As a non-native English speaker, I find it utterly confusing to read text containing figures like "1,234 m", which in English means 1234 m while for a person used to read e.g. French or German (person's native language) it'll mean 1234 mm at first, only to be re-read to reinterpret correctly.

There seems to exist a cure for this: according to Wikipedia,

Since 2003, the use of spaces as separators (for example: 20 000 and 1 000 000 for "twenty thousand" and "one million") has been officially endorsed by SI/ISO 31-0 standard, as well as by the International Bureau of Weights and Measures and the International Union of Pure and Applied Chemistry (IUPAC), the American Medical Association's widely followed AMA Manual of Style, and the Metrication Board, among others.

But, as the posts on StackExchange contain lots of numbers with comma as digit grouping symbol, I'm reluctant to change them to narrow spaces when editing, because I worry that this may be regarded as a question of style, although I suppose using spaces will help non-native-English audience avoid mistaken reading of the numbers — especially 4- to 6-digit long ones, where there's only one comma (two or more commas is a hint already that it's not integer-fraction separator). So far I've approached this conservatively, only fixing numbers lacking integral part in a decimal number (like .1345 instead of 0.1345) by prepending zeros to them, to make it more obvious that there's a point before the digits.

My question is, is it acceptable, when fixing formatting of a post, to replace commas in numbers to narrow no-break spaces (U+202F)? Or are there any good reasons why I shouldn't do it?

This question has also been asked on general Meta.SE, where a suggestion is that the answer may be site-specific, thus this question on Physics.SE.

  • $\begingroup$ I prefer 62626 over 62,626 and 62 626. $\endgroup$ Commented Aug 17, 2018 at 9:43
  • 4
    $\begingroup$ If the only thing you're changing is commas to spaces, then I'd reject it asno improvement whatsoever. If it's something you've added to other changes, I possibly would be fine with it. $\endgroup$
    – Kyle Kanos
    Commented Aug 17, 2018 at 10:32
  • 3
    $\begingroup$ If you're replacing commas with spaces then I would reject that as clearly conflicts with author's intent. If you're adding spaces to unformatted numbers as part of a larger edit, I'd probably be fine with it. $\endgroup$ Commented Aug 17, 2018 at 10:46
  • $\begingroup$ @KyleKanos I'd actually encourage you to submit that as an answer, since it is a legitimate alternative to the existing answers and I think we should see how the community thinks about it. $\endgroup$
    – David Z
    Commented Aug 28, 2018 at 18:57
  • $\begingroup$ @EmilioPisanty Same thing, I'd suggest making your comment an answer. $\endgroup$
    – David Z
    Commented Aug 28, 2018 at 18:57
  • $\begingroup$ @DavidZ it seems to me that rob's post largely contains what I said (cf. 1st full paragraph under the 2nd quote) $\endgroup$
    – Kyle Kanos
    Commented Aug 28, 2018 at 19:07
  • $\begingroup$ @KyleKanos Hm, I thought you were saying something different. Anyway, that's fine. I was thinking of deleting the comments (at least yours and Emilio's), though, since they seem to be more answer-like. $\endgroup$
    – David Z
    Commented Aug 28, 2018 at 20:27
  • $\begingroup$ @DavidZ i can see the difference (one is before the action, the other is after) but it's still the same idea. Feel free to clean up my comments! :) $\endgroup$
    – Kyle Kanos
    Commented Aug 28, 2018 at 20:36
  • $\begingroup$ @DavidZ I also find rob's answer to supersede my comment. Then again, the vote counts on Kyle's comments and mine are nontrivial information, but I don't see a huge need to clean-up or move them to answers given the existing, much more expansive ones. $\endgroup$ Commented Aug 29, 2018 at 18:19

2 Answers 2


From our editing guidelines:

Edits are expected to be substantial and to leave the post better than you found it. Common reasons for edits include:

  1. To fix grammar and spelling mistakes
  2. To clarify the meaning of the post (without changing that meaning)
  3. To include additional information only found in comments, so all of the information relevant to the post is contained in one place
  4. To correct minor mistakes or add updates as the post ages
  5. To add related resources or hyperlinks

There's always some tension in meeting all of these guidelines. For instance, there's a software-enforced restriction where single-character and few-character edits can't even be submitted, on the assumption that they are not substantial. That's not always correct: sometimes an author will accidentally a key word, and the omission of that key word renders the meaning of an important section of their text ambiguous or incorrect. In those cases different editors choose different techniques to evade the software limit. Some find some other cosmetic improvements to make to the post (good); others fill out the character count with invisible whitespace changes (not so good).

The current answer to your Meta.SE version of this question contains this nice guidance:

"Which version of English do we use?" [...] whichever the author wishes. For example, we do not edit "favorite" to "favourite" (or the reverse)

which is used to justify not editing intranumerical commas into spaces or periods.

Here's what I would suggest. If you are reading a post and this internationalization issue is the primary thing that you would like to fix in an edit, then in general it's better to let the post stand as it is rather than coming up with some other "padding" to "substantialize" the edit. But if you're reading a post that has several minor issues --- maybe some grammar problems or LaTeXing ambiguity --- and one of those issues just happens to be ambiguous or missing interdigital separators, then you might as well replace the ambiguous separators with a space. If the numbers are in MathJax-typeset text, the TeX thin space \, is nice, and less confusing for future editors than an invisible Unicode character with strange behavior.

Along the same lines, please don't make a project of digging back through the archives and fixing minor problems like this on inactive posts; that does distort the front page. If you're looking for a project, pick a tag that you like, look for old unanswered questions, and write good answers to them. That's the main reason we're here, after all.

  • 3
    $\begingroup$ "That's not always correct..." Well played. $\endgroup$
    – Kyle Kanos
    Commented Aug 18, 2018 at 1:35
  • 2
    $\begingroup$ The point about confusing non-standard Unicode is extremely important. It's extremely hard to detect that those are in play even if you know that they're a thing and you have the toolset (say, this) to diagnose them correctly. All posts should be left in a state where they're editable by any other user - and frankly that means that Unicode shenanigans are right out. Particularly since there's a transparent MathJax solution. $\endgroup$ Commented Aug 18, 2018 at 14:18
  • 1
    $\begingroup$ Does the LaTeX/MathJax \, way achieve the goal of avoiding splitting the number into two? I chose narrow no-break space exactly for this reason. If it does, then indeed on Physics.SE it's a much better option. $\endgroup$
    – Ruslan
    Commented Aug 18, 2018 at 17:26
  • $\begingroup$ @Ruslan Seems so. A little experimenting by writing $1\,000\,000\,\cdots$ and watching the preview suggests that short MathJax sections will get moved to the beginning of the next line rather than broken, no matter what they contain. Very long inline MathJax has more complicated linebreaking rules. $\endgroup$
    – rob Mod
    Commented Aug 18, 2018 at 17:41
  • 1
    $\begingroup$ "there's a software-enforced restriction where single-character and few-character edits can't even be submitted" true, but only for users with less than 2,000 reputation. $\endgroup$
    – TylerH
    Commented Aug 27, 2018 at 15:20

My opinion: I am perfectly ok with editing posts to comply with stylistic standards and, in particular, I agree commas look silly in science. But:

  • If you are a <2k user, don't suggest an edit just to remove commas. A reviewer will have to go through your edit and, if minor, it will only waste their time. Make the edit significant or don't do it.

  • If you are a >2k user, your edit will not require reviewing, so in principle you are only wasting your own time; you do you. But please don't just remove a comma on old posts -- this would bump the question into the front page, thus wasting reviewers time. Make the edit significant or don't do it.

  • 3
    $\begingroup$ Let's add to this that they should first be sure the author of any post intended, for example, "1,234" to be "1234" and not "1.234" because changing something meant as "1.234" to "1234" is actively harmful $\endgroup$
    – Jim
    Commented Aug 17, 2018 at 19:11
  • $\begingroup$ Just to be sure, you're also implying that such tiny edits made by >2k users are not inappropriate when the post is already in the top 5-8 active questions on the front page, right? $\endgroup$
    – user191954
    Commented Aug 19, 2018 at 3:48
  • 1
    $\begingroup$ @Chair Yes, kind of. I would say, anywhere on the front page, not only in the top 5-8. I'm not sure how long the front page is, but the top 20 seems fine to me. Of course, other people might draw the line somewhere else. $\endgroup$ Commented Aug 19, 2018 at 13:32

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .