986982 - Set browser.backspace_action to 0 by default on Linux so Backspace key navigates back like Mac and Windows

Status: NEW

(Firefox :: Keyboard Navigation, defect)

Description

[Yunier J.]

User Agent: Mozilla/5.0 (X11; Linux i686; rv:29.0) Gecko/20100101 Firefox/29.0 (Beta/Release)
Build ID: 20140318013849

Comment 1

[Yunier J.]

In Linux the user must click on Back button for go to previous page.
Is very easy and comfortable press backspace key for do it.

Comment 2

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/integration/fx-team/rev/3960ad0617ff

Comment 3

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

(In reply to Wes Kocher (:KWierso) from comment #2)
> https://hg.mozilla.org/integration/fx-team/rev/3960ad0617ff

This was supposed to go to bug 986932, but it landed with the wrong bug number. Sorry for the bugspam.

Comment 4

[YF (Yang)]

Bug 358764.

Comment 5

[Yunier J.]

Somebody can answer to this request? Please.

Comment 6

[Panos Astithas (he/him) [:past] (please ni?)]

I think it is time to revisit the past decisions, based on how the ecosystem evolved in the last many years. There is a long history in bug 358764, bug 325541, bug 262905, bug 208035 and bug 36922 that explains why things came to be like this.

It is a great design choice that there is a pref for selecting what is best for everyone, but defaults should be catering to the less experienced users as those are less likely to know how to tweak the browser to suit their needs. After all these years I expect both sides of long-time users to have changed the do-nothing default to either back or page-down, so they wouldn't be affected. Those who liked the do-nothing behavior are obviously not using the backspace key in this context and therefore are also not going to be affected.

Apart from the constituency priorities, there is also the matter of product consistency and Firefox has been using backspace to go back in every other platform but Linux for a long time. It is important to note that both Chrome and Chromium on Linux also map backspace to the back button, so we are the odd one out there. As for the argument about the already available alt-left shortcut to go back, I would say that if anything deserves a single-key shortcut that must be the largest button in the browser UI.

In the discussion we had in the Firefox desktop call yesterday nobody was against this change, so I will submit a patch to implement this in a bit.

Comment 7

[Panos Astithas (he/him) [:past] (please ni?)]

Attachment: MozReview Request: Make the backspace key behave consistently on all platforms (bug 986982). r?dolske

Review commit: https://reviewboard.mozilla.org/r/51899/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/51899/

Comment 8

[Chris Peterson [:cpeterson]]

Chrome just landed the opposite change: disabling backspace as a keyboard shortcut for back navigation.

https://codereview.chromium.org/1854963002

https://bugs.chromium.org/p/chromium/issues/detail?id=608016

The response on Hacker News is mixed:

> Nooooooooo!

> THANK YOU!!!!!! Progress

https://news.ycombinator.com/item?id=11729287

Comment 9

[Justin Dolske [:Dolske]]

Comment on attachment 8751202 [details]
MozReview Request: Make the backspace key behave consistently on all platforms (bug 986982). r?dolske

https://reviewboard.mozilla.org/r/51899/#review50718

Yeah, I saw the HackerNews post and was amused at the polarized responses.

I looked through the various Firefox and Chromium bugs a few days ago, and am somewhat loathe to poke at this sleeping dog... All options have a set of users vocal about why a particular behavour is clearly wrong, and (IMHO) none of the options are clearly correct. But I think Panos' point about consistency with other platforms still has merit. It might be that, when the dust settles, we might want to consider following Chrome's change (if it sticks!)  that on all platforms. But for now the question is what to do on Linux.

If we do change the pref, I think we should add a migrateUI() step to retain the existing behavour for existing users. I feel like this is a classic bikeshed, and that most of the discussion is driven by omgchange. So let's avoid that by not changing anything for existing users.

And, it's a bit heretical, but I wonder if it might even be worth adding a UI pref (to Prefs -> Advanced -> General -> A11Y) to allow changing behavour. At least as a temporary measure, and perhaps with telemetry to see if there are actually significant users who care either way. (It would be interesting to see how many users have changed the default today, and how many change it after this patch.) Although I strongly suspect the answer is "it just doesn't matter either way for most users."

Comment 10

[Kestrel]

Google search overrides backspace navigation, Safari 6 disabled it and now Chrome, given that it is a common annoyance it seems like browser.backspace_action=2 is the better direction to be going here.

Comment 11

[Yunier J.]

(In reply to Panos Astithas [:past] from comment #6)
> I think it is time to revisit the past decisions, based on how the ecosystem
> evolved in the last many years. There is a long history in bug 358764, bug
> 325541, bug 262905, bug 208035 and bug 36922 that explains why things came
> to be like this.
> 
> It is a great design choice that there is a pref for selecting what is best
> for everyone, but defaults should be catering to the less experienced users
> as those are less likely to know how to tweak the browser to suit their
> needs. After all these years I expect both sides of long-time users to have
> changed the do-nothing default to either back or page-down, so they wouldn't
> be affected. Those who liked the do-nothing behavior are obviously not using
> the backspace key in this context and therefore are also not going to be
> affected.
> 
> Apart from the constituency priorities, there is also the matter of product
> consistency and Firefox has been using backspace to go back in every other
> platform but Linux for a long time. It is important to note that both Chrome
> and Chromium on Linux also map backspace to the back button, so we are the
> odd one out there. As for the argument about the already available alt-left
> shortcut to go back, I would say that if anything deserves a single-key
> shortcut that must be the largest button in the browser UI.
> 
> In the discussion we had in the Firefox desktop call yesterday nobody was
> against this change, so I will submit a patch to implement this in a bit.

Thank you for your explication and support it idea.

Comment 12

[Yunier J.]

(In reply to Kestrel from comment #10)
> Google search overrides backspace navigation, Safari 6 disabled it and now
> Chrome, given that it is a common annoyance it seems like
> browser.backspace_action=2 is the better direction to be going here.

(In reply to Chris Peterson [:cpeterson] from comment #8)
> Chrome just landed the opposite change: disabling backspace as a keyboard
> shortcut for back navigation.
> 
> https://codereview.chromium.org/1854963002
> 
> https://bugs.chromium.org/p/chromium/issues/detail?id=608016
> 
> The response on Hacker News is mixed:
> 
> > Nooooooooo!
> 
> > THANK YOU!!!!!! Progress
> 
> https://news.ycombinator.com/item?id=11729287

I don't think that do same of Safari and Chrome will be the best option.
Let to a survey and see what think our users.
Sometimes the people not use the features or keyboard shortcuts because they aren't know that it exists (see Panorama).
And, if no body use backspace key to back previous page, always must exits a option to set it.

Comment 13

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Shouldn't we change Platform/OS to All/All and mark bug 1041377 as dup of this bug?

Comment 14

[Amin Al Hazwani [:amin] (Firefox UX)]

Attachment: chome-toast-prev-page-shortcut

Do you we have metrics about how often people use the backspace to go back to the previous page? I would also be interested in understanding how often people press the backspace (by mistake) while filling a form. I was reading this article https://www.engadget.com/2016/05/20/google-tests-killing-backspace-chrome/ and it seems that in Chrome "0.04 percent of page views are currently triggered by going back with the backspace button. In addition, 0.005 percent of page views are from people using the backspace key after a form interaction". Today when you hit the backspace in Chrome a toast shows up above the location bar and displays the new shortcut for getting back to the previous page (see attachment). 

I would suggest to change the default to a new shortcut and allow savvy users to revert it back to the backspace. What do you think?

Comment 15

[Panos Astithas (he/him) [:past] (please ni?)]

Unfortunately, I don't think we have telemetry on how often backspace is used to navigate back, although it wouldn't be hard to add it. Here is where the probe should go:

https://searchfox.org/mozilla-central/rev/7a8c667bdd2a4a32746c9862356e199627c0896d/browser/base/content/browser.js#2085

I like the toaster approach, but it might be overkill for us as we don't have to explain a change of behavior to existing users. Or are you suggesting we switch to meta-left for back-in-history everywhere (browser.backspace_action=0)?