What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 133 Stack Exchange communities.

Stack Overflow (at least, and probably other SE sites) just recently got a slick new "inline" editing mechanism for questions and answers, where instead of taking you to a new page for editing, it just modifies the page in place with the editing interface, using JavaScript and dynamic HTML.

This is great. It didn't feel broken to me before, but it's cool, it's fast, and I bet it reduces server load, which is a Good Thingtm. But:

If you edit something of any size, particularly a question, so much changes (virtually everything in the window) that it gives a very strong impression that you've changed pages — I mean, you clicked a link, and everything seemed to change. If the question was of any size, it probably scrolled the page up to the top (since you were scrolled down a bit to click "edit") and just about everything you see is slightly different. The impression of page-change is very strong.

This doesn't happen editing smaller things, particularly shorter answers. There enough of the surrounding page stays put that it's clear(er) that you've just changed modes, not pages.

Because the page-change vibe is so strong with larger items, you immediately find yourself doing the natural thing if you decide not to edit or didn't mean to click that link or whatever: Clicking the Back button. But you don't find yourself back at the question, you find yourself wherever you were before that. This is surprisingly jarring.

My first thought was to set a browser history point. This is readily done in modern browsers (and not-so-modern browsers), and it would preserve the natural reaction to such a large change when clicking a link if you want to "go back" — using the Back button. As Jon Skeet said in a comment when a "cancel" button for edits was raised, "The back button pretty much is the web UI for "I didn't mean to do that.".

But when you're editing smaller things, that may not be the UX people expect, although I've seen it used to good effect (and I'm told Facebook does this with their image pop-ups).

Barring a history point, just something making it more obvious that you haven't changed pages would improve the UX. If we can get rid of the "page change" vibe on larger items, I don't think people will necessarily reach for the Back button. I never have when editing a comment. But I have, repeatedly now, when editing larger questions or answers.


Suggestion: Experimenting with this more, it seems to me the worst of it happens when the top of the thing being edited is off the top of the browser window. The jump makes everything move, even if many things stayed the same, giving the impression of a page change. The fadeout/fadein doesn't help, because it gives the impression of page refresh as well.

It may well be as simple as smooth-scrolling up to the top of the item (if the top is scrolled out of the window) before replacing it with the edit interface. I'll say frankly that I'm not usually a fan of that sort of thing, but I think it would break the illusion, because page changes don't scroll like that. I'd like to come up with something more complete, something that happens regardless of where the top of the item being edited is, but that would make a start.

share|improve this question
3  
But how would this suggestion handle the case where users realize that they've opened an inline editor and want to use the "Back" button to go back to the actual previous page that they've visited? –  Cody Gray Jul 11 '11 at 13:11
    
There's a cancel link right next to the "Save Edit" button, but I take your point about a change in behaviour. –  ChrisF Jul 11 '11 at 13:11
4  
@Cody: For something more clearly not a page change, fair enough. But the page changes radically. It feels like you've gone somewhere, hence using Back to go back. –  T.J. Crowder Jul 11 '11 at 13:15
    
@TJCrowder - believe I found other new edit hiccup: meta.stackexchange.com/questions/98132/… –  amanaP lanaC A nalP A naM A Jul 11 '11 at 15:27
    
@Neal: Bound to be the odd niggle with new functionality. :-) I expect it'll get sorted out, now it's been reported. –  T.J. Crowder Jul 11 '11 at 15:30
    
@TJCrowder. related to your question: meta.stackexchange.com/questions/97283/… see what the mod keeps doing to comments on their answer here... meta.stackexchange.com/questions/98137/… –  amanaP lanaC A nalP A naM A Jul 11 '11 at 15:39
    
What browser does it look like a page change? To me, it looks about as much of a page change as clicking "add comment" does. There's no "navigation blip" or anything else that signifies to me a change of page rather than a change of UI. I suspect this may be more due to how my browser renders things, though. –  Grace Note Jul 11 '11 at 15:50
1  
@Grace: I'm using Chrome, but I don't think it's a browser thing. Editing short answers isn't too bad, a bit like inline comment editing though not nearly so clear as that. But editing anything of any size (particularly questions), the whole viewing area changes, leading to the impression that one has changed pages. –  T.J. Crowder Jul 11 '11 at 15:54
1  
@Grace - I would say what makes it feel like it has changed is that the content you were looking at is no longer anywhere on the screen. It's not like clicking edit only adds to the screen, it also takes away. –  NickC Jul 11 '11 at 17:06
2  
You can hit Esc to also cancel the edit. –  Jarrod Dixon Jul 11 '11 at 18:18

2 Answers 2

up vote 5 down vote accepted

I just changed it so editing is way less jarring.

We now keep the comment section and the vote controls.

I am not strongly opposed to html 5 history, but IE does not support it and there are a bunch of complications.

The whole point of inline editing is to make editing seamless, if we are to populate the stack with every transition, you will be stuck pressing back 3 times just to get to your previous question.

share|improve this answer
    
Cool. When will that be on SO? (Just did a hard refresh and tried it.) Agreed that too many history points would be troublesome, but even after living with this and actively trying to adjust to it, on anything more than a few lines long, it still feels very much like a page change. Hopefully your changes will mitigate that. :-) –  T.J. Crowder Jul 12 '11 at 7:33
    
Re history: The new history API is just one option. There are plenty of IE-friendly solutions to adding history points. –  T.J. Crowder Jul 12 '11 at 7:36
    
@T.J. I know we use those hacks on the user page. SO is deployed now –  waffles Jul 12 '11 at 7:37
    
for the record @t.j. I always wanted the vote controls and comment sections to remain in place, there were just development issues preventing it from happening immediately. –  Jeff Atwood Jul 12 '11 at 8:18
    
@Jeff + waffles: Those changes really help, thanks. I would still prefer a history entry, but it's purely in the preference category now as opposed to something that seems wrong, if you follow me. Great work, waffles, nice one. –  T.J. Crowder Jul 12 '11 at 9:11

I don't agree that inline editing looks like or acts like a page change at all.

Adding a browser history point here would cause extreme confusion because the user has not navigated away from the question page in any way, shape, or form.

(also consider what would happen if you clicked edit on two answers -- would you then have to click back three times to leave the page?)

share|improve this answer
    
How about if you're inline editing two posts and click back you're beset with a multitude of popups? –  random Jul 11 '11 at 15:59
    
@Jeff -- what about this issue: meta.stackexchange.com/questions/98132/… –  amanaP lanaC A nalP A naM A Jul 11 '11 at 16:05
2  
@Neal: That's a separate issue, let's not conflate things. They're both just niggles that need to be sorted. –  T.J. Crowder Jul 11 '11 at 16:17
1  
@Jeff: Let's see. I didn't navigate? Hmmm. I clicked a link. The page completely changed as far as I can see (I'm talking about editing a question of any size). That's navigation in my book. –  T.J. Crowder Jul 11 '11 at 16:19
2  
Not to say you can't have your own opinion on this, but even Facebook's photo-popups (which definitely aren't a page "change") add a history point. Clicking the back button closes the popup. –  NickC Jul 11 '11 at 16:20
    
@TJCrowder, sorry, i thought they were a little related. –  amanaP lanaC A nalP A naM A Jul 11 '11 at 16:22
    
@t.j. when I click edit here, it does not look at all like "the page completely changed" -- just my answer area on the page changed, so perhaps 30-40% of the page? See for yourself... i.stack.imgur.com/oYTyy.png –  Jeff Atwood Jul 11 '11 at 17:10
1  
@Jeff: Yes, as I said to Grace, smaller things don't cause the same UX. But click "edit" on my question (I think it's big enough, don't have enough edit rights here to check), and everything seems to change. I have a reasonably-sized browser window, but so much changes it gives a very convincing "the page has changed" effect with similar-sized Q's on SO (I have enough rep there). All of this suggests to me that since small things and big things differ, and inline is cool, the latter rather than former part of my suggestion would be best: Making it clearer that the entire page hasn't changed. –  T.J. Crowder Jul 11 '11 at 17:20
2  
@t.j. even when I edit your question, above, the entire header is unchanged, the sidebar is unchanged, the vote gutter spacing is the same, nor does the entire page "flash" as a typical full page load would. I think we'll have to agree to disagree on this one. –  Jeff Atwood Jul 11 '11 at 17:24
1  
@Jeff: I think you'll see what I mean if you have your window scrolled down just a tad when you click edit. (And I've edited the question.) The jump is a big part of it. (I get a similar effect if the top of an answer of any real size is off the top of the window when I edit that.) –  T.J. Crowder Jul 11 '11 at 17:42
    
@TJCrowder Can you suggest how it can be made clearer that the entire page hasn't changed? Especially if it's not looking like a page change to some, it's not exactly easy to find out what should be changed when it seems plain as day. Some starter point on what should be looked at, and what can be done to make it less jarring, would probably get some more ground. –  Grace Note Jul 11 '11 at 17:59
    
@Grace: Yeah, agreed; the history point was the original suggestion but I don't currently have an alternative one. I'll give it some thought. The odds of my coming up with something Jeff will like are low, :-) but I still hate to just offer no suggestions. Most of the ideas immediately coming to mind are easily rejected, though. –  T.J. Crowder Jul 11 '11 at 18:07
    
@Grace [ and Jeff -- but I think I heard somewhere, Jeff, that you'll get notified regardless ;-) ] Fooled around with it some more and added a suggestion. Perhaps not a complete answer, but possibly a start (and who knows, possibly enough). –  T.J. Crowder Jul 11 '11 at 18:14

You must log in to answer this question.

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