Why are trivial edits discouraged?

Question

Looking at Edit questions and answers privilege, I find this to be a surprising requirement:

Try to make the post substantively better when you edit, not just change a single character. Tiny, trivial edits are discouraged.

What motivated this, or rather, what does this mean exactly? Does that, for example, include removing stuff like "Hi, all", and "thanks in advance"?

[update] I was led to a blog post talking about this topic, but it has a million Comments, and doesn't really justify the rule (unless it's hidden in those million Comments). However, looking at Jeff's Comments gives me an idea that this was meant as a general guideline, instead of a hard-and-fast rule: If you edit a post, check elsewhere on that post if there are other improvements that can be made.

Answer 1

I've addressed this several times already... But I'll try for a little bit different perspective here:

Keeping it simple

There are two common models for editing in most widely-used software packages. One doesn't track changes at all (at least not publicly), while the other tracks every change. Let's call them the Forum model and the Wiki model:

Forum edits

Forums and message boards grew out of email and USENET, and as such many didn't (and still don't) offer any sort of editing support: you post your message, it enters the timeline and stays there, static, forever. The big downside here is that most of us make lots of mistakes, and seeing those ensconced on the 'Net FOREVER kinda sucks.

So some forums let authors (and maybe moderators) make edits, at least for a short while after the message has been posted. Generally though, there's no revision history to speak of, beyond perhaps a simple "edited at [time] because [reason]" footer. If I post something like,

Obama has forever ruined Life in America as we know it.

...then wait for you to respond and change it to,

I like waffles!

...there's not much you can do to prove I wrote something else. If a moderator, drunk on power as is their nature, goes in and changes it to,

I like waffles, but don't have any. THANKS, OBAMA!

...then there's not much I can do about that either.

Generally speaking, forums have a pretty crappy (or non-existent) editing culture.

The wiki model

This is the other extreme: every revision is kept for an indeterminate amount of time; if I want to see what you changed when you revised your post at UTC 7:31, I can look that up. If I want to see what you changed at UTC 7:32:19, I can look that up too. If I want to see everything that changed on a post between UTC 5:00 and UTC 12:59, there are tools that'll let me select a range of revisions to compare as well.

The advantage here is that nothing is hidden - if someone decides to play silly games, their shenanigans will quickly be noticed by one and all. As such, it's easy for a robust editing culture to develop: no one needs to be afraid of Being Bold with their edits, since there's no chance of inadvertently attributing their work to someone else.

The disadvantage is that it quickly becomes very noisy. If I want to see what's changed on a given page since the last time I looked at it, I might have to walk through multiple revisions - perhaps dozens or more - before I'm able to get a clear picture of what happened. The tools get more complex to combat this; pretty soon, someone implements a "minor edit" feature to allow hiding revisions that aren't really that interesting... Which then starts to erode the advantage outlined above.

Stack Exchange edits

The Stack Exchange model for editing is a hybrid of these two approaches. While it falls somewhat closer to the wiki model than the forum model, it deviates from the latter in a few key areas:

• Edits made by the same editor in a short period of time are collapsed: no matter how many times a post is edited within a 5-minute window, only one revision is stored (as long as only one editor is involved). This largely eliminates the need for a "minor edit" feature.

• Tools for comparing revisions are much more limited: only revisions directly adjacent in the chronology can be compared.

• Outside of Community Wiki posts, an original author is always maintained and clearly identified even if substantial changes have been made by other editors since the post was created.

Together, this hybrid system allows for a much simpler, much easier-to-read, easier-to-navigate user interface. However, this comes at a cost: trivial one-off edits must be discouraged in favor of more comprehensive edits.

If merely maintaining a simple UI were the only concern, it might be worthwhile to consider an optional "expert" mode that allowed minor edits at the cost of a more involved UI. However, there's a much bigger concern...

The cost of ownership

As much as it can act like a wiki at times, Stack Exchange posts are not wikis. The most important distinction is the one I noted above: nearly all posts have a clearly-identified author, whose name and image always appear prominently near it, whose account gets full credit for the post and whose reputation score increases or decreases according to how the post is received by the community.

The advantage of this was perhaps debatable in the early days, but has since become clear: folks will put an awful lot of effort into building their personal portfolio. Selfish? Sure - but we all benefit from the results.

The downside of this is... Well, folks can get a bit tetchy when you mess with something they consider a reflection of their expertise or personality. Early on, 3rd-party edits were a very controversial part of the system; even today, the notion that someone - anyone - can come in and mess with your words can be very disconcerting to new users.

There were wars... Fought over very minor edits.

And it quickly became clear that it is hard to sell someone on the value of 3rd-party edits when all they see is pedantry. If someone takes your barely-legible, heavily-downvoted question and turns it into an easy-to-read, easy-to-answer showpiece, well... The proof is in the pudding! But when someone else's name suddenly appears on Your Masterpiece and you look only to find that they've been obsessively adding Oxford commas and nothing else... Well, shucks - that can be a bit hard to swallow.

Conclusion: the system gently encourages behaviors that make life easier for all concerned

No, that wasn't a title; that's the actual conclusion here. There are other factors involved that I haven't touched on, but you can find them linked to in the first paragraph.

Answer 2

I think this is an excellent question and I'm unsure about the answer myself.

On the one hand, a website where all questions use "I" instead of "i" and all sentences end with proper punctuation looks and feels professional and serious, which is very good. People often match their behavior to the behaviors of those around them, so newcomers are more likely to use "you" instead of "u", or take a minute to proof-read their posts, when posting in such a site. This is especially true for question titles, which have a much higher visibility than just content.

On the other hand, there's no real upper bound to how many times you can improve a post. I've had posts of my own that I carefully proof-read and fixed all mistakes, only to encounter a few more grammatical mistakes later on. Is it worth the bump just to fix it? Would anyone care? Would anyone care if I wrote "starcraft" or "StarCraft"?

Personally, I try to edit formatting, spelling or grammar mistakes or omissions, but avoid editing for capitalization issues. I also try to avoid small corrections on posts that already give a high-quality impression. If someone posted a 4-paragraph + diagram answer, I'm not going to edit in a period after the last sentence of paragraph 3. It's just good enough, gives the proper feeling.

Also, Arjan raises an important point – I avoid editing within the author's grace period, especially if it's not a new user.

Answer 3

I think those wiki pages were written by us and vetted by the staff.

Since basically everyone here on meta is surprised by this I don't know how authoritative that sentence is; probably it simply reflects the opinion of the person who wrote it rather than the community's, and no one noticed it before.

Answer 4

I assume the motivation is twofold:

• Edits bump questions back on the frontpage, which is not always liked.
• When the body has been edited by 5 different people, a post becomes Community Wiki.

Answer 5

The privilege pages are editable (to those with enough rep) here on Meta and the changes pushed out to the other sites - see this question So to see who wrote what and what changes have been made you need to look here.

Looking at the revision history here on MSO you can see (assuming you can see this of course) that most of the post was written by Jeff.

So I think it's fairly authoritative.

Other answers detail why trivial edits are to be discouraged.

Answer 6

I think trivial edits are not the issue, but the misconception of 'editing' to improve an answer against just editing to correct an error. Neither are 'trivial', but there is no clear difference by using such derogatory term.

Answer 7

Note: This answer wasn't originally written on this question and was merged from another question.

If the issue is rep, just don't give any for such edits. If the issue is bumping, don't bump such edits.

Who decides, which edits are rep-worthy? I don't think giving this privilege to the reviewers is a realistic option, and I don't think moderators would like it either.

If the issue is that too many people have to review the edits, lower the rep limit for minor edits.

How does the code know what is a minor edit, and what should be reviewed?

If you think that 2k rep users will somehow find it on their own or that flagging is the proper solution, you're introducing unnecessary bureaucracy.

The problem with minor edits is not that they're small, but that they fail to improve most issues with a post instead of just one or two.

If you think that everyone can see a grammatical error without cringing, you don't know developer types.

You don't need to be able to see all mistakes, but while you're at it, improve as much as you can.

---

If there's only one minor issue with a post to fix and everything else is fine, I don't think fixing that one issue is a too minor edit, but if the post is littered with issues and you only fix a few letters and perhaps retag a question, then it is too minor.

The goal is to reach the perfect post in fewer edits. Rather than having a person suggest 10 edits on the same post to gain 20 rep, we can force them to do it all in 1 edit. And ideally, if there's something minor the suggested edit missed, but it improved most of the post, the reviewer would improve the post to fix the last issues.