I recently rejected some edits to previous posts of mine that used Gb, F# notation where the edit was just replacing these notations with G♭ and F♯, using Unicode code points U+266D–U+266F.

Personally I don't see a point to this — the unicode chars are harder to type and leave more room for error on the rendering side, so I think we should leave it up to the author and seek internal consistency. I rejected the edits as not substantive.

Not wearing the mod hat anymore, I'm not too worried about this, but seeing that the proposed edit was from an enthusiastic new user, I think we should have some guidance on the books: should we proactively be mopping up old posts to use the new unicode glyphs, a la Wikipedia? Yea or nay?

5 Answers 5


I have accepted these sorts of edits before, on the grounds that the actual music notation symbols look better than # and b, and thus are improvements to the original post. But....

Your comment that use of the Unicode code points may "leave more room for error on the rendering side" has caused me to think again about this. I use the Unicode characters all the time, but I have seen some users complain in the comments that the Unicode characters aren't rendering on their systems. Further, I have noticed that the Unicode characters seem to render as smaller characters, which are often not as legible as their simpler Latin alphabet approximations (especially on smaller screens). E.g. # vs ♯, b vs ♭, x vs 𝄪, bb vs 𝄫. I don't know how to best communicate a natural with a Latin character, but ♯ and ♮ can be difficult to distinguish, so it may be best to use # for sharp and ♮ for natural.

All of this to say that, upon reflection, I don't think that changing these symbols to their Unicode counterparts improves the clarity of a post, and it may even reduce clarity. I won't be accepting these edits any longer, and I may even go back to using Latin characters, except for naturals, in the interest of improved clarity and access.

  • 2
    My go-to for natural has always been Bnat. Anyhow my main issue is that I don’t think we can expect the majority of writers of posts to be using these Unicode characters when the usual ascii chars are right there on the keyboard within easy reach. Maybe if there is special support added to the editor in the future it would be more reasonable to expect, but not until then.
    – NReilingh
    Aug 29, 2018 at 3:16
  • 2
    Well, it's not just the platform supporting unicode, it's that the typefaces present on the machine need to include the glyph! Case in point, the double sharp/double flat symbols in your post render just fine in Safari on macOS, but don't render in Safari on iOS. Same platform, but the font on the mobile device doesn't have those symbols.
    – NReilingh
    Aug 29, 2018 at 3:29
  • 3
    I note with slight ironic amusement that the double sharp and double flat unicode characters in this answer do not render on the Windows computer I'm currently using. Aug 31, 2018 at 15:35
  • 1
    Case in point, I'm also supportive of the improved looking Unicode versions, however, in this very answer, two of the characters rendered simply as squares on my browser. I've noticed more rendering problems of late as Apple, Microsoft, Google, and others all seem to be doing their own thing.
    – JYelton
    May 2, 2020 at 0:10
  • 1
    @JYelton -- When I originally wrote this answer, I used the Unicode code points because they looked nice; but after writing the answer and thinking about it, I quit doing that (although I have not written any answers here in a while, anyway....) If the goal of the site is to be as inclusive as possible, I think that Unicode is a no-go here when it offers no real benefit over #, x, b, bb, nat, etc. We aren't distributing pdfs, after all, we are trying make accurate information accessible.
    – user39614
    May 2, 2020 at 0:25
  • It's unfortunate, but I agree. I usually (on Electronics SE) favor the correct symbols for units (µ vs u, for example) but if there's a good possibility that it won't render at all, then I suddenly think the approximations are better. Damn those rendering engines!
    – JYelton
    May 2, 2020 at 0:27
  • @JYelton -- I don't know this for a fact, but I imagine that Greek letters are better-supported than musical symbols in many fonts. This seems like it would be a good question for a web development guru. Have you ever noticed an issue with Greek symbols on Electronics SE? Or seen any complaints about rendering, or questions on Meta there?
    – user39614
    May 2, 2020 at 0:44
  • 1
    I haven't noticed or seen complaints about them; you bring up a good point. I'll be more on the outlook for them though, and ultimately more careful about edits involving them.
    – JYelton
    May 2, 2020 at 0:48

I'd have to agree that an edit like that is non-substantive and should be rejected.

I get the utility of unicode chars, and if an edit was fixing numerous other problems, then sure, using the correct symbols for sharp and flat would be fine, but the b and # work well enough.

In the case of a new user, I think just reiterative the requirement for substantive change should be enough.


Yes, I agree that simply changing the notation of accidentals is not sufficient grounds for an edit. As long as one method of typing is not clearly superior or something, I think any form of clear expression of accidentals should be okay. Similarly, edits for spelling/grammar issues, in my opinion, should not be accepted unless they are so egregious as to obscure comprehension.


Merely changing the characters for sharps flats and naturals doesn't harm the question or answer. It possibly improves the look, but that's about all. It seems one contributor feels duty bound to 'improve by editing', and it costs the rest of us nothing. There are more pressing problems around.

  • I agree, but I also wonder if this user has created a script that automates this process.
    – Richard Mod
    Apr 27, 2020 at 18:12
  • 1
    @Richard - if he has, is he keeping it a secret, or could he share it..?
    – Tim
    Apr 27, 2020 at 18:26
  • Just based on that user's propensity for these edits both here and elsewhere, I'm guessing it's either a script or a team of users using a single account. But for all that we'd have to ask that user, I think!
    – Richard Mod
    Apr 27, 2020 at 18:27
  • @Tim -- but there is a cost; the unicode musical characters may render as gobbledygook on some systems. I have had complaints about the Unicode flats and sharps not rendering on phones in some instances; in a comment above, ToddWilcox noted that some of the symbols used in my answer here did not render on the Windows system he was using at that very moment. The cost is accessibility, and for little gain.
    – user39614
    Apr 27, 2020 at 19:21
  • Please don’t confuse different characters with different fonts. The distinction is highly significant for accessibility purposes.
    – NReilingh
    Apr 29, 2020 at 1:49
  • @NReilingh - thanks for that. I (obviously) did confuse the two. Although I need someone to explain the difference between with regard to a sharp sign. Please!
    – Tim
    Apr 29, 2020 at 6:19
  • @Richard this user is pretty active across all sites and is a very prolific editor and a mod on another SE. I would not be surprised if he uses the data exporter for this. On the whole most of his edits are pretty good, some of the just accented edits still trickle though but those are definitely in the minority and I reject them when I see them.
    – Dom Mod
    Apr 29, 2020 at 15:25
  • @Dom My thoughts exactly.
    – Richard Mod
    Apr 29, 2020 at 15:28
  • @Richard - I could do with access to a faster script than '♯', but maybe I'm stuck with that on old-fashioned Windows 10. Maybe do it all wrong and wait for an edit to manifest itself?!
    – Tim
    Apr 29, 2020 at 15:31
  • @Tim I've treated accidentals in at least four ways: Bb and F#, Bf and Fs (which I admit could be confusing), B-flat and F-sharp, and B♭ and F♯ (which don't render in chat). I've never had any problems with people understanding me, so I think you'll be fine either way!
    – Richard Mod
    Apr 29, 2020 at 15:36
  • @Richard - we never have trouble understanding you! I'm fine with any old # and b, it's nat that gets me!
    – Tim
    Apr 29, 2020 at 15:42
  • Heh, I just stumbled upon this thread. I have indeed edited quite a couple of posts for this reason, but mostly because I encountered them in the First Posts review queue and (on old posts) always if the Q&A was already active recently, as not to cause needless bumps to the front page. Since I now have 2k reputation, I'm not 'wasting' reviewers' time anymore. (FWIW, I do use the data explorer, the API and regular search for other types of edits on older posts, such as repairing broken images.)
    – Glorfindel
    May 8, 2020 at 17:24

Another reason not to accept these edits (or perhaps even use the symbols at all) is that it can radically change search results.

Try individual searches for C#, 'C&sharp', and C♯.

As of this post, I got 2,941, 1,809, and 11,747 results, respectively. That suggests a search including one representation will miss (or erroneously include) posts using a different one.

Possibly we could make a feature request that # and '&sharp' (and b and '&flat') should be equivalent in searches, but I wonder if that's technically feasible.

You must log in to answer this question.

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