# Should we write every constant upright?

This post comes out because some of my edit was rejected.

I honestly think we should use upright letter for specific mathematical constant.

For example I think $$\operatorname i$$ is the immaginary unit and $$i$$ is current density, hence I would write the famous Euler's formula as

$$\operatorname e ^{\operatorname i \text{π} } +1 = 0$$

I think π also should be written in upright font, because it represent the trascendental number $$3.1415\dots$$, instead $$\pi$$ represents the geometrical quantity : $$\pi=$$ π $$\operatorname{rad}$$; therefore I would prefer to write $$\sin(2$$π$$x)$$ instead of $$\sin(2 \pi x)$$.

Moreover the charge of electron $$e$$ should be distinguished from the base of natural logarithms $$\operatorname e$$.

So my question are:

1. Is it a way to write upright Greek character without copy-pasting them from an unicode table?

P.S. : sorry for my english, it is not my native language.

EDIT: I am not a "MathJax nazi", I just wanted to know if there was a standard convention in this site, since some of my edit was revised or reject many times. I put the "feature-request" tag, because I did not find a clear way to write $$π$$ in math mode (the \text command seems to not work)

• You specifically used \operatorname{e}, e.g.. $e, i$ are not operators, they are constants, just as $2\in \mathbb Z$ is a constant. Jun 8, 2020 at 11:38
• You can do that in your own posts, but don't force your stylistic preference upon others. I find the usual way to write things (that is, not upright) looks far better. Jun 8, 2020 at 11:38
• @amWhy Actually they are not operator, but operator too should be written upright, for example $\sin$ and not $sin$. I could just use \mathrm instead. Jun 8, 2020 at 11:40
• You were suggesting unnecessary edits to another's post, not your own post. Jun 8, 2020 at 12:03
• RE: "If I see $\int ln(1+e^x) dx$, I edit $\int \ln(1 + \operatorname e^ x) \, \operatorname d x$. Is this unnecessary?". Yes. I'd say even disruptive if done in others' posts.
– Aloizio Macedo Mod
Jun 8, 2020 at 12:25
• @T.S one usage is appropriate the other is inappropriate. The imaginary unit is certainly not an operator.
– quid Mod
Jun 8, 2020 at 12:48
• Nor was \int \ln(1 + \operatorname e^ x) \, \operatorname d x rendering as $\int \ln(1 + \operatorname e^ x) \, \operatorname d x$ an appropriate use of \operatorname. Don't use it on $e$, $i$, or $dx$. Jun 8, 2020 at 12:51
• "I think the imaginary unit can be represented" alright I take the "certainly" back. This is not common though, and especially in the edit you actually proposed it hardly semantically represents that. Further you also use it for $e$. Now maybe you'll tell me that this is also an operator since after all $e^x$ is the value of the exponential function at $x$, but then one would not denote it like this. This notation is most naturally read interpreted as the real number $e$ to the power $x$. Likely it does not change all that much there but it is still not a good idea.
– quid Mod
Jun 8, 2020 at 12:56
• This is tagged (feature-request), but it is not clear what feature (if any) should be implemented. Also did you mean to write "reach an agreement" rather than "enreach an agreement"? Jun 8, 2020 at 13:06
• Regarding the "feature" there is a LaTeX package "uprightgreek" but the same does not seem to be available for MathJax at least not here.
– quid Mod
Jun 8, 2020 at 13:10
• A agree your edit was properly rejected. Do not merely change the conventions of one country into the conventions of another country. Similarly, do not change "3.14159" into "3,14159" or vice versa. Jun 8, 2020 at 13:44
• Others have pointed out that \operatorname is wrong, but no one else seems to have explicitly pointed out that the spacing in \operatorname{d} x is quite wrong, e.g. $\int\operatorname{d} x$. If one is going to use an upright "d" for the differential (which, frankly, is what I do), then the correct syntax is likely \mathrm{d} x, e.g. $\int \mathrm{d}x$. I would also argue that $$f : \mathbb{R} \to \mathbb{R}$$ looks better than $$f \colon \mathbb{R} \to \mathbb{R},$$ particularly if one wants to give a mapping in the same line: ...
– Xander Henderson Mod
Jun 8, 2020 at 14:40
• ... For example, $$f : \mathbb{R} \to \mathbb{R} : x \mapsto x^2$$ vs $$f \colon \mathbb{R} \to \mathbb{R} \colon x \mapsto x^2.$$ Of course, this is a matter of style, and each user should write with the style they feel is appropriate. What is inappropriate is editing other's questions with the sole intention of changing their style to match your preferences.
– Xander Henderson Mod
Jun 8, 2020 at 14:41
• @Xander Okay, seeing your edited comment: I think its important to remember that some conventions are standard, for example $\sin$ vs $\sin$, so (depending on context), such edits are good. (Similar discussions on editing formatting have occurred before, in particular this highly-voted question is about editing in $dx$ vs $\mathrm{d}x$.) Jun 9, 2020 at 13:53
• Visually the default style of upright operators and variables in italics looks quite pleasant. What I can't stand really is operators not written upright. Often this is because of lack of starting backslash and I usually fix such typos when the post is interesting.
– Paramanand Singh Mod
Jun 10, 2020 at 6:45

While I think there is some "standard" (ISO maybe) that recommends this, in my experience this is rather rarely done in mathematics, and (or therefore) many will find this usage rather unusual.

If you prefer this style I will not edit your posts and would tell others to leave your posts alone, but please do not change this in others' posts, not even when you edit them anyway and definitely not as the only change.

Also do not use \operatorname for things that are not operators, as it will give the wrong spacing.

This is for the upright letters. For the example in the comment I do consider an edit from $$ln$$ to $$\ln$$ as appropriate (though if it is the only change, I recommend not to suggest an edit or edit an old post). But that is not the same thing; the difference is that in $\ln$, $\sin$, $\log$, typeset as $$\ln$$, $$\sin$$, $$\log$$, the letters form one "unit", a word if you like, while $sin$, typeset as $$sin$$, are three separate letters.

Let me add one more general remark, which is somewhat orthogonal to the rest: whatever one might think is best, this site is not a good place to indulge in finer details of typography. In a context where there are potentially many different people editing and some users do not know MathJax overly well, there is some intrinsic value in keeping the source as simple as possible. Thus, in cases of doubt and arguably even in the case of minor improvements, I strongly recommend to err on the site of keeping the source simple.

To answer the added question what the standard of the site is explicitly.

There is not much of a "house style", yet standard-etiquette is the author of the post decides and these choices are then accepted (within reasonable but generous limits).

• Thank you for your words. I really like interested in typographycal convenction. For example, in my Master thesis I writed every vector, matrix in italic bold (e.g. $\pmb v$, $\pmb A$), set in boldface ($\mathbf I$) and tensor in sanserif ($\mathsf T$). Maybe I am a bit "maniacal" about this ^^' Jun 8, 2020 at 13:32
• @SewerKeeper In the early drafts of my master's thesis, every vector was bold. My advisor objected, claiming that the vectors were too heavy on the page. As I needed his signature, I changed all of the vectors to a plain typeface (though I covertly put all the Oxford commas back in right before submission... bwahahahah!). Looking back on that document many years later, I am glad that I followed his advice. In retrospect the, the bolded vectors were quite ugly.
– Xander Henderson Mod
Jun 8, 2020 at 14:47
• @SewerKeeper thank your for the reply. I agree it is a fascinating subject. Maybe you find content that interest you about it on TeX - LaTeX or Graphic Design too.
– quid Mod
Jun 8, 2020 at 16:23

This is mostly a matter of style. See this Wikipedia article on typographical conventions in mathematical formulae. In particular,

The rules of mathematical typography differ from country to country; thus, American mathematical journals and books will tend to use slightly different conventions from those of European journals.

If a post is already well written in one acceptable style, it is not necessary to enforce another one. For instance, I find the following two both OK: $$\int_a^b\sin (x)\;dx,\quad \int_a^b\sin(x)\;\mathrm{d}x$$

Of course, rejecting your (significant) improvements in a post simply because your own style would be nonsense. If you are

• fixing MathJax in other's posts,

you can certainly use your own style. I don't find it reasonable to reject your edits; minor edits like only one or two instances of ln(x) to \ln(x) will be rejected automatically already by the system.

For font styles one can use in math mode see this question. In particular, \mathrm is the normal upright Roman font. You may also use \operatorname. 1 Of course, MathJax is not TeX; it may only support a subset of those.

1 There is some subtle difference between these two commands; see What's the difference between \mathrm and \operatorname?

• Thank you! I didn't know the difference between them. Jun 8, 2020 at 12:54
• While some specific rejects can be objectionable, the general statement that "If you are (...) fixing MathJax in other's posts, you can certainly use your own style." is problematic. If I have a post consisting of, among other things, seven $\int fd\mu$ and in one of them I write \innt f d\mu, fixing it to \int f d\mu and then changing all instances to $\int f \operatorname{d}\mu$ can be reasonably considered disruptive. If you change your post to "(...) you can certainly use your own style if it doesn't conflict with the one already in place", then it seems to be a good guideline.
– Aloizio Macedo Mod
Jun 8, 2020 at 13:31
• This seems reasonable, Thank you Aloizio. Jun 8, 2020 at 13:39