3

Both questions and answers often need to be edited. In this case there seem to be three different behaviors:

  • Editing the question/answer to make it look like the new version was here from the beginning;
  • Editing the question/answer to make it look like the new version was here from the beginning and leaving a comment what was changed;
  • Clearly marking in the text what was edited.

What is the preferred behavior and why? Should this be included into the faq?

(I open this question, because of the discussion that started in the "Etiquette on more than one correct answer". In particular Juan's and Andrew's answers already express opinions on this topic.)

7

The editing history is always available to someone who looks for it (by clicking on the "edited [time]" at the end of an edited answer), so your question is: Should we make the edit history explicit in the answer itself?

My answer is No. Simply folding the edits appropriately into the answer makes for less clutter, and a better, concise answer. This is useful to everyone, both the original questioner and those who arrive at the question in the long term, and (to borrow Juan's phrase) "people just don't want to read the whole saga of how the answer was improved over time". (Disclaimer: This here is a remark about hypothetical excesses of historiography, not about the fine answer that provoked the previous thread.)

However, in order to be fair, it's good to briefly mention your edit whenever (1) you borrow or steal from another answer — so that it doesn't seem the other answer was redundant, and to give credit, or when (2) you edit the question to invalidate one of the answers previously posted — so that their answer doesn't seem unduly inapt. Most edits are not of this form, though, and I feel that when they are, a brief but clear mention of the change suffices.

Andrew suggests it's better to preserve edit history by marking off edits separately, "since that person knows what I originally wrote and is only interested in seeing what has been added". This is a nice idea, but I'm unconvinced. Whom does it help? Someone who

  • reads the answer in the interval after it was originally posted and before it was edited, and
  • visits the thread again after the edit, and looks at this particular answer, and
  • remembers what was written and doesn't want to reread it, but
  • knows to look for "Edit" to see what's new.

Not very frequent, in my view. Certainly, you must mark edits when you make a major change, something that may easily be missed by a reader of the previous text and which is substantial and important enough that a previous reader's attention ought to be drawn to the new material with a bold "Edit" (e.g. you remove something bad, or when your original answer was so long that it won't be reread unless you draw attention). But in most cases, this is not necessary, and even when this happens it's good to eventually go back and clean up the answer to remove 'history clutter'. (Also, when the changes introduced in the edits are logically distinct from the previous, it's better to eventually use descriptive titles like "Explanation" or "Caveat" rather than "Edit" and "Second Edit".)

  • Note that although my answer ("No") may seem diametrically opposite to Caramdir's ("Yes"), we seem to agree that major edits must be marked, and (hopefully) that when there's no good reason to keep them separate, the edits may just be merged in. Apart from some minor quibbles about what's "major" and how frequent its occurrence, I guess both approaches boil down to the same thing, and are adequately covered by the simpler "Use common sense". – ShreevatsaR Aug 1 '10 at 22:28
2

First, I must say that do not have much experience with SO. I followed MO a bit (not really actively participating) and used SO only via google.

For minor edits (spelling, etc.) I think we can agree that no particular marking of the edit is necessary. So the following only concerns major edits (like adding new information or major reformulations).

Usually I not just want the solution to a question, I also appreciate any surrounding information, that technically does not belong to the question. For example on the question "How do I underline text?" there is a simple technical answer. However, someone might write a comment that one usually shouldn't underline text (and explain why). For this reason I usually also read at least the upvoted comments and some additional answers.

Now imagine the questioner has some really good reason for wanting to underline text and edits this into her questions without marking the edit. When I come along and read the question and comments, I'll be rather confused why somebody recommended not to underline the text when there is a really good reason stated right in the question. Sure I could read through the editing history, but how often do really do that. More importantly, how many people that find the question via google know that there is an editing history and how to access it?

The same goes for clarifications in the question when it already has answers. Consider this question and Will's answer (which currently is the most upvoted one). Without the "EDIT" in the question anyone coming along now will probably be quite confused by this answer as it does not answer the question.

For this reason, I usually appreciate it when (major) changes to a question and answer are clearly marked; either directly in the text or as footnotes.

Interestingly this does not necessarily lead to convoluted and unreadable answers. Consider this answer by Andrew. First (nearly) the original answers which just gives the solution. Then the first edit, two paragraphs with an explanation. Then the second edit, a paragraph with additional information. So it is a well-structured answer where the most important information comes first and not just a wall of text. (I'd only change the two headlines to "Edit: Explanation" and "Edit 2: Additional Information".)

1

I always go for readability first and foremost.

So usually, I try to edit new information seamlessly into the post (if I want to notify people that I edited the post, I post a comment under it)

Sometimes, the edit is in direct response to a comment or such, in which case I might put it in a separate paragraph at the bottom, following an "*Edit: *"

I never mark specifically what has been edited. It's not really relevant, and if people want to see that information, they can check the post's edit history anyway. Posting that information just clutters my post, making it less readable.

Sometimes, I find my post is more readable when I integrate edits seamlessly. Sometimes, it is important to make the distinction that "this was edited in afterwards", typically because it is in response to some other answer or comment.

But the edit history should never be there, and the explicit "edit" indication should only be there when it makes the post clearer.

Remember that the SE software already marks your post as "edited X hours/days ago by Y". Readers can easily see that you edited, even if you don't point it out. Often, that is all they need to know.

0

I do think that there is a case for making the edit clear when not doing so would lead either the original questioner or a casual browser (ie one not inclined to click "edited" before making their judgement) is likely to come to the wrong conclusion as to how things have developed. Given that people generally don't read the full history before voting, I'd like to avoid the situation whereby people vote down an answer that was initially helpful. Although the reputation system shouldn't be taken seriously, it is part of the system and helps impart that warm glow of knowing that you've helped someone so it shouldn't be ignored. Getting an unjustified down-vote for one of one's first answers can easily deter someone from using the site. I want this place to have a reputation for friendliness as well as for usefulness! (Compare with the reputation of MO!)

That said, I probably did go a bit overboard in my answer that provoked this whole discussion. I've cleaned it up a little, whilst preserving the overall structure. I wonder if that is now to everyone's liking or not!

  • 2
    as to how things have developed — Well, the point exactly is that this is unnecessary to most readers. :-) It does not add to the usefulness of the answer, and is of only secondary importance. Besides, when it matters (which is only when some other answer is likely to be downvoted), it just needs a clear mention, not a preservation of the entire historical record. (This mainly matters when questions are edited, too... for questions it may be worthwhile to point out edits more explicitly.) – ShreevatsaR Aug 3 '10 at 2:41
  • Aside: if we stayed true to the real function of the votes — which is to rank answers by how useful they are, and not to reward or punish users — we would downvote an answer when it becomes irrelevant or redundant. We tend not to do that, because many people are attached to their reputation (or at least sensitive to downvotes as if they were being voted on and not a particular answer), and it seems "unfair", but note it's now become about people rather than answers. – ShreevatsaR Aug 3 '10 at 2:53
  • 1
    @Shreevatsar: Actually, I'd challenge that. The reputation is meant to be a mark of how much the "community" values you so it is linked to the user in the system and not just in people's heads. If you're someone who goes round picking up on early vague questions and helping clarify them, then that's a very valuable thing to do and should be rewarded. For such answers, I'd simply edit them to include "Note: This answer referred to an earlier version of the question" and hope that this effectively removed them from the voting palaver. – Loop Space Aug 3 '10 at 8:12

You must log in to answer this question.

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