The backspace key pages up instead of going back in history

I can confirm this.

Upstream changed the default behavior of Backspace on Linux builds in Firefox 2.0 betas.

You can change it back by opening about:config and changing browser.backspace_action to 0.

This is a really *annoying* change, because:
1. It changes one of the most-used shortcuts with no good reason, and most people won't find the about:config hack (especially not those that will be confused the most).
2. You need two hands to press Alt+LeftArrow on many non-English keyboards (i.e. those that have an AltGr key).

Can this please be reverted?

maybe this should be sent upstream?

I can also confirm this; although I found out about the about:config hack, it still seems like the default breaks expected behavior. I think the reason they changed it is that sometimes when entering form data, backspace could go back a page instead of deleting a character as expected - but are there people who prefer the new behavior?

It is annoying. Even en_GB keyboards have AltGr; I guess this is especially annoying to everyone outside the USA. Fortunately, my hand can stretch across my laptop's keyboard between the "Alt" and "<-" keys, but anyone with a smaller hand or bigger keyboard will be at a loss. It's uncomfortable to me, as it involves unusual wrist movement. This is a large usability regression.

I was checking around to see if they would fix this, and, as Marius Gedminas points out, they see this as a feature (and I started to support that after reading arguments). Upstream won't fix this. I doubt Ubuntu will fix it as well...

It seems this is when they changed the behavior: https://bugzilla.mozilla.org/show_bug.cgi?id=325541 (and GOD they have sooo many bugs on this). I'll try to find the upstream bug report that wants this to go away (which probably be marked as "wontfix")...

I have read several of their arguments, and none of them are convincing IMHO. Why do you think this is a feature?

In my opinion, if hardcore unix users prefer the backspace = pageup behaviour, they can easily change it, as they are experienced users and will find the way how to do it, but breaking existing behaviour for inexperienced users to please the proverbial two or three geeks is stupid.

Ubuntu isn't released with the Caps Lock = Ctrl hack either.

Exactamondo.

I think you've got it the other way round Jan, the avarage user will benefit from the backspace = pageup and it's the hardcore unix users that are currently complaining.

You avarage user doesn't use keyboard shortcuts, they use the mouse to click the back button and so they won't notice the change that much. In fact the only time they really notice the shortcut is when their form window accidentaly loses focus and they loose their form data because the browser decided to go back a page.

On reading the firefox bug report you'll find that this was done for better platform (every other app does a page-up) consistancy, so once again this is a feature, not a bug. Live with it.

What about Windows users migrating?

Most people I know use backspace in Windows in Firefox and IE.

Firefox is it's own platform; it runs on many systems. Why change aboriginal behaviour?

I personally had to use the hack. Especially on a laptop this feature is dreamy.
I was shocked to see it removed together with them killing the Auto search feature.

@ Arwyn Hainsworth
Firefox doesn't go back to the previous page if you press backspace while filling out a form, so there is no data loss. And saying that shortcuts are not meant for average users is denigrating.

Also, consistency is something that often improves usability, but not if you use it without thinking. To give you one example: going back to the previous page in lynx, links, elinks, info, etc. is done by pressing the LeftArrow key, so if we change Backspace to PageUp we should add this LeftArrow shortcut to Firefox too. Because we need consistency, you know...

@ Toby Smithe
This is not only about Windows users, but also about those users that have been using Ubuntu for 2 years now (or linux even longer), and suddenly it starts doing weird things.

Jan Claeys wrote:
> @ Arwyn Hainsworth Firefox doesn't go back to the previous page if
> you press backspace while filling out a form, so there is no data
> loss.
>

to reproduce data loss:
1. start filling out a form
2. click somewhere outside the form (one can always miss a form field
while trying to click on it)
3. (accidentally) press Backspace (as if you were gonna delete a
character in the -missed- form field)
4. unless data was kept in the memory (try going forward, weirdly enough
that's not supported by all forms), oopsie ;)

> @ Toby Smithe This is not only about Windows users, but also about
> those users that have been using Ubuntu for 2 years now (or linux
> even longer), and suddenly it starts doing weird things.
>

I see your point. I've been using linux now for about 1,5 years (maybe
less). I still think this backspace bug is a feature (hehe). I saw a
comment that demonstrated this extremely well in one of the bug reports
(someone was citing an example of data loss, where his/her customer was
trying to delete part of a form but instead lost data in the form), but
I can't find it now. that was the comment that changed my mind. before
reading that comment, I strongly objected this pageup thing as well.
sounds completely weird at first...

Note that I don't know about the usability of the new shortcuts (is it
easy to use them with one hand? etc)

> And saying that shortcuts are not meant for [used by] average users is
> denigrating.

I agree.

Okay, you can have data loss if you make two accidental misses in a row, one with your mouse, the other with your keyboard. It's much easier to get data loss by making one accidental click on the back button or the close button or by pressing Ctrl+W or Alt+F4, or ...

I think that Firefox should ask if you really want to lose the data in the form, but IMHO that's an entirely different bug (Epiphany asks in some cases, but not all).

(BTW: whether pages keep their data after a go forward depends on caching headers and maybe sometimes JavaScript used.)

And the new shortcut requires the use of 2 hands for many users that don't use a US keyboard (if you have a US keyboard, try pressing LeftAlt+LeftArrow, as many non-US keyboards have an AltGr key where RightAlt is on US keyboards, to be able to access additional characters, accents, etc.).

Jan Claeys wrote:
> And the new shortcut requires the use of 2 hands for many users that
> don't use a US keyboard (if you have a US keyboard, try pressing
> LeftAlt+LeftArrow, as many non-US keyboards have an AltGr key where
> RightAlt is on US keyboards, to be able to access additional characters,
> accents, etc.).
>

that was what I meant as a usability issue in my previous reply...
making firefox US-centric is a no-no...

leftalt+left arrow is bad for laptops as well (my left arrow is on the
right side, leftalt is on the left side, and my hand is not big enough).

then, for me, the question is, should that go to upstream as a usability
loss bug (if there is such a thing) or should it be fixed only by
ubuntu. that, of course, would have to be answered by a devel :)

Upstream already rejected this AFAIK, and I don't think a fix by Mozilla.org could be get before Edgy release anyway.

If this change was made to prevent data loss when someone mis-clicks a form field, then backspace should just do nothing, right? Won't the person be a bit confused/annoyed when the browser pages up rather than deleting a character from the field? I imagine if nothing happened they might quickly figure out that the field doesn't have the focus.

@Brad Pitcher: Agreed, having backspace do nothing would be the best default behariour. That way a misplaced click will not yield unexpected results.

Arwyn, do you really want to disable all potantially dangerous keys and buttons? Firefox would become completely useless very fast if you go that route...

The possibility for data loss thing is a bug that should be fixed in a general way that doesn't harm usability; e.g. by asking the user if they really want to lose what they wrote in the form whenever they do something that might have that effect.

Throwing away day-to-day usability to mask (not fix!) a bug is never a good solution.

While I agree with the commenters that say there is are probably better ways to fix this, from my perspective this is definitely a step in the right direction. At least if I accidentally get outside the data entry field and then I backspace I end up on a different part of the page rather than a different page all together.

@Jan:
All those buttons that can cause data loss and are easy to click accidentaly should be removed. This is standard usability procedure, data loss should not be easy to cause by accident.

That said, in FX this is the only one I've come across, if you know of others, please tell me, so I can file bugs for thier removal.

This has nothing to do with what keyboard you use (I have a GB keyboard with AltGr) and everything to do with a bug that causes data loss.

There are indeed a number of ways to fix this bug, but removing the behaviour that causes the bug is by far the most elligant way of solving it in this situation.

If you consider the fact that there is now only one shortcut that is not easy to do on non-us keyboards a bug, you should file a bug to that effect and not try to repleal a usability bug fix. There are a number of shortcuts possible that are valid for both non-us keyboards and us keyboards that are both easy to use and won't be pressed by accident, how about using one of those?
or how about adding the shortcut AltGr+back = 'go back a page' to get identical behaviour on both us and non-us keyboards?

Arwyn, your proposal doesn't fix anything, as accidentally pressing AltGr+Backspace instead of AltGr+- or AltGr+$ or whatever other keys are next to it on your keyboard would cause the same dataloss (actually AltGr+$ = "]" on Belgian keyboards, so this isn't an really uncommon combination to press for almost anyone using a web forum).

And the same potential for data loss is there for every possible way to implement this functionality (including the current Alt+Arrow ones and the back/forward buttons on the toolbar).

There are probably hundreds of different ways to get the same type of data loss in Firefox currently; you can't remove all those keys, key combinations and buttons, because there would be no useful keys/buttons left.

I agree that this sort of data loss should be prevented, but this should happen in a sane and clean way (and it can be done; the Firefox developers just have to take the time to do it instead of adding new "features").

Read my comments carefully, I have nothing against keys that can cause data loss, sometimes that is the required behaviour. It would be a bit troublesome if we couldn't quit a program when we wanted to because we have changed some setting.

No, what I'm against is keys that make it easy to cause ACCIDENTAL data loss. In none of the other examples you provided is it easy to accidentaly loose data, only on purpose.

And this _is_ the sane, clean way of preventing data loss, all other ways are non-trivial, will take up developer time and will provide the same or worse results.

If you would read what I have written, you would know that I have got a lot more data loss by those other keys than by the backspace key. Not everybody has the same habits, so not everybody makes the same accidents.

I did read what you said and all I can say is that you must be a very careful typer to lose more data to pressing Alt+F4 than to the backspace bug.

After all it is a lot easier to hit backspace when typing than to click the back button or to hit Alt+F4...

I'm going to withdraw from the argument now, since I believe my point has been made and all I can do is hope that the devs in charge agree with it.

It wasn't broken, I don't see why it was "fixed", and why so many users of a known feature had to file or find bug reports to discover it. This is precisely the sort of thing that ticks off Joe Average as well as power users. Consistency should be considered a far more important priority, IMHO. At least tell the user when their settings are changed for them.

</rant> :p

this is not invalid, but won't fix.

ok, i added ubufox (the ubuntu firefox extension) package as a valid bug target ... I will consider to enable the default backspace action for users that have ubufox installed (which should be most as ubufox is shipped by default).

I targetted this bug for gutsy tribe-6 milestone so I get remembered to come up with a decision by then.

On Tue, Aug 21, 2007 at 08:41:04AM -0000, Kevin P. Munn wrote:
> I have read the commments, and arguements, both pro and con.
> As an "End User",, who prefers to keep things simple (KISS Disciple), Having just one key to press is much easier. for me, and colleagues who HAVE to use "Ubuntu" OS (Corporate Mandate) with Firefox. Pressing multiple keys can be confusing as you have to remember which combo does what. Playing "Chase The Cursor" while trying to type work related documents is also annoying, as it is the only viable option.

BTW, this can be irritating especially for novice end-users that don't
understand the concept of focus ... consider that you accidentally
leave a text fiel where you typed a lot of text, then you press
backspace and suddenly you go a page back. I imagine that this can
really confuse users at best ... or even shock them (hey i have lost
all the text i just typed).

 - Alexander

Alexander Sack wrote the following on 08/21/2007 05:15 PM:
> On Tue, Aug 21, 2007 at 08:41:04AM -0000, Kevin P. Munn wrote:
>
>> I have read the commments, and arguements, both pro and con.
>> As an "End User",, who prefers to keep things simple (KISS Disciple), Having just one key to press is much easier. for me, and colleagues who HAVE to use "Ubuntu" OS (Corporate Mandate) with Firefox. Pressing multiple keys can be confusing as you have to remember which combo does what. Playing "Chase The Cursor" while trying to type work related documents is also annoying, as it is the only viable option.
>>
>
> BTW, this can be irritating especially for novice end-users that don't
> understand the concept of focus ... consider that you accidentally
> leave a text fiel where you typed a lot of text, then you press
> backspace and suddenly you go a page back. I imagine that this can
> really confuse users at best ... or even shock them (hey i have lost
> all the text i just typed).
>
>
> - Alexander
>
>
I was thinking the same thing. Although I administer a network of about
35 users who run Firefox exclusively many on Windows where the
functionality exists by default. None of them have ever mentioned
having problems with it.

-Joe Baker

Here's the Firefox documentation page which says that on Linux Backspace should take the user back a page in history. Here's the footnote:
"For this shortcut to work on Linux, use about:config to set the preference browser.backspace_action to 0."

The Documentation link:
http://www.mozilla.org/support/firefox/keyboard#f1

Thanks for your attention to this matter. It really is much simpler to just hit backspace button.

The documentation page shows that both IE and Opera also use Backspace to go back in history as well.

I can confirm this is still present in the Gutsy release candidate. Everytime my girlfriend uses my computer she complains that backspace doesn't work, because it's the behavior that she's used to on Windows. Opera, Internet Explorer, and Windows Firefox all have backspace go back. Has anyone tested Konqueror, Epiphany, or Safari? I suspect their behavior is similar. I'm not sure what the developers mean when they say this is for more consistency, and frankly since it's not immediately apparent to me as a user, it's not a consistency I want or care about so their reasoning is irrelevant. I would support Ubuntu going against the upstream default. If enough distros do this, it might convince upstream.

1. Alt+Left still works, so going back is consistent in Gnome. Backspace also working doesn't conflict with that.

2. Backspace going up may make sense for browsing a filesystem (even that is arguable), but it doesn't make any sense for the web.

3. Normal users are going to see that it's the one key on their keyboard that has an arrow going backwards. It's fairly intuitive to expect it to go... back. Not up.

4. Page up is like a centimeter away. If I wanted page up I'd press page up.

5. It's more consistent with other browsers.

6. Did I mention my girlfriend complains?

7. If the data loss argument was persuasive, Mozilla would be adopting it for the windows port as well. But they're not.

Backspace doesn't work like this in Epiphany, but (according to its developers) that's because they can't implement it, because GNOME & Gecko aren't compatible in the way they can intercept keypresses (again, that's what I've been told by some Epiphany devs).

That's really annoying.

What if Volvo changed their car design so that you would switch gears with the steering wheel? Sorry, but that's how #&)#"/¤&/#¤ this is. Shortcut keys like this one is a de facto standard.

Arguments about old time unix users is void because A) on Ubuntu there is a lot more "new linux users" these days compared to the number of "old time unix users", and consider also B) in some ancient time some people used CTRL-Y and CTRL-V for page down/up. But guess what? More or less every single keyboard ships with a dedicated button that is labeled "PAGE UP", maybe it's time to evolve and start to use it? For the ultra hardcore people, I suggest "apt-get install old_unix_style_keybinding".

Arguments about "data loss" is void because: A) surf to google.com, B) type in some query, C) press the normal GUI back button in the toolbar, D) data lost. E) big deal get over it.

Arguments about "platform consistency" is void because: A) start the GNOME filemanager, B) scroll down to the last item and select it, C) press BACKSPACE and notice that it does not mean page up.

On a personal level; for now I'll go with the about:config hack, but give me more of these kinds of bugs and then it's bye bye new distro for me. I mean a missing sound driver is one thing because I know everyone *wants* to fix it (albeit it might take a long time). But with bugs like this one are big time showstoppers.

I hope that people can agree on a sensible fix to this bug so that ubuntu can continue to grow.

A more apt analogy would be changing the gearbox so that shifting into reverse is the *same* as shifting into first, because as we know, shifting into reverse while driving forward can cause some major "data loss". Then putting the actual reverse control not on the steering wheel (where it might even be somewhat convenient), but rather on a separate lever in the passenger glove box that you really have to reach for.

Sadly, yesterday was the ONE YEAR anniversary of this bug existing, and being unresolved (at least to our satisfaction), and I don't see Ubuntu or Mozilla developers eager to fix it at all. The Mozilla bug was reported in Feb 2006 and still nothing. Mozilla changed my original bug on reversing this, to "On Linux, backspace should do NOTHING", which was really irritating seeing as that is not what I wanted at all. I very much agree with you that this makes no sense, and I will not accept it until they also change the default behaviour on Windows and Macintosh platforms.

"Mozilla changed my original bug on reversing this, to "On Linux, backspace should do NOTHING", which was really irritating seeing as that is not what I wanted at all."

Agreed. This is a really irritating non-solution. If they're concerned about data loss, they should implement a solution that fixes *ALL* such data loss issues. Many pages use javascript to pop up a "Are you sure you want to navigate away from this page?" dialog. This should be just be standard on any page with user-changeable elements. Then the backspace/data loss "issue" would be moot.

If you enter any data into a form, then accidently press backspace
and goes backwards, when you press forward, the data you entered is
still there. So there isn't any dataloss.

I submitted a patch to fix this (make backspace go back on ALL platforms) . Please go vote for the bug so it gets in firefox 3.0!
https://bugzilla.mozilla.org/show_bug.cgi?id=400518

Also, Ubuntu should consider patching their packages as this is a pretty straight forward patch. No need to wait for mozilla, which can be pretty slow when it comes to fixing Linux bugs.

Go vote! -> https://bugzilla.mozilla.org/show_bug.cgi?id=400518

Yeah good luck with that. The Firefox developers and the Ubuntu Firefox package maintainers have all proven that they refuse to listen to user demands, and are going to require each of us to change our config manually. I've all but given up. I tried the exact same thing a year ago, trying to get people to vote for my bug, but my bug report was hijacked and changed the behaviour to do nothing instead of page up. I guess it doesn't hurt to keep trying, though...

Everyone keep trying, please. :-)

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

hackel wrote:
> Yeah good luck with that. The Firefox developers and the Ubuntu Firefox
> package maintainers have all proven that they refuse to listen to user
> demands, and are going to require each of us to change our config
> manually. I've all but given up. I tried the exact same thing a year
> ago, trying to get people to vote for my bug, but my bug report was
> hijacked and changed the behaviour to do nothing instead of page up. I
> guess it doesn't hurt to keep trying, though...
>

I wouldn't mind hand patching a single config file if it would affect
all the users as their defaults on a multi-user system. I want to
maintain a desktop system for thousands of users.

- --
Joe Baker
Computer and Networking Solutions Consultant
Specializing in Open Source tools that make our world better!

http://www.dcresearch.com

V O T E R O N P A U L F O R P R E S I D E N T !
http://www.ronpaul2008.com

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFHirpS7J1dPd3sAmARAqoUAJ9BWIx5UEoo4HFHXn7TxfsCjOZ67ACfd6Z8
x7QRYxz8IgT7LylSlI6AMjI=
=cMfw
-----END PGP SIGNATURE-----

we will not add a tweak for this.

Too bad, now I'll have to edit Firefox's about:config for every ubuntu install I do... :-(

I'm glad to see that Gentoo already does The Right Thing by setting the proper pref in their Firefox packages in order to reverse the decision of the Mozilla developers to break standard behaviour and cross-platform consistency. Now it's time for you, Alexander, to follow Gentoo's lead and do the same in Ubuntu's package. Thank you.

WOW, this is a very long thread! I would like to confirm this.

Now, I have not read all the posts here, but in my opinion enabling backspace would make it more intuitive for people coming from a Windows(/MAC?) world. I guess it's all about user-experience.
If ALT+LEFT/RIGHT is still enabled, I really don't see the "harm".

+1

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

The page up key provides a backwards scroll.
It would require only one hand to effect the back in history command if
it were tied to the backspace key.

- -Joe Baker

MadsRH wrote:
> WOW, this is a very long thread! I would like to confirm this.
>
> Now, I have not read all the posts here, but in my opinion enabling backspace would make it more intuitive for people coming from a Windows(/MAC?) world. I guess it's all about user-experience.
> If ALT+LEFT/RIGHT is still enabled, I really don't see the "harm".
>
> +1
>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFI1nwH7J1dPd3sAmARAr14AKCxNe2EBxlQIcYwrAuQWDe3ptEUtACgqUHt
AlcKLnss51oywbvIw8QeGuE=
=S4E1
-----END PGP SIGNATURE-----

Please don't tweak status of "wontfix" bugs. if you want this feature you should really lobby upstream and not with me.

Of course we will continue to try to get this bug fixed upstream, however as Mozilla are dragging their heels about this, Ubuntu has the ability to make this simple change to their package so that at least Ubuntu (and Gentoo who has already done this) users of Firefox will be happy.

I won't bother changing the status, but you need to listen to what your fellow Ubuntu users want here... Please don't dismiss this out of hand. There is no reason Mozilla should make this decision for Ubuntu. Thank you for continuing to release excellent packages, we simply want to help make them better.

Seems like Ubuntu users generally support this as default:

http://brainstorm.ubuntu.com/idea/9443/

We should decide based on what Ubuntu users want, not what Mozilla thinks is best.

Thankfully Chromium (and Google Chrome) get this right, so I would suggest people start using it instead.