This isn't specific to SF, but it came up recently and it's worth a brief touch point. Recently in the review queue, I stumbled upon the following edit:


How do you troubleshoot Apache redirects in production? I know there's a a logging feature somewhere that lets you log all the redirects that happen...


How do you troubleshoot Apache redirects on a production server? I know there’s a a logging feature somewhere that lets you log all the redirects that happen…

Aside from changing "in production" to "on a production server", the nuance is easily missed: the ’ and … Unicode characters were swapped into the post, obliterating their inferior ASCII counterparts. The edit comment of Grammar & formatting. suggests that this was a choice of the user, not a word processor.

Should we be performing these types of edits?

  • 7
    If you're going to perform that kind of an edit, you should at least fix the repeated word word typo too. May 13, 2014 at 14:13
  • related ; Also note that some "helpful" browsers may edit the "..." into "…" for you when they sense 3 periods followed by a space.
    – voretaq7
    May 13, 2014 at 15:57
  • I gotta say that when I first saw this title, I assumed you were talking about the Unicode (ab)use post...
    – strugee
    May 22, 2014 at 4:33
  • @strugee why would we do that on mSF ?
    – user9517
    May 23, 2014 at 18:48
  • @Iain shrugs...
    – strugee
    May 24, 2014 at 1:49

5 Answers 5


No, they don't add value. In some cases, they detract and should be removed.

It's a more common problem on SO, but I do occasionally see it on SF: A post will have code or a shell script where the quotes have been replaced with these Unicode characters in the code, causing the code or script to break. These should be edited out and replaced with their ASCII counterparts in order to fix the broken code.

When they are used in the text, as opposed to in code, it's a simple matter of style. I don't care for them, personally, but I think it's too minor to edit them either way.

  • 1
    I have this problem with wikis that autoformat quotes into "proper" quotation marks, too.
    – Andrew
    May 14, 2014 at 1:16

This is precisely the definition of an edit which is too minor. If we accept that this adds value and should be approved, we must accept virtually anything. The edit doesn't materially improve the legibility of the post, much less the correctness, and requires working around the automatic filter to submit in most cases.

More importantly, it bumps the post when no substantive change has been made, and adds the editing person's name to the post when they contributed absolutely nothing.

Edits such as this are worse even than replacing commas with semicolons, and should always be denied as too minor.

It's also worth noting that the javascript editor used to write posts doesn't appear to autocorrect or otherwise facilitate the use of unicode quotation marks, and so virtually no posts are being made with them. If it was really beneficial and worthwhile, which it isn't, the right thing to do would be to request a feature, not go through editing every ASCII quotation mark on the site.


No, not really.

Yes, the Unicode representations of quotes and ellipses are “better” than ASCII in that they're closer to their literary counterparts, but we are not a publication. First and foremost, we are a knowledgebase focused on ease of editing. The ASCII character set has its roots firmly embedded into our keyboard punctuation and isn't going away any time soon. That's the punctuation that is here to stay.

We're professional IT people. I understand that we're pedants for detail; case in point, this meta topic. Some of you may be compulsively driven to “correct” that which is in front of you, but remember that it goes both ways. Some of us are just as driven to seek and destroy that which does not conform to a system. Your good-natured attempt at Unicode defiance will be assimilated back into the fold. Resistance is futile.

  • 2
    You didn't use Unicode quotation marks in your answer!
    – Andrew
    May 13, 2014 at 6:41
  • @Andrew Fortunately, we act as a redundant pair! (also, the irony loop is now closed)
    – Andrew B
    May 13, 2014 at 6:51
  • 1
    "better" is relative. If you're stuck on a crappy machine that doesn't support unicode having posts festooned with ?s or little boxes is certainly not "better". Neither is having to manually fix every quote in a script...
    – voretaq7
    May 13, 2014 at 15:56

ASCII is more portable than unicode, and for that reason I would consider it to be preferred, whenever it does not break the meaning of the text. Given that ASCII was designed to express English, it is hard to imagine an English sentence, which cannot be expressed in ASCII.

Replacing correctly used ASCII characters with unicode in my opinion removes value. Replacing unicode with ASCII is a too minor edit, that I would consider it relevant on its own.

In places where the exact character used does make a difference, such as in snippets of code, configuration files, logfiles etc. I think it is relevant to correct mistakes, but I would expect that kind of mistake to be rare. You don't copy-paste a piece of code and accidentally change ASCII characters into unicode.


Yes, it adds value: it makes the post easier to read.

Does it add enough value to be a worthwhile edit in and of itself? No. And I don’t see any room for debate on that point. It absolutely definitely is not worth doing on its own.

However, if you’re editing a post anyway, and feel like also curling the apostrophes, I see no reason why you shouldn’t. I generally do.

  • 1
    I tend to feel the same about removing signatures, actually. To my mind, that’s also worth doing only if you’re also doing something else. I think that used to be Jeff Atwood’s policy too, but he later changed his mind and decided that signature removal was valid on its own. I think.
    – TRiG
    May 13, 2014 at 9:42
  • 9
    I dispute your it makes a post easier to read. Without peering at the edited text quite hard it's difficult to tell the difference. If when reading the text normally a difference isn't perceived then the two texts are the same so your statement cannot be true.
    – user9517
    May 13, 2014 at 12:16
  • That may depend on your font settings, @Iain. The general consensus among people who care about typography is that curled apostrophes and quotation marks are easier to read.
    – TRiG
    May 13, 2014 at 12:21
  • 1
    Perception is the only reality.
    – user9517
    May 13, 2014 at 12:22
  • 5
    The general consensus among people who care about typography is that curled apostrophes and quotation marks are easier to read and in general I agree with them totally. In the specific case of websites that display technical content that can be badly affected by someone 'correcting' quotation marks, etc. I think there's an adequate case for just say no because that's easier to enforce than anything more complex and nuanced.
    – Rob Moir
    May 14, 2014 at 16:58
  • I don't see how "do it in text, but not in code" is particularly nuanced, @RobM. Seems fairly straightforward to me.
    – TRiG
    May 26, 2014 at 11:05
  • And to me too, because like the majority of people here, we're sensible. But in any large website group there's a minority of idiots that need to be catered to as well. You can already see on any of the stack exchange sites where a group of people are prepared to invest 10 times as much energy into why something should be allowed to remain broken than they are into just fixing it.
    – Rob Moir
    May 28, 2014 at 9:17

You must log in to answer this question.

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