6

Given that

We now automatically remove salutations from posts as they are entered. -- Jeff Atwood♦

what else would we like to automatically remove (or correct) as it is entered?

Any suggestions?

One thing which currently irritates me is the misuse of parentheses

1
  • 2
    Looks like title says it all
    – random Mod
    Commented Jun 7, 2011 at 14:45

13 Answers 13

15
tks

thanks in advance

??????

plz help

I specially find the multiple question marks annoying.

EDIT

Wow, I'm off the site for an hour and this happens! :)

Just to be clear, I'm talking about the text that appears at the end of the question. This applies to all the suggestions I made (this is also the reason I left the ... out of the suggestions because it appears in too many places).

Most of my experience revolves around the PHP and JS tags and this kind of signing off is fairly common.

I totally agree that it's better to err on the safe side when auto filtering as the community can always remove the ones that slipped through the filter. The filter should only apply to the most flagrant violations and not for anything even appearing borderline.

13
  • 5
    +1. The multiple question marks, and the multiple dots for that matter is annoying the bajeebers out of me.
    – Mia Clarke
    Commented Jun 7, 2011 at 14:02
  • +1 for the question mark... It's trivial to remove multiple question marks from the end of the question title, and only leave one.
    – fretje
    Commented Jun 7, 2011 at 14:03
  • 3
    can we add multiple commas to that as well?
    – tombull89
    Commented Jun 7, 2011 at 14:25
  • 2
    What about questions regarding the ?? operator in C#?
    – hammar
    Commented Jun 7, 2011 at 14:29
  • @hammar: is there one where that operator is actually at the end op the question title?
    – fretje
    Commented Jun 7, 2011 at 14:30
  • @fretje: I guess that case might be safe, but my point is that in programming we frequently use symbols and combinations of letters that while not considered proper english are perfectly OK in some programming language. Any kind of automatic transformations done to questions will have to be written with extreme care to avoid breaking code. Limiting such transformations to avoid code sections won't work, as people consistently forget using them; adding missing code sections is probably the most common edit.
    – hammar
    Commented Jun 7, 2011 at 14:37
  • @hammar I guess the "salutations remover" will skip code samples, blockquotes and hyperlinks.
    – yms
    Commented Jun 7, 2011 at 14:46
  • @yms: Read the last part of my comment again. People forget using those properly all the time.
    – hammar
    Commented Jun 7, 2011 at 14:52
  • @hammar yes, you are right about that...
    – yms
    Commented Jun 7, 2011 at 14:54
  • @hammar: I agree. I was just pointing out the "question mark at the end of the title" case. I agree that other stuff shouldn't be removed automatically. Maybe this answer should be split up in two different ones, as I'm guessing most of the up-votes are for the question mark thing.
    – fretje
    Commented Jun 7, 2011 at 14:55
  • @fretje: Just make it clear that you're only talking about removing multiple question marks at the end of a question title.
    – hammar
    Commented Jun 7, 2011 at 14:57
  • 4
    To be clear, I think content should NEVER be removed automatically, unless it's really really clear what it is. For instance salutations at the beginning of the body, which are already automatically removed, and multiple question marks at the end of the title. And I'm sure someone is clever enough to come up with an algorithm to define a signature at the end of the body. Much more than that would be overkill I think.
    – fretje
    Commented Jun 7, 2011 at 14:59
  • @hammar @fretje answer updated with more details on what I actually meant.
    – JohnP
    Commented Jun 7, 2011 at 16:00
15

I don't think that we should be blindly stripping content from posts.

I'd like to use the stance of 'Innocent until proven guilty.' I don't think that we should assume that specific words can be removed from posts. I think it's a bad idea.

4
  • 5
    +1. I agree that this must be done with the utmost finesse if it is to happen. I believe that substitution is better than deletion in this case.
    – Mia Clarke
    Commented Jun 7, 2011 at 14:10
  • 3
    or at least some sort of indication that text was removed. People won't fix their habits until they realize their content is being removed
    – j_syk
    Commented Jun 7, 2011 at 14:16
  • 4
    I think it is somehow safe if they are at the very begining or the very end of a question, and I think the user should be notified that those words have been removed, otherwise it will be very confusing, even for experienced SO users.
    – yms
    Commented Jun 7, 2011 at 14:20
  • 1
    Currently, most of these suggestions are part of the bad-question filter anyway.
    – mmyers Mod
    Commented Jun 7, 2011 at 16:10
7

I think that removing salutations is probably enough. If the site tries to get too clever, the rate of false positives is likely to go up.

On an entirely lighter (and only semi-serious) note, I'd like the following automatically removed/improved:

  • i > I IIF it's not contained within a codeblock
  • teh > the
  • gimme teh codez > I'm too bone idle to even try, please ban hammer me
  • thx > thanks

Anything that auto-removed "txt spk" would be a blessing (but also a curse as it'd hide the fact that the OP is just plain lazy)! :)

2
  • 2
    Your last sentence says it all...
    – Benjol
    Commented Jun 7, 2011 at 14:02
  • Personally, I would much prefer posts containing "txt spk", "1337 $þ33|<" just not make it through the quality filters and be rejected (if it can be reliably detected). We can't catch all forms of possible words so if someone was really trying hard, we might miss a few. Commented Jun 7, 2011 at 20:52
6

It's urgent!

could also be removed. Specially if it is written in capital letters.

1
  • 5
    Au contraire. That text is an excellent marker of an overall crappy question/answer in need of editing or nuking Commented Jun 7, 2011 at 16:08
6

One thing that might benefit from some auto-correction is the habit of some users (it must be a rule somewhere in the world, but I've never managed to figure out where) to start paragraphs with four spaces, leading to code indentation:

I have a problem ......

I imagine it would be difficult to implement, though - one would have to find out whether the first paragraph is code, or plain text.

1
  • 2
    In addition to the spaces, I also strip "I have a problem" from posts when I see it.
    – Bill the Lizard Mod
    Commented Jun 7, 2011 at 14:06
5

Things that I'd like to be flagged about for editing, but not necessarily removed:

Posts containing the following words:

  • noob
  • newbie
  • guys
  • ...
  • :)
  • :-)
  • :-(
  • (:
  • (-:
  • )-:
  • :/
  • :-/
  • plz
  • thx
  • More than one ? (E.g. ?????)

And yes, I'm being serious. I've edited enough questions to see that when these words are involved, the question generally needs editing.

Something I was working on was an API app that would rank questions by the number of misspellings they have, so I'd be able to go in and rank the questions to edit. I really need to dust that off and finish it. It'd be helpful.

6
  • 9
    I think that smilies should be allowed in posts.
    – jjnguy
    Commented Jun 7, 2011 at 14:03
  • 2
    @jjnguy didn't say they weren't; just that the smilies have been indicative of a post that needs further editing. They're a herring, though not necessarily a red one. Commented Jun 7, 2011 at 14:06
  • 3
    Removing smilies would for example break this Ruby code: [2, 1].inject(:-). IMO it's too hard to do this automatically without breaking code examples (we can't count on them always being in code segments), and might go unnoticed when pasting large code blocks.
    – hammar
    Commented Jun 7, 2011 at 14:51
  • 3
    @Hammar @jjnguy: Let me say this more clearly. I said, not necessarily removed in my post. There is a correlation between smilies and a post that needs to be edited for other reasons. Did you read my post? Commented Jun 7, 2011 at 15:42
  • How about adding "Greetz" to that list? Damned if I know what it means.
    – razlebe
    Commented Nov 9, 2011 at 11:52
  • Need to be careful if including the ellipsis bot to flag a lot of questions concerning variadic functions in C or C++
    – razlebe
    Commented Nov 9, 2011 at 11:55
4

Not removing, per se, but I'd like it if things like plz, im, thanx and codez were automatically replaced by the correct words please, I'm, thanks and code.

1
  • 1
    See the end of Rob's answer: I'd prefer these to be left in, as they can be useful markers of bad question quality: i.e. one that wants nuking more than babying.
    – Benjol
    Commented Jun 7, 2011 at 19:37
4

An other nice idea would be to convert complete uppercase posts to lowercase.

2
  • Do those make it past the quality filter in the first place? Commented Jun 7, 2011 at 14:28
  • @Daniel DiPaolo, I still see them once in a while. So probably most don''t pass. Commented Jun 7, 2011 at 14:49
2

If the last line contains not much more than "thanks" (or any of its derivatives, tnx, thanx, ...) and the users name, I'd say it's save to remove that line.

2

This doesn't add anything to the question:

Sorry for my English, but [some reason]

Being sorry doesn't help if we cannot understand a word. As my mom told me: don't be sorry, be cautious.

And this is another one (with many variation):

[Any] Answers [and comments] are welcome

Of course they are welcome, it's the purpose of asking a question on this site.

1
  • 2
    @Pierre.Vriens Do you genuinely feel that changing "don't" to "doesn't" on 6 year old questions with virtually no views about obsolete topics really helps anybody? How are you even finding these? Please look at the bigger picture before editing.
    – Jason C
    Commented May 20, 2017 at 0:56
1

I think caution is warranted when automagically removing text from question, lest the context is accidentally changed and the well-meaning are disaffected by our obsessive auto-correcting of our collective pet hates.

That said, as long as we're collating a list of pet hates, I'd like to nominate the word "So" when it appears as the very first word in the question body, as in:

So I have some code that does this and that, and the problem is that...

Damn, that's irritating.

0

I do wonder whether any first line which is less than (say) 20 characters shouldn't just be zapped.

And even if it isn't markdown-conformant, I really think that \ns should get turned into </BR>s, or whatever. It would save a lot of formatting time...

0

codes

Easy shibboleth. 6chars

You must log in to answer this question.

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