3
$\begingroup$

Sometimes after posting an answer I want to edit it—for example to correct typos, word something better, or plug some gap.

Usually I just go ahead and edit, but I'm never quite sure what to do if the answer has already been accepted: since I might be changing it from what the questioner needed.

Example: in my answer to Need for rigorous meaning in English for definitions from my reference book I wrote

"The $A$ of all $B$ 's" is unambiguous as long as you know the mathematical meaning of an $A$ and a $B$. There is only one $A$ (because of the), and it contains every possible instance of a $B$.

Now, I don't like contains in the last sentence. I would prefer to write

There is only one $A$ (because of the), and every possible instance of a $B$ belongs to it.

to cover cases where "containing" isn't the correct relationship. But the questioner asked precisely about interpreting phrases in English definitions, so rephrasing one of my key points might affect its clarity for them.

In general, what is appropriate?


Edit: In the case of the linked post, I've now made the edit (I do feel the original wording was too restrictive, and in particular that this made the language point misleading.) But I've also noted that editing an answer bumps the post (I thought it didn't), so did a proper copy-edit while I was at it to try and avoid extra edits.

$\endgroup$
  • 3
    $\begingroup$ For the illustrated post, with a relatively minor edit, it is permissible to make such a change to your own post if you feel it makes even a minor improvement (regardless of whether it is an Accepted Answer). That said, if you plan on "polishing" a post with a number of such minor formatting or expositional changes, it is recommended to batch them up rather than submitting them one at a time. The latter approach would cause excessive "bumping" of the post in the opinion of many Readers. $\endgroup$ – hardmath Dec 23 '18 at 18:19
  • $\begingroup$ @hardmath Ah—I've never been sure what bumps the post and what doesn't. I knew adding a new answer has that effect, but didn't realise editing an existing one does. $\endgroup$ – timtfj Dec 23 '18 at 18:29
  • 2
    $\begingroup$ Do you think "x belongs to y" is any less ambiguous than "y contains x"? Just please don't make trivial edits. In this case, if you want to be unambiguous, then use "for all $B\in A$" or "for all $B \subset A$," whatever your intended meaning of your answer. If you mean neither of these things, then "A contains all B" vs. "all B belong to A" makes no difference, the use is still somewhat ambiguous in either case. $\endgroup$ – Namaste Dec 24 '18 at 0:32
  • $\begingroup$ @AmWhy The point was that questioner wants to know about definitions in general, no necessarily just definitions of things which relate to each other like sets. It's basically "How should I treat these phrases when they come up in definitions?" Contains wrongly restricts my answer to sets and their elements/subsets, and also misrepresents the English usage that he's asking about—since of doesn't necessarily imply inclusion. "Belongs to" is therefore much more accurate. $\endgroup$ – timtfj Dec 24 '18 at 15:44
  • $\begingroup$ Actually belongs to is too specific as well. Suppose someone defines the "ratio limit" of a sequence $(s_n)$ as $\lim_{n\rightarrow\infty} (\frac{s_n+1}{s_n}) $ then defines $\phi$ as "the ratio limit of all Fibonacci-like sequences". $\phi$ isn't a set of sequences, and saying that a sequence 'belongs to $\phi$" is decidedly odd as well. $\endgroup$ – timtfj Dec 24 '18 at 17:07
3
$\begingroup$

A post that needs to be edited, needs to be edited. A green tick doesn't change that.

If this makes it no longer the most useful answer, the asker can remove the tick or change it to the answer that does help them. But more likely, the edits only make the answer more useful, so the reason for the tick isn't affected.

Consider the other side of this, too. If the asker is most helped by an answer that is ambiguous or confusing to the point it needs to be fixed, is the question even reasonable in the context of Math SE? Nonsense answers only help nonsense questions, of anything, so the whole lot should be removed and the effort spent better elsewhere.

$\endgroup$
  • 1
    $\begingroup$ Also the edit history is there for all to see (if they know where to look), so one doesn't need to fear that the OP is being deprived of good information. $\endgroup$ – hardmath Dec 23 '18 at 22:46
  • 2
    $\begingroup$ And if a post doesn't need to be edited, or isn't improved by an edit (change in one term used), then the edit is a bad idea, because such an edit needlessly bumps the post. While this might be the goal of the editor, it is not ideal for this site. $\endgroup$ – Namaste Dec 24 '18 at 0:34
  • 1
    $\begingroup$ @amWhy The bumping is a bit of a nuisance to the one doing the editing, too—because we can't just base edits on a perceived need to edit but also have to consider the unwanted side effect. On my part, I had no idea that edits to an answer bumped a post. I thought the only things that did that were adding a new answer and editing the original question. (I've tried to check a few times after making an edit, and have never managed to find the bumped post, so I've assumed they weren't bumped.) I've noticed other posts being bumped, but never ones I've edited myself. It's puzzling. $\endgroup$ – timtfj Dec 24 '18 at 15:58
  • 1
    $\begingroup$ @hardmath Good point, and in this case I think the questioner had been here long enough to have found out how to that. (The reason I was cautious was that the questioner explicitly said English wasn't their first language, and was asking about language—so a change in my language might introduce a new problem in the form of another phrase in need of explanation!) $\endgroup$ – timtfj Dec 24 '18 at 16:04
  • $\begingroup$ @amWhy Any edits to answers bumps the post to at least the top of the list in the "Active" tab in the Questions page. $\endgroup$ – John Omielan Jan 25 at 23:38
  • $\begingroup$ @John, exactly, as I stated in my comment. Thanks for confirming. $\endgroup$ – Namaste Jan 25 at 23:40

You must log in to answer this question.

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