8

Today I was reviewing this suggested edit and I wondered.. what changed?

Note: right-click and view image to see the full size image, or view the live thing linked above

spot the difference rendered output

MIME-Version was replaced with MIME-Version?
So i viewed the 'markdown' version:

spot the difference markdown

Could it be that the - was some uni-code character instead of a plain simple 'ascii' -, resulting in an (hard to track-down) error after copy-pasting the code from this answer?
Upon further investigation I found that the original code actually was MIME Version.. with a space.

By the time I finally figured out what what the intended change was and validated that this was correct, the suggested edit was rejected (understandable, I almost did the same, but I recognized the 6-char minimum frustration expressed by the editor in the comment).

Obviously, had the editor explained his change in the comment (besides his (in my opinion) reasonable frustration), I (and I think my fellow peers) wouldn't have had to scratch their head wondering what (apparently important thing) changed.
However, in the editor's defense.. how could he know this simple edit is so difficult to see ?

I understand that this could be considered an edge-case, but a - is also used in math/formulas and suddenly this 'edge-case' becomes more common (at least on SO where it is all about coding).

Difference highlighting is there to help the reviewers (giving them a fighting chance, given the vast amount of review-tasks available every minute) and it usually does a very good job..
We generally just compare the difference and (using our expertise or google) validate the edit (or skip it and leave it to someone who can make right call).

So then.. shouldn't the reviewers be aware of such cases where one can't see the difference?

The coder in me would instinctively protest (to having to remember and look for known edge-cases) and would first ask: can't we just have the computer just show us the difference so we can continue to focus on the task of validating the edits?

So here is the essence of the question/suggestion/feature-request to discuss:
Can the difference-highlighting be improved (in a reasonable way)?

If the deleted/replaced text's decoration wasn't strike-through, then the red vs green color-highlighting alone would have clearly depicted the difference.
But I can imagine that people with color-blindness (or overly cheap displays) would appreciate the extra visual indication of the strike-through text. Even I like the extra visual hint.

One possible solution might simply be to change strike-through to line-over.. I honestly never saw a line-over (character) in code (and hardly ever in plain text..).
A simple style-rule is all it would take.

Example:

suggestion

and

enter image description here

However.. it would need to be tested thoroughly on humans and I realize there are many different types of formatting in posts..

Alternatively, one could also implement a little extra script allowing us to click both original (left) or suggested edit (right) post block that toggles highlighting..
That way one doesn't have to scroll to some kind of check-box (I did think about it haha).

Or... share your solution as an answer!!

7
  • 2
    +1 This can get annoying. Also, red freehand circles. Nov 12, 2013 at 4:01
  • 1
    +1 for caring enough about review quality to track this down.
    – Undo
    Nov 12, 2013 at 4:02
  • The more I think about it, I'm starting to prefer the alternative solution I posted: the toggle highlighting script..
    – GitaarLAB
    Nov 12, 2013 at 4:31
  • Had the editor just said something like "Changed an incorrect space to a dash. Added some whitespace as a workaround for the 6 character limit.", people would've actually known what's going on. Nov 12, 2013 at 8:06
  • 2
    Side note - I don't think @ notifies a user that suggested a rejected edit. Nov 12, 2013 at 8:08
  • @Dukeling I'm sure I read somewhere that it doesn't. Nov 17, 2013 at 1:03
  • I tried that before - it doesn't notify the user (I hate the use of the term "doesn't work"). Nov 17, 2013 at 1:04

0

You must log in to answer this question.

Browse other questions tagged .