Can (should) we get rid of multiple punctuation marks?

Question

Disclaimer: I admit that this is a personal pet peeve of mine, so feel free if to close this with extreme prejudice if I'm alone in this.

Some people have a tendency to end their sentences with an over-abundance of punctuation marks. The most common offences to this happens with exclamation marks and questions marks and more often than not it's in the title.

And opinion on this aligns closely with that of Sir Terry Pratchett. But even if you don't subscribe to that particular train of thought, I think it's clear that this construct doesn't really add anything to the clarity or expressiveness of the question.

Can we please get rid of them? I think it would be fairly easy to reduce any case where multiple question/exclamation marks are at the end of the title to a single instance.

Ideally we could also do this in the body (or inside the title), but obviously this is a lot more dangerous (e.g. "What does the !! operator do in C" should not be reduced to "What does the ! operator do in C", even if !! is not an operator in C).

What do you think?

Note that according to this query, there are 427 questions that end in "!!", 235 questions ending in "!!!", 28 questions ending in "!!!!" and 9 questions ending in 5 or more exclamation marks. There are also 4331 questions ending in "??", 1546 ending in "???", 221 ending in "????" and 61 ending in 5 or more question marks (note that the query is not perfect, as the numbers for 2 to 4 punctuations marks also include the higher counts, I'm too lazy to write the correct query).

As far as I know, this query also only finds questions that have not been edited to remove this. I guess that there are a lot more such questions posted than that, but most of them will be edited.

Answer 1

Great idea, I totally support this, making it so:

--- make sure all question titles end with at most one question mark
update
Posts
set Title = dbo.RegexReplace(Title, '\?{2,}$', '?')
where 
PostTypeId = 1 and 
Title like '%??'

7315 rows affected

--- make sure all question titles end with at most one exclamation point
update
Posts
set Title = dbo.RegexReplace(Title, '\!{2,}$', '!')
where 
PostTypeId = 1 and 
Title like '%!!'

1005 rows affected

--- make sure no question titles end with any number of periods
update
Posts
set Title = dbo.RegexReplace(Title, '\.+$', '')
where 
PostTypeId = 1 and 
Title like '%.'

89696 rows affected

--- make sure no question titles have extraneous spaces before the ?
update
Posts
set Title = dbo.RegexReplace(Title, '\s+\?$', '?')
where 
PostTypeId = 1 and 
Title like '% ?'

45201 rows affected

--- OK now you're just pissing me off
update
Posts
set Title = dbo.RegexReplace(Title, '[?!]+$', '?')
where 
PostTypeId = 1 and 
(Title like '%!?' or Title like '%?!')

~1500 rows affected

Also run for SO, SU, SF and the top ~15 SE 2.0 sites. It is also now enforced as a silent, automatic edit at time of question ask.

Answer 2

Yes, I think this is annoying as well when it is overused. But I can imagine certain cases where two (but not more than two) exclamation marks would be acceptable to truly emphasize something. I've done that before, and I wouldn't want them removed.

And in addition, as you point out, it would potentially cause problems for operators in certain programming languages.

I think this is another example of the more general case that edit privileges and suggested edits were designed to solve. We really don't need to handle this with a regex.

Titles are pretty visible, and easily corrected by other users. This problem seems like it solves itself. Automatically replacing problematic grammatical constructs is a slippery-slope, and not one that I'm convinced is worth going down.

Answer 3

I correct most double+ punctuation when I'm editing for other problems, though I usually wouldn't edit just for that. I do think that all the cleanup of these should just be done manually, it's way too hard to automate all the special cases on a programming site.

The one I also hate is the extra space in front of the question mark.

What's up with that ?

Answer 4

The default position of the system should be to leave content as its author intended. If a human being decides to alter text, that is acceptable per the terms of using this site (ironically that item in the FAQ includes "?!"). However, that doesn't mean we should whimsically write scripts to alter content. The great power of automatically editing content should not be used lightly for trivial stylistic issues but saved for more egregious issues (if there are any).

People really, really don't like when they create content and then the computer changes it to something else. Look at the reaction the @lert removals caused. There shouldn't be automated editing of human-generated content without a really good reason, and your personal stylistic preferences don't meet that standard.

Answer 5

This would be best if it performed the correct substitutions instead:

• Given any string of ! and ? outside a <code> block, reduce the length to at most two punctuation marks, retaining one of each if both are present.

• Replace double exclamation points "!!" with the proper U+203C DOUBLE EXCLAMATION MARK: ‼

• Replace double question marks "??" with the proper U+2047 DOUBLE QUESTION MARK: ⁇

• Replace both "?!" and "!?" with the proper U+203D INTERROBANG: ‽

Precision is, of course, an important quality in a programmer, so I think it's clear that using the correct punctuation would show a certain attention to detail and make the whole site feel more professional.

The current situation, with people just littering punctuation anywhere and everywhere, is dreadful. Four exclamation points in a row? Who would do such a thing‽ It's barbaric, I tell you.

---

Sigh. I guess I'm the only one who finds ridiculous unicode characters inherently funny. Oh well. Yes, for the record, the above is a bad idea. Making fiddly changes to posts to correct stylistic issues is also a bad idea. There's any number of minor tweaks that would be easy to add, but other than stuff like stripping out the obvious salutations it's just going to confuse people when their text gets altered out of nowhere, and false positives will create massive headaches

Think about how the @lert-stripping in comments collides with some actual programming syntax, forcing people to use backticks in comments when they otherwise wouldn't need to. While it's well and good to say that code should be in code blocks or backticks, do we really want to require that in random cases just because it looks like poor style? If I'm talking about nullable types in C#, would it make sense that I could say 1 + 1 = 2 directly, but not num ?? 0 without an extra hurdle?

Interrobangs are still pretty cool, though.