Talk:Group (mathematics)/Archive 7

Archive 1

←

Archive 5

Archive 6

Archive 7

Archive 8

Group homomorphisms - Cayley's theorem

The last paragraph of the Group homomorphisms section describes Cayley's theorem applied to finite groups:

One useful theorem about finite groups is that every finite group admits an injective group homomorphism into the symmetric group of a finite set. This can be constructed by taking the finite set as the underlying set of the group {\displaystyle G} G and sending a {\displaystyle g\in G} g\in G to the set-automorphism given by {\displaystyle g\cdot } {\displaystyle g\cdot }.

This paragraph seems out of place given the level of the rest of the article. If we want to keep it, we should find a way to re-write it to be as readable to the lay-person as the surround text. I also think that (if we keep this), we should add an explicit mention and link to Cayley's theorem and remove the restriction to finite groups. Having said that, ignoring the readability issues, Cayley's theorem seems like it is still a bit to advanced to warrent inclusion in an otherwise 2 paragraph long introduction to homomorphisms. Homura1650 (talk) 07:23, 5 December 2017 (UTC)

I agree with you. The edit was certainly done in a constructive mood, but we do mention Cayleys theorem in the § about finite groups (where it belongs). I don't think a more detailed explanation why Cayleys theorem holds is in order there.

I suggest reverting the edit you quote. Jakob.scholbach (talk) 10:02, 5 December 2017 (UTC)

External links modified (January 2018)

Hello fellow Wikipedians,

I have just modified one external link on Group (mathematics). Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:

• Added archive https://web.archive.org/web/20140222213905/http://gdz.sub.uni-goettingen.de/index.php?id=11&PPN=PPN235181684_0020&DMDID=DMDLOG_0007&L=1 to http://gdz.sub.uni-goettingen.de/index.php?id=11&PPN=PPN235181684_0020&DMDID=DMDLOG_0007&L=1

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

This message was posted before February 2018. After February 2018, "External links modified" talk page sections are no longer generated or monitored by InternetArchiveBot. No special action is required regarding these talk page notices, other than regular verification using the archive tool instructions below. Editors have permission to delete these "External links modified" talk page sections if they want to de-clutter talk pages, but see the RfC before doing mass systematic removals. This message is updated dynamically through the template {{source check}} (last update: 5 June 2024).

• If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
• If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—InternetArchiveBot (Report bug) 19:10, 27 January 2018 (UTC)

Finite fields

I have twice reverted the inclusion of Finite field as the main article for the section on modular arithmetic. It is true that some finite fields are of this type, but not all! When the modulus is a proper power of a prime, the field of that order and modular arithmetic are not related. Furthermore, the modulus in modular arithmetic is not limited to primes or powers of primes, so again there is no general correspondence with finite fields. Indicating that finite fields are the main sources of modular arithmetic is just plain wrong! --Bill Cherowitzo (talk) 18:50, 26 June 2018 (UTC)

1) text on this part of page is too short;

2) text on this part of page is similar to part "Finite Field" of page "Field", the picture is same;

3) can You (Bill Cherowitzo) redirect this part of Page "Group" to page "Finite Group" ?

With the hope for a constructive solution,

   Grigory Bilchenko

Ggbil2 (talk) 19:42, 26 June 2018 (UTC)

P.S. The first part of the page "Finite Ring" is named ... "Finite Field" ;)

Ggbil2 (talk) 20:00, 26 June 2018 (UTC)

I agree with Bill Cherowitzo. Also, according to Template:Main, because subsections are often written in summary style, this template is used after the heading of the summary, to link to the subtopic article that has been summarized. In this case the template is not appropriate, because the subsection is specifically about modular arithmetic—not finite fields. There is currently no part of the Group (mathematics) page that should be redirected to the Finite field page. However, the Modular arithmetic subsection already links to the Modular arithmetic article.—Anita5192 (talk) 21:04, 26 June 2018 (UTC)

Definition (again)

the requirements ea = a = ae and a ^-1 a = e = a a ^-1 are redundant. sufficient is: ea=a and a ^-1 a = e. This is posed as an exercise in Group Theory .. by Norton Hamermesh. With the reduced definition one can prove that none of the axioms is redundant (id est, none of the axioms can be proved by the other axioms) Should this be included in the article? Jacob.Koot (talk) 16:07, 28 September 2018 (UTC)

This is already mentioned in Elementary consequences of the group axioms.—Anita5192 (talk) 16:51, 28 September 2018 (UTC)

My mistake, sorry. Nevertheless, I don't see mention of the fact that it can be proven that none of the axioms can be omitted. Also that the identity e in ea=a must be the same as the identity e in a ^-1 a = e, which is essential too. Jacob.Koot (talk) 17:17, 28 September 2018 (UTC)

If any of these elementary principles are not in the article, you may want to include them.—Anita5192 (talk) 17:21, 28 September 2018 (UTC)

How about:

It can be proven that none of the axioms can be omitted and that it is essential that the identity e in axiom

e • a = a

is the same one as identity e in the axiom

a^-1 • a = e

Jacob.Koot (talk) 18:08, 28 September 2018 (UTC)

First of all, this is obvious, since the same variable, e, is used in both equations. Second, I avoid using the word one in mathematical discussions, as it can be ambiguous. I would have said, "is the same identity as e in the axiom . . ." Third, I think it is more important to state in general that the identity e is unique, which is easy to prove: assume the existence of two identities, e₁ and e₂, and then prove e₁ = e₂.—Anita5192 (talk) 19:02, 28 September 2018 (UTC)

My point is that it is essential that in both axioms e is the same. It is possible to construct a system that satisfies all axioms except that the (or a) identity is not the same in both cases. For example:

e	e'
e'	e'

It satisfies all axioms, except that it has two distinct identities and is not showing a group. I think mentioning that the two identies must be one and the same is essential. This is clear from the definition, but I think it should be mentioned two distinct identities not necessarly produce a group. Consider it as part of showing that the axioms cannot be reduced to less restrictive axioms. Jacob.Koot (talk) 19:19, 28 September 2018 (UTC)

Please look at your example again. Your element e' is not an identity unless e' = e (since e' ⋅ e = e'). If all the group axioms are satisfied, you can not have two distinct identities, as Anita5192 has pointed out. In fact, this can be shown in much weaker systems. A groupoid having a left identity and a right identity has a unique (two sided) identity. Most authors would not consider it necessary to stipulate possibly different identity elements in the statement of the axioms, since it is completely trivial that they would have to be the same and this statement would only lead to unnecessary confusion. --Bill Cherowitzo (talk) 20:29, 28 September 2018 (UTC)

Look at it as follows:
1: forall x,y in G : xy in G (composition)
2: the composition is associative
3: exists e in G : forall x in G : ex = x (left identity)
4: exists e' in G : forall x in G : exists y in G : yx = e'
5: e' = e

Axiom 5 cannot be omitted.
There are systems (not groups) that satisfy 1, 2, 3 and 4 but do not satisfy 5. See my example above.
Hence axiom 5 is necessary.
Notice that I left out the right identity.
Its existence can be proven as well that it necessarily is the same as the left identity. Jacob.Koot (talk) 11:00, 29 September 2018 (UTC)

What you just defined is not a group.

However, looking at the definition in the article, I noticed several deficiencies: 1. the operation should be referred to specifically as a binary operation. 2. the result, ab, should be specified as unique, that is, the binary operation is well-defined as well as simply defined. 3. the line under Identity element indicating that "Such an element is unique . . ." should be removed, as this is a result of the definition—not part of the definition. If nobody has any objections, I will make these changes.—Anita5192 (talk) 16:30, 29 September 2018 (UTC)

More simply stated, having 4 in this form is convoluted. This axiom defines the inverse, not a new unity and an inverse. It should simply say:

4: forall x in G : exists y in G : yx = e

where the e is implicitly the e of the previous line and 5 becomes meaningless. −Woodstone (talk) 18:05, 29 September 2018 (UTC)

I agree. Your #4 does not define a right inverse for each element unless e' = e. A semigroup with a left identity is a group if and only if each element has at least one right inverse (Bruck, 1971). Your approach is being overly convoluted and I am not seeing any advantage to it. --Bill Cherowitzo (talk) 19:08, 29 September 2018 (UTC)

The point I wanted to make is that in the two axioms
forall x in G: ex = x
forall x in G: exists y in G: yx = e
it is essential that both axioms have the same e.
Just to show that the axioms cannot be simplified. It is not very difficult to prove that no part of the axioms can be omitted, the rule that e must be the same in the two axioms above included. That is my point. Jacob.Koot (talk) 15:21, 30 September 2018 (UTC)

What you are saying is clear to me, but I do not believe that it is making the point you think it is making. The problem is that your axiom 4 is not a group axiom as you have stated it. Let me try to expand on this. Call a system that satisfies your axioms 1-4 a Beastie. You have a theorem that says a Beastie is a group if and only if the statement of axiom 5 holds. Every group is a Beastie, but there are (proper) Beasties that are not groups (as your example shows). A Beastie is a generalization of a group. Now we have to leave the realm of mathematics and ask, is the concept of a Beastie useful? Is it interesting in some way? There are no mathematical answers to those questions as they are based on subjective judgments. I would say that they are not since proper Beasties are not quasigroups (do not have multiplication tables that are Latin squares) and can not be embedded in any group (although they could contain a group as a substructure). Others may view this differently and find them interesting, but it is clear that they are not groups nor even structures that are in some sense precursors of groups. Your claim that this theorem about Beasties shows that axiom 5 is essential as a group axiom does not follow. Take any theorem whose conclusion was that something is a group and any hypothesis of that theorem. By weakening the hypothesis you can lose the conclusion and in this situation, anything that re-strengthens the hypothesis, you would call an essential axiom in the definition of a group. This is the argument you are making in this specific case and when looked at from a general viewpoint you can see that it is not convincing.--Bill Cherowitzo (talk) 19:15, 30 September 2018 (UTC)

It convinces me. BTW, I am not suggesting to add axiom 5 as a separate axiom.

I only like to emphasize that in the axioms ex=x and exists y: yx=e, the same element e is essential.

Consider the set {0, 1, 2, 3, 4} with associative (even abelean) composition f(x,y) = min(x+y,4).

We have f(0,x)=x and f(4,x)=4, but do not have a group here. It is a Beastie, as you call it. Jacob.Koot (talk) 11:49, 1 October 2018 (UTC)

Switch to LaTeX?

I've been switching Wikipedia math pages to LaTeX to aid presentability, but I've been advised that I should write on the talk pages before doing so.

I think this article should render math equations via LaTeX. It currently is mostly not using LaTeX, which is surprising because this is an extremely important article. I can do this in a pretty systematic way using Visual Code Studio.

Proposed version on the left; ugly current version on the right.

My reasons:

1) You don't have to do anything and only the math symbols will be changed.

2) TeX can be rendered on any modern browser and mobile device. I checked so myself.

3) TeX is easier to type out. Compare ''F'' : ''C''<sup>2</sup> → ''D''<sup>2</sup> which produces F : C² → D² with :<math>F:C^2 \to D^2</math> which produces $F:C^{2}\to D^{2}$.

4) The math rendering using apostrophes, for example ''F'' : ''C''<sup>2</sup> → ''D''<sup>2</sup> which creates F : C² → D², is extremely limited. For example, how are you going to create an integral? A fraction? Are you going to use some kind of ridiculous unicode symbol?

5) TeX looks nicer, and presentability is extremely important for math pedagogy. Modern mathematicians are starting to wake up and realize that we need graphs, figures, diagrams, colors, and better formatting to explain concepts and to overall write better books. The ultimate goal is to help people learn these concepts, and right now this article suffers from styling issues which I believe can make it harder to understand.

6) Overall, I think anyone will be sort of turned off when visiting this article and discovering that the math has been formatted in a horrible, ugly and unreadable manner, and will then go somewhere else to read about groups.

An example of my proposed changes can be compared side by side; see the thumb pic on the right. If nobody has any objections/no one responds, then I will go ahead and implement the changes since it's nothing risky anyways (as far I have researched anyways; correct me if I'm wrong). — Preceding unsigned comment added by Lltrujello (talk • contribs) 23:38, 17 March 2020 (UTC)

Generally no. I do usually support change to {{math}}/{{mvar}} for simple, inline stuff, and <math>...</math> is generally okay for more complicated inline stuff and anything that's set apart on its own line, but changing every little bit of math on a page to <math>...</math> mode isn't ideal. It tends to hinder accessibility; you can't copy it to the clipboard; etc. –Deacon Vorbis (carbon • videos) 23:42, 17 March 2020 (UTC)

Could you provide different reasons? I'm not sure what you mean by accessibility. As I pointed out, TeX can be rendered on modern web browsers and mobile devices. I also think it'd be more readable for someone with an eye disability; for example, as it is now. It is difficult to distinguish math characters with regular words because they're so closely similar.

Also, I'm not sure what you mean about copying and pasting. I'm a workaholic math student, so I read math notes, books, and surf wikipedia math articles and stack exchange sites every day. But I cannot remember a single time I've ever wanted to copy and paste a math expression. The only possible reason I'd want to do that would be to google something, but it's generally not fruitful to google specific unicode characters. Lltrujello (talk) 02:31, 18 March 2020 (UTC)

• Full mathematics formatting is needed for formulas in the "Rationals", "Galois groups", "Finite groups", and "Topological groups" subsections. My position is that we should not mix and match formatting: a single style should be used consistently within any single mathematics article, because otherwise we get the same thing formatted in different ways, unnecessarily confusing the readers. So because we need LaTeX somewhere in the article, I am in favor of using it everywhere in the article. —David Eppstein (talk) 23:48, 17 March 2020 (UTC)

Oppose The use of <math> . . . </math> tags is useful for complicated mathematics typesetting, for example, integrals and fractions, but I disagree with (3) above. For simple inline variables like a, b, etc., apostrophes are much easier and less error-prone. And retroactively changing everything is a lot of work to do, a lot of edits for other editors to check, and a lot of opportunities for mistakes. We have more important editing to do. — Anita5192 (talk) 00:00, 18 March 2020 (UTC)

• Let me give you a better example. Observe that "If {{nowrap|1=''f'' : ''a'' → ''b''}}, {{nowrap|1=''g'' : ''b'' → ''c''}} and {{nowrap|1=''h'' : ''c'' → ''d''}} then {{nowrap|1=''h'' ∘ (''g'' ∘ ''f'') = (''h'' ∘ ''g'') ∘ ''f''}}" (From the Category Theory page) which creates

if f : a → b, g : b → c and h : c → d then h ∘ (g ∘ f) = (h ∘ g) ∘ f

can be compared with "If f: a \to b, g: b \to c and h: c \to d then h \circ (g \circ f) = (h \circ g)\circ f" which produces

if $f:a\to b,g:b\to c$ and $h:c\to d$ then $h\circ (g\circ f)=(h\circ g)\circ f$

Clearly, the previous expression with the ridiculous apostrophes is extremely time consuming to write and error prone. I don't see how anyone could possibly believe the opposite. Furthermore, the previous equation requires 178 characters with 192 bytes while the LaTeX-ed equation requires only 83 characters and 83 bytes. I've come across many expressions like these, and have felt sorry for the previous authors who probably spent a few minutes on an equation which could have been written in a few seconds with LaTeX.

Ultimately, the goal here is to make mathematics readable. I'm trying to make math better for other people just like everyone else on these sites. But times are changing, LaTeX is becoming more widespread, and the novice math people, which this article targets, are not going to like the hard-to-read, apostrophe-written equations that are here, and will prefer LaTeX. Lltrujello (talk) 02:31, 18 March 2020 (UTC)

"If {{math|1=''f'':''a'' → ''b''}}, {{math|1=''g'':''b'' → ''c''}} and {{math|1=''h'':''c'' → ''d''}} then {{math|1=''h'' ∘ (''g'' ∘ ''f'') = (''h'' ∘ ''g'') ∘ ''f''}}" creates

"If f:a → b, g:b → c and h:c → d then h ∘ (g ∘ f) = (h ∘ g) ∘ f

Using {{math}} produces better-styled formula text than not using it, as you can see. Using <math>...</math> for complex formulas is fine. Using {{math}} for simple formulas is also fine. There's a grey area in between where it's less clear. <math>...</math> generally shouldn't be used for simple stuff in running text if it can be avoided, especially for single letters. The Media Viewer also won't render <math>...</math> in image captions, so it should be avoided there if at all possible as well. As I said before, it hinders accessibility and prevents copying to the clipboard. Formulas can be of mixed type in different parts of a page. Simple stuff is often just as easy to type up either way. These comparisons of byte counts are fairly meaningless. –Deacon Vorbis (carbon • videos) 02:43, 18 March 2020 (UTC)

I see that adding {{math}} can make the equations nicer; however, that doesn't address that they're still incredibly hard to type out and therefore are error prone. That's why I brought up bytes and characters; you're pressing the keyboard a hundred times more, and inputting more information for the computer, than you should be for something which looks worse than if you had decided to use LaTeX. Overall, it's time consuming. Now, if someone wants to go ahead and waste their own time typing out poorly formatted math equations, that's fine. But that is at the expense of the interested reader. And another thing: this doesn't address is, again: how are you going to write out an integral? A fraction? LaTeX clearly saves the day. But if you use both math styles, then the issue of mixed math rendering arises which is yet again at the expense of the reader. I can personally say that, years ago when I was first starting out as a math student, I was very confused by the mixed math notation on Wikipedia and was wondering why on earth people were rendering math by italicizing letters of the alphabet. But then I would always be relieved to see LaTeX-ed equations; just like they appear in books, notes, research papers, i.e. literally everywhere a professional mathematician needs to communicate their thoughts through some kind of external medium. Lltrujello (talk) 21:59, 19 March 2020 (UTC)

Some comments

1. Changing systematically from html to latex is highly time consuming. Althoug the process could be automatized, this is not the case. Moreover, this change requires experimented editors who understand the formulas. Their time would be much more useful, if spent for improving the phrasing of the articles. Surprisingly, this is often the most elementary articles that are badly written (and also that have badly formatted formulas).
2. Changing from raw html (''x'') to {{math}} is fast and easy (select the formula, click on the button {{math}} in the "math and logic" menu, and check special characters like =, |, {, })
3. {{mvar|x}} is easier to type than <math>x</math> (10 vs 14 characters)
4. Inline latex has some alignment issues: $x$ for example

Therefore, my practice and my recommendation are

• For new content, use {{math}} for very simple inline formulas, and <math> otherwise.
• For existing content, add {{math}} to all formulas in raw html that appear in the edited section.

D.Lazard (talk) 10:37, 18 March 2020 (UTC)

Using Visual Studio Code I can copy and paste the existing text and systematically change the math to LaTeX, so it wouldn't be too time consuming. I see and agree with your points, but it is unfortunately more time consuming to swap back and forth between different types of math rendering than just putting everything to LaTeX once in VS Code. I don't want to spend time on this page editing every math equation, deciding whether or not to use {{mvar}} or LaTeX, hence why I proposed a systematic clean up to LaTeX. And LaTeX is pretty and easy to write. But I'll just follow your advice and edit more niche math articles that need more help than just styling updates. Perhaps maybe in the future people will be more open to a full LaTeX switch, but for the now the article will stay a bit ugly, which is unfortunate for the future interested readers. Lltrujello (talk) 21:59, 19 March 2020 (UTC)

Example for associativity in 2nd Example wrong?

In the example for the associativity in Group_(mathematics)#Second_example:_a_symmetry_group, it seems like the right before left rule hasn't been followed? (f_d ∘ f_v) gets simplified to r₃ while it should probably be r₁ if I understood the article correctly. If not then the beginning example explaining the right to left rule is probably wrong. 2A02:908:617:1580:F9FC:1277:BB22:AB42 (talk) 02:30, 16 September 2020 (UTC)

It's right as is. The way the world writes function composition, especially when those functions are group elements, is just plain confusing. Follow vertex 1 for example: $f_{\mathrm {d} }\circ f_{\mathrm {v} }$ first applies a vertical flip, so vertex 1 moves to vertex 4, and then a diagonal flip, where vertex 4 stays fixed. So in the end, vertex 1 has moved to vertex 4. Following the other vertices, one finds they're all also rotated by 90 degrees counterclockwise, which corresponds to $r_{3}$ in this notation. –Deacon Vorbis (carbon • videos) 02:48, 16 September 2020 (UTC)

Thank you for the quick explanation, I can follow your reasoning (and the examples on the page) looking at the images. I assumed the cayley table also needs to be read right to left, which is where my confusion stemmed from. 2A02:908:617:1580:9CB6:8597:745A:1032 (talk) 03:33, 16 September 2020 (UTC)

rectangle 8 symmetries

The page states rectangle has 8 symmetries an describes them, but no where explains why there are only these eights and no other ones. Clarification added.

If someone can improve the clarification, please do it, but do not delete it without cover the previous gap in the content. — Preceding unsigned comment added by 88.6.183.73 (talk) 20:09, 21 November 2020 (UTC)

I have reverted again your edit with the following edit summary: This is not an article about the symmetries of the square. The important aspect of this example is that these 8 symmetries form a group, not that the square has no other symmetry. So this paragraph disrupt reading.. So I support the reverts done by two other editors. Please note that you have done four times the same edit. This violates the WP:3RR rule, and you may be blocked for editing because of this violation. By WP:BRD, for being accepted, your edit must reach a consensus here. This is improbable as three editors think that your edit does not improve Wikipedia. So please, read WP:Edit warring and stop edit warring. D.Lazard (talk) 20:55, 21 November 2020 (UTC)

Some feedback that someone more qualified might be able to use to improve the article

As someone only just beginning to learn what a group is in mathematics, here's some feedback on the article: 1) Multiplicative and additive groups aren't defined (unless of course their definitions are merely and *entirely* notation related); 2) The article surely can't claim that, "There is no mathematical difference between a multiplicative group and an additive group; the difference is only in the notation," and then go on to claim that "An abelian group may be notated as a multiplicative group or an additive group, but a nonabelian group is always a multiplicative group," because the latter implies that there *are* indeed mathematical differences between additive and multiplicative groups. (Right??)

I think the article would be improved by clarifying what's meant by "composition" in "When the group law is composition" (under Notation and terminology). I later figured out that it means applying more than one valid operation in sequence (like function composition, right?) as per the "Second example: a symmetry group". Perhaps someone more knowledgeable than me could make this clearer / more explicit at that first mention of composition?

51.219.141.160 (talk) 16:11, 2 January 2021 (UTC)

Thanks for the feedback. I agree that section "Notation and terminology" requires to be rewritten. The issues that you quote are not the only ones. Among other issues, one can remark that the lead and section "Definition" call group a set equipped with an operation, while that section call group an ordered pair of a set and an operation. This is not the same, and requires explanation.

I'll try soon to improve the section. D.Lazard (talk) 18:40, 2 January 2021 (UTC)

 Done D.Lazard (talk) 17:34, 3 January 2021 (UTC)

I think that your changes will be very helpful to beginning readers. I agree with almost all of them. (I will make a few small wording changes.)

The one thing that seemed odd to me was to call ≤ an operation. I think it is more common to treat it as a relation instead of an operation taking values in {true,false}. Moreover, calling it an operation might be confusing in the context where operations are supposed to map a pair of elements to an element of the same set. Ebony Jackson (talk) 18:11, 3 January 2021 (UTC)

Maybe, I was influenced by my experience in computer algebra. I agree to restrict the explanation to the two classical operations. Also, this could make clearer what follows. However, I would certainly not do the modification myself today... D.Lazard (talk) 18:23, 3 January 2021 (UTC)

OK, thank you; I tried implementing a version of this. Ebony Jackson (talk) 21:28, 3 January 2021 (UTC)

cref/cnote templates

The {{cref}} / {{cnote}} system of templates seems to be painful, and I suggest that it should be replaced: we have better ways of doing this, at the cost of having the notes in the text stream (which we already have for references). In particular:

• Hovering over the note tag does not seem to show the text of the note in a pop-up. You actually have to click on it to navigate to the note to read it. Unnecessarily painful for the reader and unnecessarily disruptive to reading flow.
• The correspondence between reference and the note has to be maintained in two different places, and is a maintenance headache. For example, tags [b], [c], [f] do not exist, leaving three unreferenced notes.
• Sequencing is a maintenance headache. Keeping alphabetic sequence corresponding to the text sequence, avoiding gaps in the sequence, dealing with re-ordering the text, avoiding duplicate tags – unless we don't care (except for the last point: duplicate tags cannot be used).

Or does someone have a different perspective? —Quondum 21:12, 7 May 2021 (UTC)

Hm, I agree it is a bit prone to rot... What template would you suggest instead? Jakob.scholbach (talk) 13:09, 8 May 2021 (UTC)

{{efn}} with {{notelist}} seems to work well. See an example in Ricci calculus. There are also related templates such as {{sfn}} that seem to be compact equivalents of {{harvard citations}} that we already use, though there is no need to change these. —Quondum 14:10, 8 May 2021 (UTC)

OK, that looks better. I started implementing this change - would you be willing to help out with some more? Thanks! (Please don't delete the ones that are not referenced, those should be reinserted at appropriate spots in most cases, I believe.) Jakob.scholbach (talk) 15:33, 8 May 2021 (UTC)

Yes, no problem. I don't promise that it will be immediate, but I'll probably complete this within a few days. —Quondum 01:36, 9 May 2021 (UTC)

Done. That was quicker than I expected. —Quondum 02:18, 9 May 2021 (UTC)

Thanks! I reintegrated two that still made sense, and removed one that was never used. Jakob.scholbach (talk) 12:13, 9 May 2021 (UTC)

Formatting

The article is a bit inconsistent w.r.t. formatting and in need of cleaning up, and it seems that we should get a consensus on format choices. Group-theoretic articles do not generally need fancy formatting, and often achieve a uniform look by using minimal HTML formatting, even forgoing LaTeX in stand-alone formulae, and sidestepping many of the problems associated with LaTeX. It is not a big deal to get this consistent with any given formatting choices, once these are settled. Some choices:

• Inline formulae: LaTeX or HTML?
• Standalone formulae: LaTeX or HTML?
• HTML formatting: {{math}} or {{nowrap}}?
• Known set symbols (reals, rationals, etc.): bold or blackboard bold? (Blackboard bold is not really acceptable in HTML: it creates a nasty mess of sizing in some common browsers. So either avoid BB or use LaTeX here as an exception in HTML.)
• Specific elements of sets: italic or roman?

—Quondum 12:47, 4 May 2021 (UTC)

It is bad practice to choose different formats for inline vs standalone formulas, because then the variables have different appearances and usually in mathematics different appearances of the same letters are a marker that they have different meanings. This intentional use of different meanings for different formats can be seen already in this article's "Second example: a symmetry group" section, where we have both a roman c (representing a counter-diagonal reflection) and an italic c (representing an arbitrary group element). So since we have some formulas that are too complex for non-LaTeX formatting, for instance the one in the "Galois groups" section, I think we should make them all LaTeX to make the formatting more consistent than it is now. The math templates try to mimic LaTeX appearance but just don't succeed, and plain-html formatting is right out. I also prefer blackboard bold for the symbols that can be formatted that way; I think that's more standard these days and it makes clear that those symbols have a specific meaning. As for "specific elements of sets: italic or roman": I think you are seeing an inconsistency where there is none. We are using roman for specific geometric operations whose roman letters stand for the name of the operation, and italic for abstract elements of sets; that is deliberate and a good use of distinctions in notation to make distinctions in meaning. —David Eppstein (talk) 16:06, 4 May 2021 (UTC)

On "specific elements of sets: italic or roman", there was inconsistency until I recently made changes to this, and am checking whether consensus is aligned with this. I'll wait for more comments, though I think perhaps I've stirred up a hornet's nest. —Quondum 17:01, 4 May 2021 (UTC)

Yes, the article was (and still is) inconsistent. When this article became an FA in 2008 we had every math bit formatted as ordinary text (italics for variables, R for reals), except where this was not possible or not reasonably possible. This was reasonably consistent and looks acceptable. Then, apparently some editor thought "let's make it better", and forgot to go all the way down to the end. Now, the majority of math symbols is formatted using the math template (except where this is not possible), which also looks OK to me.

My conclusion is that the minimally invasive thing would be to replace all italics by { { math ... } } templates.

I personally (apparently contrary to David's preference) am not at all in favour of displaying every stupid math symbol by LateX, this breaks a lot of the reading flow. Jakob.scholbach (talk) 20:41, 4 May 2021 (UTC)

Well, the likelihood that someone would express a contrary opinion is one reason I didn't just rush ahead and do it the way I would prefer. —David Eppstein (talk) 04:54, 5 May 2021 (UTC)

Yeah, this stupid formatting is a safe way to enrage everybody involved... That said, I don't have a strong opinion. As long as it is uniform I'm OK with pretty much anything. Jakob.scholbach (talk) 14:08, 5 May 2021 (UTC)

I think I'll leave it be until some consensus is apparent. I'm a little surprised, given that this is nominally a FA. —Quondum 14:19, 5 May 2021 (UTC)

Well, 2008 was plenty long enough ago for inconsistencies to creep in. I'm in the "make them all LaTeX" camp with David Eppstein, I think. XOR'easter (talk) 15:43, 5 May 2021 (UTC)

(unindent) OK. What about this compromise?: we use LateX for

• all stand-alone expressions,
• all inline expressions involving sub/superscripts (":..., a⁻³, a⁻², a⁻¹, a⁰ = e, a, a², a³, ...,")
• all inline expressions using other "serious" notation e.g. $\mathbb {Z}$ etc.

We use html for

• inline symbols that can just as well be rendered using html ("For each a in G, there exists an element b in G ...)?

I don't really see the point in using Latex for those. Jakob.scholbach (talk) 08:57, 6 May 2021 (UTC)

I have uniformized the notation in removing all {{math}} templates (and replacing them by {{nowrap}} where needed). I have replaced the Z by $\mathbb {Z}$ etc. Some of the stand-alone formulas are not yet in Latex, I will do this soon. Jakob.scholbach (talk) 18:44, 7 May 2021 (UTC)

You might want to consider the formatting of "literal" elements. For example, f_c versus f_c. —Quondum 19:40, 7 May 2021 (UTC)

I still think that we should make an effort to make $f$ in displayed formulas look the same as the f in inline formulas, which html (as you can see in this sentence) does not even come close to. If you insist on not making everything <math>, at least use the {{math}} templates rather than plain html: f. Jakob.scholbach's removal of the math templates is in this regard a horrible step backwards. If you're going that far, why not

write everything in monospaced ascii instead of even using formatted text

? —David Eppstein (talk) 21:21, 7 May 2021 (UTC)

I agree that the difference is more marked than ideal. Regarded on its own, that is easily solved (make standalone formatting as for inline, with the single awkward example of the quadratic root equation handled separately; the actual formula's relevance is too obscure anyway, so could be removed). I deliberately am staying away from actually expressing a preference; I am just observing that one can achieve consistency of format fairly easily in an article such as this, since it can get away without fancy formulae if we wish (as do many abstract algebra articles). —Quondum 21:40, 7 May 2021 (UTC)

It's not just that the html formatting is horribly inconsistent with the displayed-math formatting (although it is). Other problems with the html formatting include bad spacing around operations and relations (too wide around the centered dot, too narrow around equality; too close to variable in "b/a" in "Division" section), a centered dot that in my default rendering is so tiny that it can barely be seen (the one in the math template is better) and a raised circle for composition that cannot be distinguished from a centered dot, greater likelihood of forgetting to italicize variables (there currently is a roman f in "Second example: a symmetry group", for instance); insufficient distinction from article text (visible most clearly in the phrase "an element a in a group G"; there are also some examples where the text is italic, exactly matching the variables); inability to show nesting of parentheses more clearly by varying their sizes (fortunately the examples here are small, "Φ(φ(g))": whose brilliant idea was it to use a roman and italic phi for inverse homomorphisms?). —David Eppstein (talk) 00:33, 8 May 2021 (UTC)

I guess any such discussion will quickly open Pandora's box. I believe there are arguments in favor of any of the several options. This is, probably, why WP:MOSMATH says "For inline formulae, such as a2 − b2, the community of mathematical editors of English Wikipedia currently has no consensus about preferred formatting". Everyone has their opinion, and is entitled to their opinion. My personal opinion is based on the fact that it looks quite OK in my browser, and that the article passed FA scrutiny with such a formatting.

David Eppstein, if you have such a strong opinion about this, would you be willing to change the entire article to Latex? As I said earlier, I don't personally really like it (people who compare it to the usual Latex typesetting process are usually neglecting that in a usual math paper the text and the formulae are rendered using the same font, which is not at all the case if we use Latex here). But in the interest of a reasonably productive editing on the article, I wouldn't dig my heels in. It is just that someone needs to do that work. I don't see the clear benefit of having Latex all over the place (on the contrary), so I won't commit to such a time-consuming edit which is not forced upon us by MOS. If others have a strong opinion either way, go ahead (but only do it if you do it all the way through). Jakob.scholbach (talk) 07:20, 8 May 2021 (UTC)

If there's consensus to make this change I'm willing to put in the effort, but I probably won't have time until this Sunday. —David Eppstein (talk) 07:22, 8 May 2021 (UTC)

LaTeX is neater within itself, but does not embed well into the surrounding text. HTML does not work well for expressing math, but formats coherently with the surrounding text. Compounded by WP defaulting to an arbitrary browser font, we have that every compromise is poor, inviting disagreement and instability. If a stable consensus on the chosen style can be clearly recorded (e.g. at the top of the talk page of the article), I think editors of all stripes will put in the effort to conform the article to that style. —Quondum 12:52, 8 May 2021 (UTC)

@David Eppstein: IMO, go ahead. (I would probably suggest trying it out first with the table of group elements, where it might create quite some bigger width.) Jakob.scholbach (talk) 12:15, 9 May 2021 (UTC)

Reliability of Becchi

Is Becchi "Introduction to Gauge Theories" a reliable source? It seems to have been published only on arXiv, and that's not peer-reviewed. —David Eppstein (talk) 20:50, 9 May 2021 (UTC)

The claim is covered by the standard textbook referenced at the end of the next sentence, so I removed the Becchi citation. XOR'easter (talk) 23:37, 9 May 2021 (UTC)

math display=block

The markup <math display=block> behaves a bit poorly, very slightly on a computer browser, but more markedly on a mobile phone browser. In particular, vertical spacing before and after seems to be erratic and quite different from that on a normal browser, including that a single blank line (normally no effect) before or after it has quite a marked effect. I suggest that we stick to the better-behaved :<math>. —Quondum 17:45, 10 May 2021 (UTC)

Using colons for indentation violates accessibility guidelines (MOS:INDENTGAP) and is discouraged by MOS:MATH. The display=block markup is intended to provide a more-accessible way of formatting mathematics displays. The blank lines have an affect because (unlike :-indentation) it matters whether the displayed block is part of a surrounding paragraph or stands alone as its own paragraph. I suggest we stick to the guideline-approved and accessible display=block. In particular, part of the featured article criteria include following style guidelines, and these are style guidelines that we should be following. —David Eppstein (talk) 17:56, 10 May 2021 (UTC)

Okay, whatever. It is possibly just a quirk of the browser I'm using, as widespread as Safari on mobile is. I'll just have to put it down to another of those things where WP formatting just does weird things, I guess. Which just is another thing to dampen enthusiasm for getting things in WP nicely polished. —Quondum 18:16, 10 May 2021 (UTC)

I agree both that mathematics formatting on Wikipedia is subpar compared even to ten years ago on the rest of the web, and that the vertical spacing around display=block math is not right. But I think it's the best we can do under the circumstances. The Wikimedia developers have a long-term pattern of being incredibly unresponsive to these issues, and of actively discouraging local efforts to work around them. —David Eppstein (talk) 18:20, 10 May 2021 (UTC)

We seem to be on the same page there. I guess keeping the HTML semantics correct, notwithstanding formatting quirks, is probably a worthwhile objective. —Quondum 18:28, 10 May 2021 (UTC)

<math> on mobile devices

Tracked in Phabricator
Task T129054

I've just noticed that the <math> markup in footnotes displays as blank white boxes on a mobile device (or at least on Safari on iPhone). We'll have to abandon this markup in that context. —Quondum 17:52, 10 May 2021 (UTC)

It looks fine for me on the Android app. Are you sure this is a widespread and long-term issue and not just an issue with your particular setup or something that we could reasonably expect to be fixed? —David Eppstein (talk) 18:00, 10 May 2021 (UTC)

I am using the latest on iPhone. As such, it may be universal on all iPhones and thus most likely widespread with probably no known fix date (it would depend on Apple), but I have not checked. It would be helpful if other editors with Safari on iPhones could give feedback on this. —Quondum 18:07, 10 May 2021 (UTC)

Math markup seems unavoidable in extended footnote L as written ("The same is true for any field F instead of $\mathbb {Q}$.") because of the use of blackboard bold. The html/template/unicode substitutes for that are so badly inconsistent across browsers that MOS:MATH now explicitly discourages their use. —David Eppstein (talk) 18:14, 10 May 2021 (UTC)

It seems that this problem appears in many browsers: I have tried Mozilla Firefox, Edge and Brave all on Windows 10, plus Safari on iOS 14, all with the same effect. Navigate to https://en.m.wikipedia.org/wiki/Group_(mathematics)#Definition and then select the footnote [a] (click or tap) at the end of the first paragraph. All the <math> comes out as blank rectangles. We have one negative datapoint for the Android app. Was this to the .m. site? —Quondum 17:30, 11 May 2021 (UTC)

It was the Android Wikipedia app. An app, not a site. I can see the problem you describe for popups on footnotes on the mobile site in an ordinary browser on an ordinary computer. The footnotes themselves are rendered just fine at the bottom of the article. The problem is not in the footnotes at the bottom of the article, but only in the appearance when you click on them and get a popup. The problem appears to be that the math markup is rendered as black text on a transparent background and that the popup uses white text on a black background, causing the math to become invisible. —David Eppstein (talk) 17:54, 11 May 2021 (UTC)

Turns out to be a bug known and listed as a bug for five years. I think the snail-like progress is another example of the Wikimedia developers' poor prioritization of acceptable mathematics rendering. —David Eppstein (talk) 18:06, 11 May 2021 (UTC)

Thanks for digging that out and clarifying it. I'll leave it be, since it is nothing new. —Quondum 18:12, 11 May 2021 (UTC)

Incidentally, there is another math rendering bug that seems to be even lower priority for the developers: in mobile view, if you tap on an image to go to a full-screen view of the image, math in the caption is rendered incorrectly (T263572). —David Eppstein (talk) 18:14, 11 May 2021 (UTC)

TFA nomination

I have nominated this article to run as today's featured article for an unspecified date. Editors may join the discussion for this nomination at Wikipedia:Today's featured article/requests/Group (mathematics). Z1720 (talk) 20:04, 17 January 2022 (UTC)

Modular arithmetic

There's been a bit of back-and-forth on the modular arithmetic section lately, which in light of the FAR I thought worth discussing here. Before the FAR, it discussed the modular arithmetic operations as acting on the numbers $0$ through $n-1$. Someone recently tried changing the elements being operated on to be equivalence classes of numbers, with a cumbersome notation for distinguishing them from the numbers they represent. Then another variation kept the equivalence classes but discussed them in terms of their "representants" (why not representatives, a much more common word with the same meaning?). I have restored something more like the original version, which performs the arithmetic directly on representatives and doesn't mention equivalence classes at all (although it does mention the equivalence relation from which the classes come). Using equivalence classes as group elements is not in any way a more rigorous treatment of this topic; all one needs is a clear statement that one is using a modified arithmetic operation on numbers that differs from the usual addition and multiplication operations by reducing to representatives. The use of equivalence classes makes the treatment both unnecessarily detailed and unnecessarily WP:TECHNICAL, something I think we should be trying to avoid in a featured article. Equivalence classes may be preferred in some modern treatments aimed at mathematics students because of its symmetry (avoiding the need to choose representatives) or because using equivalence classes makes the proofs of associativity etc more direct, but those are not good reasons for preferring them here. In fact, from the advanced mathematical point of view equivalence classes come with some unwanted baggage of their own (far beyond the scope of this article): handling them involves second order arithmetic while there is no reason to go beyond first order merely to define modular arithmetic.

Relatedly, a comment by User:Quondum in an edit summary suggested that, aside from this issue, the modular arithmetic section is overly detailed, and that it could be trimmed while directing readers to the main article for the missing detail. I tend to agree, but rather than just slashing it myself I thought it would be better to discuss here first. My suggestion would be to cut the detailed justification for why the multiplicative group is a group (everything after footnote [n] except for the final sentence of the section about notation and applications); what do others think? —David Eppstein (talk) 00:55, 14 May 2021 (UTC)

The section is almost entirely redundant with the cyclic group section that follows it; its merit is that the additive group in modular arithmetic might be a familiar manifestation of the finite cyclic groups, and this can be conveyed in a mention in the next section. My inclination is to have the article only point out that modular addition corresponds to the finite cyclic groups with a link (the multiplicative group of a ring being too involved for those it is intended to help), and maybe keep the clock diagram for those who need a picture. Others may think differently from me. —Quondum 02:30, 14 May 2021 (UTC)

If one wants a less radical approach, defining a "+_{mod n}" operation on an actual set of integers {0, ..., n−1} is what people are familiar with from programming and is rigorous. Maybe closer to what David has in mind. —Quondum 14:46, 14 May 2021 (UTC)

I didn't have the time to follow the recent edits in that section closely, but what we currently have strikes me as the correct approach. Equivalence classes should be kept to an absolute minimum.

About trimming this: I am decidedly against that idea. Yes, for a pro reader this is entirely redundant. On the other hand, the purpose of the modular arithmetic section is to tell people "you already know this concept, maybe without knowing its name". The purpose of the cyclic group section is to turn the screw one turn more, by embedding what people (likely) know into a bigger, and more abstract context. In my mind these two sections are an example of a well-done pedagogical approach, and the length we spend there is not in vain. Jakob.scholbach (talk) 18:58, 14 May 2021 (UTC)

I'm confused. "Wikipedia is an encyclopedic reference, not a textbook. The purpose of Wikipedia is to present facts, not to teach subject matter.": a 'pedagogical approach' is expressly not the style in WP. —Quondum 19:22, 14 May 2021 (UTC)

Well, I'm not advocating writing a text-book (and this is not what the article is doing). I'm just pointing out that the arrangement of the material is a very reasonable choice from a perspective of someone who doesn't know these things yet.

You could make the same objection to the very first section about the integers, the symmetry group and the definition: in a way this is highly repetitive, basically repeating itself three times. Yet, this is on purpose, and BTW was very much requested throughout the FA process. (Take a look at the FA nomination discussions, if you want to convince yourself!). Jakob.scholbach (talk) 19:29, 14 May 2021 (UTC)

Many mathematics articles on WP are not a useful encyclopedic reference because they did not think about the reader. Presenting facts is difficult if you're not allowed to explain the context. I agree with keeping this simple, without equivalence classes. FemkeMilene (talk) 19:36, 14 May 2021 (UTC)

@Jakob.scholbach: Cyclic groups are simpler to introduce than is modular arithmetic. Those who are not familiar with the latter would be better introduced to this via the former. I would advocate introducing the cyclic groups and then take the approach of "but hey, you know these already in the form of addition in modular arithmetic."

@Femkemilene: We're all agreed on presenting the information in the simplest form possible, just not at the expense of accuracy. No-one is arguing to bring back the description through equivalence groups. I would like to see the terms "equivalent" and "representative" removed entirely. I suggested an approach to do so above. —Quondum 20:04, 14 May 2021 (UTC)

I firmly disagree with the idea that cyclic groups are easier than modular arithmetic. Cyclic groups require to first know groups, then internalize what it means to take iterated powers and then get an idea "what that actually means". Modular arithmetic needs ordinary arithmetic, and the process of division with remainder. Jakob.scholbach (talk) 20:08, 14 May 2021 (UTC)

Needless to conclude which one is easier! Jakob.scholbach (talk) 20:09, 14 May 2021 (UTC)

The concept of the possible rotations of a n-gon that leave it the same, visualized through equally spaced points on a circle with a reference point (or else directed arcs starting at one point)? One does not need any understanding of groups to understand this. —Quondum 22:17, 14 May 2021 (UTC)

I agree that I think geometric examples are easier than those arithmeic-based examples where you need machinery to describe the operations. We almost have the group of rotations on the square in the article already: it is the subgroup of the symmetry group we give as our second example that you get by removing reflections. — Charles Stewart (talk) 01:17, 15 May 2021 (UTC)

• I haven't yet formed an opinion on which presentation I think is better for the article, but I want to push back on David Eppstein's argument "In fact, from the advanced mathematical point of view equivalence classes come with some unwanted baggage of their own (far beyond the scope of this article): handling them involves second order arithmetic while there is no reason to go beyond first order merely to define modular arithmetic" - this 'baggage' is trivial, just sets of numbers (and if you care about such things, primitive recursive sets at that). It's not second-order logic, it's two-sorted first-order logic. I think this comes from a view that single-sorted logic is 'simpler' than multisorted logic, but I think that is an old-fashioned view, and one with baggage of its own. — Charles Stewart (talk) 00:59, 15 May 2021 (UTC)

Re Quondum: I think you are mixing up the complexity of two things:

• the symmetry group consisting of rotations of an n-gon
• the general concept of a cyclic group

I tend to agree that the 1st point is somewhat easier to understand than modular arithmetic. As Charles points out, this would not go far beyond what we have in the intro. (Just having rotations in the intro would mean we end up with an abelian group, which is unrepresentative.)

The notion of a cyclic group (2nd point) is certainly more abstract and thus more advanced than modular arithmetic: the latter admits a complete, non-fuzzy description accessible for everyone knowing about division. By contrast, you can't even say what a cyclic group is (in a way that is not just listing Z/n-style and Z-style examples) without knowing what a group is. This is not a reason to avoid talking about cyclic groups, but it is a good reason to introduce (in a reasonable depth) an example of an (otherwise also important) group that eventually serves as an example for a more advanced notion. Jakob.scholbach (talk) 21:04, 15 May 2021 (UTC)

What precisely you are saying eludes me. Do you feel that replacing the modular addition prelude to cyclic groups with the n-gon symmetry example would be an improvement? I never was suggesting that we replace the more abstract description of cyclic groups. —Quondum 01:40, 16 May 2021 (UTC)

I have been contesting the idea that cyclic groups are simpler than modular arithmetic. The net effect of this assessment on the article is that modular arithmetic should (continue to) precede cyclic groups.

More generally, these sections there are IMO well in place, and I see no need in changing anything there. Jakob.scholbach (talk) 13:15, 16 May 2021 (UTC)

It seems to me that the modular arithmetic example should be moved up as the "Second example," with the current second example made a third. The current second example section is tedious and hard to read. I wonder how many FAR reviewers went through it in detail. Modular arithmetic, by contrast is easy to grasp, particularly with the clock example. In my experience it is almost universally given as a first example of a group that differs from the ordinary forms of arithmetic (integers, rationals, reals). Moving it up would not alter the overall size of the article, just make it easier to understand. And I don't think we EVER need to worry about Wikipedia math articles being too easy to understand.--agr (talk) 15:36, 21 January 2022 (UTC)

Short Description

@Imaginatorium: I saw you reverted my edit of the short description. My edit removed content, but it was in-line with the purpose of the short description, see WP:SHORTDES. I made a few such edits recently and there is currently a discussion over in the Project Math talk page where I elaborate on my reasoning. To summarize here: the purpose of the short description is to briefly indicate the field covered by the article, and (explicitly) not to define the subject of the article. Notable examples exhibiting a similar degree of brevity include "American baseball player" for Babe Ruth or "U.S. State" for Florida. Feel free to add to the discussion if you wish to. Whether or not you agree with me, your opinion is welcome. Donko XI (talk) 11:24, 21 January 2022 (UTC)

Donko XI changed the short description from "Algebraic structure with one binary operation" to "Algebraic structure", and this was reverted by Imaginatorium. I support Donko XI's change for the following reasons. The previous short description is slighty too long (45 characters instead of the recommended limit of 40), but this is not the main reason for supporting the change. The aim of a short description is to allow readers to decide whether they may be interested in the article without open it. So the useful information is firstly that this is mathematics (this is already disambiguated in the article title), and to say which kind of mathematical object is the subject of the article (here, algebraic structure). Adding other details could be useful only if one could, in very few words, give an accurate definition, or, at least, disambiguate from other algebraic structures. This seems impossible. Moreover, emphasizing on the number of operations is wrong (the integers form a group that has another binary operation, the multiplication), and does not allow distinguishing groups from, say, monoids.

So Donko XI's is clearly an improvement of the short description, and I'll restore it. D.Lazard (talk) 13:58, 21 January 2022 (UTC)

I can't conceive of anyone for whom "Algebraic structure" is a meaningful description who does not already know what a group is. By contrast "American baseball player" and "U.S. state" are terms widely known to the general public. I would suggest "Set with an associative, invertible operation" (45 characters). This short description would at least give a reader with a little math background a concise reminder of what a group is. We give similar hints in the short description for Ring (mathematics), Field (mathematics) and many other articles about algebraic structures. I would also call attention to the template annotated link, which appends the short description, if it exists, to a link: "Used in lists to provide an annotated link using the short description from the linked page for annotation. Useful for disambiguation and providing an idea of what the link is about, without having to hover on the link." It's another reason to eschew absolute minimalism in short descriptions.--agr (talk) 02:54, 24 January 2022 (UTC)

You could get even closer to the 40-character soft limit by dropping the article and comma: "Set with associative invertible operation". —David Eppstein (talk) 07:26, 24 January 2022 (UTC)