# Changing $dx$ to $\operatorname d\!x$

Seems like someone has gone on a bizarre edit spree, deciding to randomly change $dx$ to $\operatorname d\!x$ (with no other edits). This is purely a matter of stylistic choice, and I find it really annoying to have ancient questions bumped for just this. Some of the edited posts are more than a year old.

Should such edits be discouraged? (I think so.)

• I have pinged the user you mention. (I think it is good to let them know about the thread on meta.) Oct 23, 2014 at 7:32
• Changing was not random, the \operatorname seemed more appropriate, how ever I want to hear what council of elders has to say on styling of differential operator. Oct 23, 2014 at 8:36
• @Arjang: Do you think we should edit $\bf R$ to $\Bbb R$?
– Asaf Karagila Mod
Oct 23, 2014 at 8:56
• @Arjang Surely it was random. The same questions you edited had several answers with italic $d$:s which you left unchanged.
– mrf
Oct 23, 2014 at 9:02
• I'm sure we've had discussions of innessential edits before. This certainly qualifies. Don't do it. Please. Oct 23, 2014 at 9:27
• @Arjang While I personally prefer the roman $\mathrm{d}$, not everybody does - but the more important thing is that even if $dx$ was categorically wrong (which it isn't), changing it bumps lots of old questions to the front page, hiding new questions, and it's not worth doing this for such a minor change.
– mdp
Oct 23, 2014 at 10:48
• @Arjang: Uniform and proper style? According to whom? I don't think there is such a thing as universally accepted style in math. Oct 23, 2014 at 10:49
• I dislike the Roman d and much prefer $dx$. I don't change other people's $\mathrm{d}x$, and I expect them not to muck about with my $dx$. Oct 23, 2014 at 11:21
• I wonder what would the reception would have been if this had been about $sin$ vs $\sin$, or $Aut$ vs $\operatorname{Aut}$. Specifically, I wonder if @Jyrki would have said the same thing? Oct 23, 2014 at 11:57
• @user1729 However, there are things that are not uniform. And anyways we all know the rules for minor edits. If someone edited an old post only to change $y=sin x$ to $y=\sin x$, I would certainly reject it. No matter I am a professional math typographer.
– yo'
Oct 23, 2014 at 12:20
• Secondly, I do not think blanket rejection is realistic: rejection should always depend on context. For example, I would reject a sin x to \sin x in an old post with few views and a negative score, but not in an old post with a high score and lots of views. So any "rules" we have should, in my opinion, be fuzzy and not concrete. Oct 23, 2014 at 13:10
• (cont'd) So it behooves me to do my Lindelof -> Lindelöf edits in bursts of 2-3 at max. Oct 23, 2014 at 13:45
• IMO, questions being "bumped" even by minor edits is a technology problem in StackExchange that should be addressed there (e.g. through a "this is a minor edit" checkbox like Wikipedia has had for years). That tiny improvements in style or typography are being officially discouraged because the software is too dumb to know when to bump and when to leave is surely not ideal.
– user139000
Oct 26, 2014 at 20:03
• @user1729 "There is no such thing as an $\mathbb{R}$-tree"? I have seen them typeset like this (and not just in papers I've written). Oct 26, 2014 at 23:41
• @pew But if the edit is not an improvement in typography, rather the opposite, like here? Oct 27, 2014 at 10:40

From typographical point of view, the question whether d in dx is upright or italic is not clear at all, it varies from author to author, from sub-field to sub-field and from journal to journal. Therefore I think that we should not unify the style across the site and rather leave it to the authors.

What is not questionable, though, is that dx has to be separated from the preceding contents by a thin space (\, in LaTeX/MathJaX), unless there already is a visual separation (which is the case more-or-less only when the integrand is a large parenthesis).

And no matter what, editing dx into \mathrm{d}x or \operatorname{d}x, as well as adding the thin space, it all certainly qualifies as a minor edit.

• Minor point: there is a difference between minor and too minor. For example, $x\mapsto -x$... is minor, but not necessarily too minor... Oct 23, 2014 at 12:04
• @user1729 No, I believe that the one you show is small (in length), but not minor (in importance) ;)
– yo'
Oct 23, 2014 at 12:15
• Touché...minor is an adjective, so really we each should have followed it by a noun. You meant "minor in importance" while I meant "minor in size". I doubt SE will change the close reason "too minor" based on this imprecision though... Oct 23, 2014 at 13:18
• @user1729 "too minor" rejection reason for edits does not exist any more.
– user147263
Oct 23, 2014 at 13:21
• Well, this tex.SE answer states ISO 31/XI suggests the upright version, which might be considered a standard. However I agree it's too minor to justify an edit - though providing a global \d or similar might improve the readability of the "correct" typesetting. Oct 23, 2014 at 19:31
• @TobiasKienzler I suggest leaving ISO for the engineers and staying with typography traditions in maths: tex.stackexchange.com/questions/14821/…
– yo'
Oct 23, 2014 at 20:09
• True dat... Personally I'd prefer spending less time with norms (or typography) and more time with contents, but somehow my job's currently quite the opposite :/ Oct 23, 2014 at 20:16
• +1 for mentioning the \  that should always precede a $dx$ in an integral. Oct 24, 2014 at 20:03
• @Arkamis It's not interword space \  ! It's the thin space \, !!!
– yo'
Oct 24, 2014 at 22:35
• Either space will do -- as long as there is a space. Oct 24, 2014 at 22:37
• I just think we need some space. Maybe take a little break, meet up at the next partition. Oct 24, 2014 at 22:39
• It doesnt cares absolutely nothing, being "official" or not it really doesnt care at all. It meaning is obvious, it doesnt lead to confusion. Oct 30, 2014 at 10:49
• @tohecz I prefer not to typeset a space every time (mainly due to my laziness), since \mathrm dx creates a visual separation by change of font already. I have even included a \dd command in my own utility package just to abbreviate the \mathrm d, but I wouldn't worry about spacing if there is some other visual separation. $\int xdx$ is barely readable, but to me $\int x\mathrm dx$ is good enough. Nov 3, 2014 at 14:48
• @AlexR There are zillions of ways how to do it in LaTeX, and zillions of people who like one and hate another one and ...
– yo'
Nov 3, 2014 at 15:39
• @AlexR: Whereas for me $\int x\mathrm dx$ requires significantly more effort to parse visually than $\int xdx$, though both are vastly inferior to $\int x\,dx$. Nov 3, 2014 at 22:39