR Macaulay2#62 on MichaelABurr/M2 is formalising), 1-lcm-needed (implemented, gcd compiled at 14.3us on 700-digit inputs, and the Binary ask declined on measurement: 9.7ms through Binary against 13.8ms for a hand fold), 1-loadPackage (PR Macaulay2#3852's warning meets the need; the file's own example is invalid today for a rank reason and fails identically with no reload). Wontfix: 1-jacobian (the intrinsic object is one line away, and jacobian-doc documents the present behaviour as "identical to jacobian generators I"), 1-list-of-ideals-of-R (asked and declined in PR Macaulay2#3328). dan/1 todo 154 -> 142, settled 351 -> 363.

joel-dodge · 2026-08-15T01:35:17Z

@MichaelABurr @ggsmith any progress reviewing this?

MichaelABurr · 2026-09-13T21:19:12Z

+    Z.formation = FunctionApplication {centerRing, R};
+
+    map(R, Z, phi^-1 matrix {ximages | yimages})
+))


Please add a newline at the end of the file.

MichaelABurr

This looks good to me for the most part. There are a few stylistic changes that should be made. It includes some tests and has documentation.

MichaelABurr · 2026-09-13T21:19:48Z

 	"capture",
 	"catch",
 	"ceiling",
+        "centerRing",


Please fix spacing

changing to 4 spaces instead of 8. the formatting of this file looks crazy to me in vscode so it was hard to tell what the right indentation was.

MichaelABurr · 2026-09-13T21:22:24Z

+        (F',r) := flattenRing(F, CoefficientRing => k);
+        if R' === R and F' === F then return null;
+        p^-1 kernel (r * f * p^-1)
+    )


Is the only difference here the removal of the final comma?

That's right. Outside of the addition of the new hook method, the only non-ws change to this file is the removal of this final comma in the definition.

MichaelABurr · 2026-09-14T02:12:53Z

-			)
-		)
-	),
+        (F, R) := (target f, source f);


It looks like there wasn't a change here - perhaps only a change to spacing.

Indeed, as i said in the pr description:

Note while editing ringmap.m2 I took the opportunity to clean up ws formatting in the block declaring different kernel strategies for RingMap. Consider toggling the "hide whitespace" option while reviewing this PR.

MichaelABurr · 2026-09-14T02:17:03Z

-		assert (not chh or G#?"rawGBSetHilbertFunction log"); -- ensure the Hilbert function hint was actually used in gb.m2
-		ideal mapback selectInSubring(1,generators G)
-	),
+        (F, R) := (target f, source f);


I'm also not seeing a change in this section either - again, it might be due to a spacing change.

MichaelABurr · 2026-09-14T02:17:34Z

-		if R' === R and F' === F then return null;
-		p^-1 kernel (r * f * p^-1)
-	),
+        (F, R) := (target f, source f);


I think that the only change here was the removal of a comma.

MichaelABurr · 2026-09-14T02:18:10Z

+assert(map(kk, kk) === centerRing kk)
+
+R = kk[a..c] / ideal {a^2 + b, b^2 + c^3, c^2 - a*b^2}
+assert(map(R, R) === centerRing R)


Missing newline at end of file

MichaelABurr · 2026-09-14T02:18:19Z

+
+-- inhomogeneous
+g = map(R, S, {a*b*c*d + 2*a*b, c*d + 1})
+assert(ker g == ideal(t^2-2*t+1,s^2))


Missing newline at end of file

MichaelABurr · 2026-09-14T02:19:20Z

+SeeAlso
+    isCommutative
+    isSkewAffineRing
+///


Missing newline at end of file

i missed that it is standard to have a newline at the end of files. Is this part of a style guide somewhere?

joel-dodge · 2026-09-14T02:52:06Z

@MichaelABurr ty so much for review and I am sorry for the confusion about the whitespace changes! I think I have responded to all of the stylistic changes you requested. I did ask one question about whether is a style guide of some sort that I have not seen. ty again!

MichaelABurr · 2026-09-14T09:59:05Z

@MichaelABurr ty so much for review and I am sorry for the confusion about the whitespace changes! I think I have responded to all of the stylistic changes you requested. I did ask one question about whether is a style guide of some sort that I have not seen. ty again!

I didn't know that you could view the changes without the whitespace, I really appreciate knowing about that option!

The newline at the end of a file isn't strictly necessary/part of our style guide, but it is part of the POSIX style. My understanding is that some software expects newlines at the ends of files (and, without it, unexpected behaviors can occur).

joel-dodge · 2026-09-15T17:48:11Z

@MichaelABurr one thing that I fussed with a bit in putting this together was whether to call the method centerRing or to just use the more mathematically usual notation of center. What I ran into when using plain center was that there are several packages that implement their own notion of center currently, each of them defining

center = method()

in their own package. I guess that if multiple of these were used at once there would be problems wiht redefinition of the global symbol center. Trying to unify all of these methods and keeping the functionality separate via the usual method of different overrides (because indeed they all operate on different types) would be technically functional but would run into the problem of all center methods having awareness of any of the options that any of them accept. For example, my new centerRing accepts a Strategy but I think none of the other ones currently do. This sort of leakiness in the abstraction was enough to make me want to just rename the method to centerRing.

Do you have any thoughts about these problems and concerns?

d-torrance · 2026-09-17T23:08:39Z

I think center is absolutely a great candidate for moving to Core

joel-dodge · 2026-09-17T23:57:06Z

@d-torrance awesome. I am very happy to rebase this on top of your change when it is merged and rename centerRing to center.

joel-dodge · 2026-10-22T04:49:48Z

Any idea what is going on here with this failure @d-torrance?

In file included from /Users/runner/work/M2/M2/M2/Macaulay2/e/mathic/mathic.h:13:
/Users/runner/work/M2/M2/M2/Macaulay2/e/mathic/mathic/DivList.h:595:36: fatal error: no member named 'back_inserter' in namespace 'std'
  595 |     std::copy(begin(), end(), std::back_inserter<TmpContainer>(tmpCopy));
      |                                    ^~~~~~~~~~~~~
1 error generated.

I see something in my local diff against development about the mathic submodule having an updated commit hash and I never really know how to think about submodules so don't know if that could be related.

d-torrance · 2026-10-22T11:16:19Z

Yeah, it looks like you accidentally committed a change to the mathic submodule, rewinding it to an earlier commit that's missing the fix from Macaulay2/mathic#28

joel-dodge · 2026-10-22T15:04:05Z

ugh got it ty so much for your help!

joel-dodge added 2 commits July 26, 2026 13:44

add method for computing the center of skew rings and use it to compu…

0d4aeee

…te kernels

some tests

7f7aa48

joel-dodge force-pushed the skew-kernels branch from 949bfc8 to 7f7aa48 Compare July 26, 2026 20:45

joel-dodge added 2 commits July 26, 2026 19:18

add center doc as subnode

6c9818c

just use centerRing

a6a82f8

joel-dodge force-pushed the skew-kernels branch from 486645d to a6a82f8 Compare July 27, 2026 14:41

joel-dodge added 2 commits July 30, 2026 11:57

move to core with lazy loading

b104b84

add comment

7724ead

joel-dodge marked this pull request as ready for review July 31, 2026 17:53

d-torrance added the Core Issues involving the Core scripts. label Aug 2, 2026

d-torrance requested review from MichaelABurr and ggsmith August 2, 2026 17:52

joel-dodge mentioned this pull request Aug 4, 2026

Compute Hom / Ext over exterior algebra via pushforward #4477

Open

d-torrance mentioned this pull request Aug 7, 2026

kernel(RingMap) cannot use a Hilbert function hint unless the map preserves degrees #4590

Open

MichaelABurr reviewed Sep 13, 2026

View reviewed changes

add final newline

96a43a8

MichaelABurr approved these changes Sep 14, 2026

View reviewed changes

formatting fixes

47b4aa0

d-torrance mentioned this pull request Sep 17, 2026

Add "center" method to Core #4724

Merged

joel-dodge added 4 commits October 21, 2026 10:04

Merge branch 'development' into skew-kernels

d003ada

define center method outside of shared.m2

d62b77e

fixup center doc nodes

d1327b5

fixup def of method options

3f285ff

joel-dodge added 2 commits October 21, 2026 13:50

fussing with doc nodes more

0b11e7f

fix center method sigs to include options

805b2e4

fix mathic submodule commit hash

c1d53d9

ggsmith approved these changes Oct 22, 2026

View reviewed changes

d-torrance merged commit 80e3b9b into Macaulay2:development Oct 22, 2026
5 checks passed

Conversation

joel-dodge commented Jul 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

joel-dodge commented Aug 15, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

MichaelABurr left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

joel-dodge commented Sep 14, 2026

Uh oh!

MichaelABurr commented Sep 14, 2026

Uh oh!

joel-dodge commented Sep 15, 2026

Uh oh!

d-torrance commented Sep 17, 2026

Uh oh!

joel-dodge commented Sep 17, 2026

Uh oh!

joel-dodge commented Oct 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

d-torrance commented Oct 22, 2026

Uh oh!

joel-dodge commented Oct 22, 2026

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

joel-dodge commented Jul 26, 2026 •

edited

Loading

joel-dodge commented Oct 22, 2026 •

edited

Loading