`LeftIdeal` replaces the old `Ideal`; new `Ideal` is two-sided #2912

antonleykin · 2023-08-07T13:30:23Z

Ideal is a type of LeftIdeal now.

Most functions are defined for LeftIdeal;
the ones that make sense only for a two-sided ideal are defined for Ideal.
The method ideal constructs an ideal typical for the given ring: e.g.,
- LeftIdeal for a Weyl algebra and
- Ideal for all other algebras: e.g., polynomial ring, skew-commutative ring, AssociativeAlgebra, etc.

Current state:

there seem to be no errors in the build process
there still may be methods that can migrate from Ideal to LeftIdeal; currently
- (clean session of M2) #methods Ideal is 223 and # methods LeftIdeal is 66
- (after needsPackage "BernsteinSato") #methods Ideal is 271 and # methods LeftIdeal is 151

Whoever is interested in noncommutative things, please look at this.

mahrud · 2023-08-08T04:43:00Z

Since the majority of Core code related to ideals is being touched, I think this is a good opportunity to move all of them into ideals.m2, with separation of those defined for LeftIdeal and those defined for Ideal.

antonleykin · 2023-08-08T20:34:53Z

Since the majority of Core code related to ideals is being touched, I think this is a good opportunity to move all of them into ideals.m2, with separation of those defined for LeftIdeal and those defined for Ideal.

I agree, but it may be best to keep the edits in-place for now, so that people reviewing them can do so easily and are convinced that these are non-drastic (and hold water).

mahrud · 2023-09-09T18:38:25Z

Totally valid point. My only suggestion is that we can add a second commit that does the moving. This way the first commit can be reviewed first (github allows reviewing commit by commit), and the second will be only a movement so should require minimal reviewing. No strong preference.

mikestillman · 2023-09-13T14:46:26Z

@DanGrayson This pull request (currently a draft request) consists of one commit. However, it shows many commits by others that we have just recently merged into development, making it hard to see what has been added in this pull request. It also contains a conflicting file now, even though the request (I think) has nothing to do with that. Any idea how to fix this problem?

mikestillman · 2023-09-13T14:47:35Z

@moorewf and @ggsmith This pull request will be changed to non-draft soon. What are your thoughts on it?

d-torrance · 2023-09-13T16:58:05Z

It looks like commits from development were cherry-picked onto this branch (there are lots of messages like "d-torrance authored and Anton Leykin committed 2 days ago" under the "Commits" tab) rather than merging development or rebasing on top of it. Assuming that the development branch is up-to-date locally, git rebase development and then pushing this branch with the --force option should fix it.

antonleykin · 2023-09-13T18:46:40Z

It looks like commits from development were cherry-picked onto this branch (there are lots of messages like "d-torrance authored and Anton Leykin committed 2 days ago" under the "Commits" tab) rather than merging development or rebasing on top of it. Assuming that the development branch is up-to-date locally, git rebase development and then pushing this branch with the --force option should fix it.

Thanks, @d-torrance !
Is there an easy way to avoid these kind of cherrypicks?
(I believe these particular ones were obtained by a combination of pulls and pushes done by magit.)

What is the simplest workflow to avoid this?
Should we

git rebase main/development (here: main is Macaulay2 as apposed to origin which is antonleykin) and then
git push --force (forcing a push every time?)

d-torrance · 2023-09-13T18:53:57Z

Thanks, @d-torrance ! Is there an easy way to avoid these kind of cherrypicks? (I believe these particular ones were obtained by a combination of pulls and pushes done by magit.)

I think just be careful when rebasing. I'm not entirely sure what happened, but it looks like something that might have happened if you were on development and did git rebase LeftModule instead of the other way around.

What is the simplest workflow to avoid this? Should we

git rebase main/development (here: main is Macaulay2 as apposed to origin which is antonleykin) and then

git push --force (forcing a push every time?)

Yes, that should work! I generally like rebasing like this instead of merging on my pull requests so we don't end up with even more merge commits than we already have.

mahrud · 2023-11-19T06:02:31Z

Just getting around to looking into this PR. I'm wondering, wouldn't the following alternative be simpler?

Let Ideal be an abstract type, meaning not necessarily two-sided
Define methods that work for any type of ideal for Ideal
Add LeftIdeal as a subtype of Ideal (i.e. opposite of this PR)
Whenever a method should do something different for left ideals, define a method which either implements the specialized routine or returns an error.

Is there a problem with this approach?

pzinn · 2023-11-21T01:21:29Z

I know this has been discussed at length (including today!), but for what it's worth, here are my 2 cents:
why do we care about left ideals anyway? I understand why we care about two-sided ideals: that's because they're needed to make quotient rings. But left ideals are really just left modules which happen to be sitting in the ring. There's not much more about them that's interesting. So what we should be focusing on is defining left/right/bimodules...

mahrud · 2023-11-21T01:44:44Z

That's an interesting perspective. It would certainly simplify the work, in that we wouldn't need to implement left modules and also left ideals, for instance.

In this case, there are 3 options for ideal(x,dx) in a Weyl algebra:

give an error, explaining that ideals in M2 must be two sided
give the corresponding two sided ideal, in this case ideal 1
give the corresponding left module instead.

I can see pros/cons for all three.

mikestillman · 2023-11-21T08:01:32Z

@pzinn I agree, an interesting point.

@moorewf What do you think about not having left/right ideals, instead just having left/right modules?

moorewf · 2023-11-21T11:49:18Z

I'm fine with that as well. There doesn't seem to be a need to distinguish between the two.

antonleykin · 2023-11-21T18:43:24Z

A sensible solution could be to have LeftIdeal to support the math habits of the $D$-modules community (a left ideal generated by a system of differential operators is as traditional as an ideal generated by polynomials), but I can see how (for instance in general associative algebras) we can get by without introducing RightIdeal... unless someone demands it.

pzinn · 2023-11-21T21:18:15Z

But couldn't you define your D-modules as left modules rather than left ideals? after all, they are called D-modules...

mahrud · 2023-11-21T21:46:48Z

You could alternatively define LeftIdeal to inherit from left modules, no? As long as nobody needs methods defined for ideals that are not defined for modules (which I don't think is the case), I think this would work.

mahrud · 2023-11-22T23:20:26Z

I realized #3003 probably caused all the conflicts here, but since all of them are just renames from Dmodules/X to WeylAlgebras/X, a rebase should be able to fix it.

… and fixed a spurious numthreads line.

Added isPseudomonomial as exported function

When checking more than one package, this was causing build failures ("Syntax error: end of file unexpected").

It's now a wrapper around substitute(RingElement, List) to avoid issues where the matrix that was constructed ended up being over the wrong ring. We also check that the number of elements in the array matches the number of generators of the ring.

We already know that we'll have ZZ objects since this is a method, so there's no reason to check their type again. Also, neither nrows or ncols makes sense to be 0, so we raise an error in that case as well.

* Various navigation buttons presumably once used in the html documentation, plus the Makefile and Scheme files used to generate them in Gimp. * Redundant 9planets.jpg -- we use 9planets.gif in the html documentation. * Several others of unknown purpose: - doc-no-buttons.css - styleswitcher.js - recbg.jpg

…fixes

…tory."

We also switch member -> isMember

Ideal is a type of LeftIdeal. Most functions are defined for LeftIdeal; the ones that make sense only for a two-sided ideal are defined for Ideal. The method `ideal` constructs an ideal typical for the given ring: e.g., LeftIdeal for a Weyl algebra and Ideal for a commutative or skew-commutative ring. Current state: FirstPackage, Style, Macaulay2Doc install without errors, there are errors in packages related to D-modules.

…, but not for BernsteinSato yet

antonleykin · 2023-11-23T15:41:00Z

This sequence doesn't seem to help

git rebase main/development 
git push --force

Experts on rebasing, what do you suggest?

antonleykin · 2023-12-13T16:39:10Z

superseded by #3020

antonleykin force-pushed the LeftModule branch from 5d1a113 to 94b4dde Compare September 13, 2023 18:38

jkyang92 and others added 12 commits November 22, 2023 22:10

Bump the 4ti2 version for cmake builds

babebbf

Bump 4ti2 version to 1.6.10 (autotools)

60c6517

Changed how skew polynomial rings are handled in AssociativeAlgebras,…

d13a1e3

… and fixed a spurious numthreads line.

Added Strategy option for rightKernel

9c9f6b3

Update prism.js symbols for version 1.22.

546cb02

Bump VERSION to 1.22.0.1

4eeeb13

Update PseudomonomialPrimaryDecomposition.m2

6a8547a

Added isPseudomonomial as exported function

Remove quotes around shell variable in HTML5 validation test

5f1ae4a

When checking more than one package, this was causing build failures ("Syntax error: end of file unexpected").

Improve genericSkewMatrix error messages

4adba89

Improve genericSymmetricMatrix error messages

4128e1d

Improve check of number of rows/columns in genericMatrix

111c791

We already know that we'll have ZZ objects since this is a method, so there's no reason to check their type again. Also, neither nrows or ncols makes sense to be 0, so we raise an error in that case as well.

d-torrance and others added 11 commits November 22, 2023 22:10

add Number * VirtualTally, VirtualTally == VirtualTally, expressions …

d3344aa

…fixes

minor VirtualTally ? VirtualTally optimization

0a1eb65

Revert "Use clang-tidy to use nullptr instead of 0 or NULL in e direc…

2009082

…tory."

Also check for "arm" for Apple silicon dlopen hack

0f34547

We also switch member -> isMember

LeftIdeal update: more packages install now; checks pass for Dmodules…

b1780de

…, but not for BernsteinSato yet

docs in packages/Macaulay2Doc/... fixed

8d7749e

LeftIdeal --> Macaulay2Doc, HolonomicSystems; breaks at B-S

0498820

working on B.-S.: stuck on colon ideal

f14c2c6

BernsteinSato; added quotientHelper quotient(LeftIdeal,RingElement)

679d34c

antonleykin force-pushed the LeftModule branch from e6cf213 to 679d34c Compare November 23, 2023 15:23

antonleykin mentioned this pull request Nov 30, 2023

LeftIdeal (take two) #3020

Draft

mahrud mentioned this pull request Dec 7, 2023

LeftIdeal as a Module #3032

Draft

antonleykin closed this Dec 13, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

`LeftIdeal` replaces the old `Ideal`; new `Ideal` is two-sided #2912

`LeftIdeal` replaces the old `Ideal`; new `Ideal` is two-sided #2912

antonleykin commented Aug 7, 2023 •

edited

Loading

mahrud commented Aug 8, 2023

antonleykin commented Aug 8, 2023

mahrud commented Sep 9, 2023 •

edited

Loading

mikestillman commented Sep 13, 2023

mikestillman commented Sep 13, 2023

d-torrance commented Sep 13, 2023

antonleykin commented Sep 13, 2023

d-torrance commented Sep 13, 2023

mahrud commented Nov 19, 2023

pzinn commented Nov 21, 2023 •

edited

Loading

mahrud commented Nov 21, 2023

mikestillman commented Nov 21, 2023

moorewf commented Nov 21, 2023

antonleykin commented Nov 21, 2023

pzinn commented Nov 21, 2023

mahrud commented Nov 21, 2023 •

edited

Loading

mahrud commented Nov 22, 2023

antonleykin commented Nov 23, 2023

antonleykin commented Dec 13, 2023 •

edited

Loading

LeftIdeal replaces the old Ideal; new Ideal is two-sided #2912

LeftIdeal replaces the old Ideal; new Ideal is two-sided #2912

Conversation

antonleykin commented Aug 7, 2023 • edited Loading

mahrud commented Aug 8, 2023

antonleykin commented Aug 8, 2023

mahrud commented Sep 9, 2023 • edited Loading

mikestillman commented Sep 13, 2023

mikestillman commented Sep 13, 2023

d-torrance commented Sep 13, 2023

antonleykin commented Sep 13, 2023

d-torrance commented Sep 13, 2023

mahrud commented Nov 19, 2023

pzinn commented Nov 21, 2023 • edited Loading

mahrud commented Nov 21, 2023

mikestillman commented Nov 21, 2023

moorewf commented Nov 21, 2023

antonleykin commented Nov 21, 2023

pzinn commented Nov 21, 2023

mahrud commented Nov 21, 2023 • edited Loading

mahrud commented Nov 22, 2023

antonleykin commented Nov 23, 2023

antonleykin commented Dec 13, 2023 • edited Loading

`LeftIdeal` replaces the old `Ideal`; new `Ideal` is two-sided #2912

`LeftIdeal` replaces the old `Ideal`; new `Ideal` is two-sided #2912

antonleykin commented Aug 7, 2023 •

edited

Loading

mahrud commented Sep 9, 2023 •

edited

Loading

pzinn commented Nov 21, 2023 •

edited

Loading

mahrud commented Nov 21, 2023 •

edited

Loading

antonleykin commented Dec 13, 2023 •

edited

Loading