The train-wreck of not telling users where their session has gone
Ben Eficent
Session' on the History menu has led developers to decide they no
longer need to (a) prompt on exit or (b) restore on startup. Worse,
they are not telling most users that their session has been saved or
what to do to get it back!
The result, if Firefox 4 ships this way, will be huge numbers of users
who currently click [Save and Quit] on exit and get their session back
on restart LOSING THEIR FIRST SESSION after upgrading, as they won't
know that it has been saved or how to get it back. Others will waste
time looking for a solution.
A restore button has been added to the default home page, about:home,
but just about anyone who doesn't use this will be none the wiser.
The people responsible for this decision are aware of the
consequences, but seem perfectly fine with them!
Well I think it's appalling. I have come up with a solution which
could be easily implemented to address this - see
https://bugzilla.mozilla.org/show_bug.cgi?id=635231 - but I'm not
holding my breath! This is because the recent changes detailed above
are part of a wider move towards form over function.
http://groups.google.com/group/mozilla.dev.apps.firefox/browse_thread/thread/23e472ae53aca942#
The more I look into this, the more it seems a case of modal dialogs
offending the sensibilities of the developers – and who cares if a
million users lose session data?
There is also a distorted view of the impact of closing a session
accidentally. Large sessions typically do not restore as they were
before – web pages update – or even get taken down, video caches get
lost, etc, etc.
The quit prompt alone is enough of an issue that people are talking
about making a plug-in just to provide that functionality. But
everyday users – the ones most likely to be affected by accidental
closings – are unlikely to know of such plug-ins should they be
created.
This situation is a scandal waiting to happen; it needs to be
addressed before the release of Firefox 4.0. Users should certainly
have the option of a prompt on exit. But far more importantly, they
need to know how to restore their session!!!
Alex Faaborg
Additionally:
1) users are saving literally hundreds of tabs in Firefox, and then
switching to Chrome because it starts up faster. The start up time
complaints aren't about 1 second versus 2 seconds, there are 1 second versus
30 seconds. Users could simply not save several hundred tabs in Firefox,
but we can't really depend on them always acting rationally.
2) asking users to decide if they need the session in the future doesn't
support undo, what if they change their mind in the future? When the future
is now, they will actually know if they want their session back or not.
-Alex
> _______________________________________________
> dev-apps-firefox mailing list
> dev-apps...@lists.mozilla.org
> https://lists.mozilla.org/listinfo/dev-apps-firefox
>
Ben Eficent
the-box home page. Do other users - I'm guessing the majority - not
count?
I can see there could occasionally be a time the user changed their
mind. But it's more likely the user would sometimes like for their
session NOT to be saved, and that's not now possible, unless they go
through their tabs one by one closing them all.
In most cases, the user is going to click [Save and Quit] if they have
lots of tabs open, [Quit] if they have a few unimportant tabs open or
[Cancel] if they closed the window by mistake (either by clicking the
X or [Ctrl+Q]ing). It would take a seriously careless user to click
[Quit] if they had valuable unsaved tabs.
And, again, if this is all about catering for careless users, isn't
actually telling them where there session has gone kinda important?
One option would be to give the user about:home on startup if they had
a saved session, (and make clicking the button open the session such
that window 1 of the session was the current window, rather than the
really untidy opening in a new window). This is not ideal, and still
leaves the issue of accidental quits (although advanced users can
change a value in about:config to get that back), but at least it
would actually tell all users where their session had gone. It would
also not require any additional translation for the non-English
versions, so would be much quicker to implement.
On Feb 22, 6:45 am, Alex Faaborg <faab...@mozilla.com> wrote:
> A massive button to restore your session is hardly a train wreck.
> Additionally:
>
> 1) users are saving literally hundreds of tabs in Firefox, and then
> switching to Chrome because it starts up faster. The start up time
> complaints aren't about 1 second versus 2 seconds, there are 1 second versus
> 30 seconds. Users could simply not save several hundred tabs in Firefox,
> but we can't really depend on them always acting rationally.
> 2) asking users to decide if they need the session in the future doesn't
> support undo, what if they change their mind in the future? When the future
> is now, they will actually know if they want their session back or not.
>
> -Alex
>
> On Mon, Feb 21, 2011 at 7:52 AM, Ben Eficent <beneficent...@gmail.com>wrote:
>
> > For some inexplicable reason, the handy addition of 'Restore Previous
> > Session' on the History menu has led developers to decide they no
> > longer need to (a) prompt on exit or (b) restore on startup. Worse,
> > they are not telling most users that their session has been saved or
> > what to do to get it back!
>
> > The result, if Firefox 4 ships this way, will be huge numbers of users
> > who currently click [Save and Quit] on exit and get their session back
> > on restart LOSING THEIR FIRST SESSION after upgrading, as they won't
> > know that it has been saved or how to get it back. Others will waste
> > time looking for a solution.
>
> > A restore button has been added to the default home page, about:home,
> > but just about anyone who doesn't use this will be none the wiser.
>
> > The people responsible for this decision are aware of the
> > consequences, but seem perfectly fine with them!
>
> > Well I think it's appalling. I have come up with a solution which
> > could be easily implemented to address this - see
> >https://bugzilla.mozilla.org/show_bug.cgi?id=635231- but I'm not
> > holding my breath! This is because the recent changes detailed above
> > are part of a wider move towards form over function.
>
> >http://groups.google.com/group/mozilla.dev.apps.firefox/browse_thread...
>
> > The more I look into this, the more it seems a case of modal dialogs
> > offending the sensibilities of the developers – and who cares if a
> > million users lose session data?
>
> > There is also a distorted view of the impact of closing a session
> > accidentally. Large sessions typically do not restore as they were
> > before – web pages update – or even get taken down, video caches get
> > lost, etc, etc.
>
> > The quit prompt alone is enough of an issue that people are talking
> > about making a plug-in just to provide that functionality. But
> > everyday users – the ones most likely to be affected by accidental
> > closings – are unlikely to know of such plug-ins should they be
> > created.
>
> > This situation is a scandal waiting to happen; it needs to be
> > addressed before the release of Firefox 4.0. Users should certainly
> > have the option of a prompt on exit. But far more importantly, they
> > need to know how to restore their session!!!
> > _______________________________________________
> > dev-apps-firefox mailing list
> > dev-apps-fire...@lists.mozilla.org
> >https://lists.mozilla.org/listinfo/dev-apps-firefox
Alex Faaborg
the majority of users haven't customized their home page. Depending on that
figure we may need to re-evaluate if we want to introduce a mechanism for
users who have customized their home page (info bar, continuing to ask on
exit). Either way this problem eventually can be resolved with the Home Tab
(basically a Firefox provided app tab that contains additional browser
functionality).
But it's more likely the user would sometimes like for their
> session NOT to be saved, and that's not now possible
>
The best approach for the user in this case is to use private browsing mode
(since it will also not store a lot of things outside of the normal notion
of session, including obscure stuff like the contents of the clipboard, and
adding entries to the recent documents list in the start menu for downloads,
etc.). Additionally we have Clear Recent History for the undo case.
isn't actually telling them where there session has gone kinda important?
>
yep it's very important,and we need to make sure we are making this decision
with good data on the percent of home page customizations (but if it's at
<5% then we can probably take the impact of the problem for a few months
before we switch over to the faster release cycle). For instance, currently
Safari doesn't provide a clear path to restore (menu only), and it works
well enough for most of their users.
-Alex
Alexander Limi
> The button you mention is only available to users who kept the out-of-
> the-box home page. Do other users - I'm guessing the majority - not
> count?
>
Based on what we see in the field, the majority of users do not change their
home page. We could use Test Pilot to verify this if needed, but it seems to
be pretty uncommon for users to do this, anecdotally speaking. If you change
your home page, that's also the exact same point in the preferences where
you can tell Firefox to always restore the previous session on startup. And,
there's an entry to restore your previous session in the History menu.
> I can see there could occasionally be a time the user changed their
> mind. But it's more likely the user would sometimes like for their
> session NOT to be saved, and that's not now possible, unless they go
> through their tabs one by one closing them all.
>
They shouldn't have to think about this at all. If you're talking about
privacy implications, Unless they clear their history, the previous session
was implicitly "saved", just can't be easily restored with a single click of
a button. If you want privacy for a session, use Private Browsing.
> In most cases, the user is going to click [Save and Quit] if they have
> lots of tabs open, [Quit] if they have a few unimportant tabs open or
> [Cancel] if they closed the window by mistake (either by clicking the
> X or [Ctrl+Q]ing). It would take a seriously careless user to click
> [Quit] if they had valuable unsaved tabs.
>
That's not the issue, and that's not how people read that dialog box.
Most people don't read dialog boxes at all, and just answer "yes" and "ok"
to get to the next step. And no amount of wordsmithing and/or scary
colors/icons can make them read instead of skim them.
We see users in the field that see the three-option dialog as "Are you sure
you want to exit?", which is what they have been conditioned to do by other
applications. If they actually read the dialog, they are unsure of what to
do, and always go for the most defensive option ("Save & Quit"), which also
happens to be the one that makes Firefox feel slow to start up, and then
they get annoyed with the browser showing them the 6 NYTimes articles they
read yesterday when they open it the day after. The core of the issue is
that you don't know whether you want to restore your previous session *on
the way out*, but you know if something is missing when you're starting the
browser and think "wait, I want the stuff I was working on yesterday".
In other words, the change makes it possible to not have to answer that
question on the way out — when you're really trying to shut down your
computer, work on something else, etc — and let's you defer that decision
until later, when you know whether you actually want to use that session
with what you're doing right now.
--
Alexander Limi · Firefox UX Team · @limi <http://twitter.com/limi> ·
limi.net
Gijs Kruitbosch
Can we auto-switch users (like myself) who've switched their homepage to
about:blank to this magical new home tab, which I presume is also local?
</selfish>
~ Gijs
Michael Lefevre
> I'll need to check the test pilot data again to be sure, but I believe that
> the majority of users haven't customized their home page.
I imagine you're right (not sure where that might be in test pilot data,
but I found a 2004 study which found that 87% of users hadn't changed
their home page).
>> isn't actually telling them where there session has gone kinda important?
>
> yep it's very important,and we need to make sure we are making this decision
> with good data on the percent of home page customizations (but if it's at
> <5% then we can probably take the impact of the problem for a few months
> before we switch over to the faster release cycle). For instance, currently
> Safari doesn't provide a clear path to restore (menu only), and it works
> well enough for most of their users.
Wouldn't it be better to consider as a percentage of users that might
actually want it? From test pilot data, if I remember correctly, the
huge majority of users only had 2-3 tabs open most of the time. If the
minority of users with dozens of tabs happen to be the same minority
that have changed their home page, then the UI may be rather useless for
the majority of people that actually wanted it (although I guess you
could say that not having the dialog getting in the way is a benefit to
everyone else, who didn't want it anyway...).
Michael
Philip Chee
> On 21/02/2011 21:27, Alex Faaborg wrote:
>> I'll need to check the test pilot data again to be sure, but I believe that
>> the majority of users haven't customized their home page.
>
> I imagine you're right (not sure where that might be in test pilot data,
> but I found a 2004 study which found that 87% of users hadn't changed
> their home page).
I suspect that the vast majority of Firefox users don't even know that
they can change their home page, or that they have any concept of a
"home page" at all. To them it's all "the internet".
Phil
--
Philip Chee <phi...@aleytys.pc.my>, <phili...@gmail.com>
http://flashblock.mozdev.org/ http://xsidebar.mozdev.org
Guard us from the she-wolf and the wolf, and guard us from the thief,
oh Night, and so be good for us to pass.
Robert Kaiser
> Depending on that
> figure we may need to re-evaluate if we want to introduce a mechanism for
> users who have customized their home page (info bar, continuing to ask on
> exit). Either way this problem eventually can be resolved with the Home Tab
> (basically a Firefox provided app tab that contains additional browser
> functionality).
We really should care that we have a solution for those people in 4.0,
and I think Home Tab isn't on the plan for 4.0 any more, right?
> but if it's at
> <5% then we can probably take the impact of the problem for a few months
> before we switch over to the faster release cycle
Well, 5% of the Firefox user base are about 20 million, if I'm right,
and that's about 2.5 the size of my home country (.at) and we manage to
largely dominate Alpine skiing and have some other good notes of our
presence on this earth, so I guess 5% of the Firefox user base, being
that factor more, could also have some significant impact on the world. ;-)
With the size of our community, arguments fast go into impacting
millions (!) of people...
Robert Kaiser
--
Note that any statements of mine - no matter how passionate - are never
meant to be offensive but very often as food for thought or possible
arguments that we as a community needs answers to. And most of the time,
I even appreciate irony and fun! :)
Alexander Limi
<gijskru...@gmail.com>wrote:
> <selfish>
> Can we auto-switch users (like myself) who've switched their homepage to
> about:blank to this magical new home tab, which I presume is also local?
> </selfish>
>
That's a very good point. When upgrading, we should probably do that, and
let the ones that *really* want a blank page turn it back on. I was in the
same boat myself; I had about:blank as my default since the remote page was
slow, but now that it's local, I've kept about:home.
It might be too late in the cycle for this change, but I'll ask around and
see if it can be done without introducing too much risk.
Mike Ratcliffe
> On Mon, Feb 21, 2011 at 2:27 PM, Gijs Kruitbosch
To be honest about:home searches google.com and I live in Switzerland
so it redirects to google.ch ... because I am British I want results
for google.co.uk so I am forced to change the home page. I have to
admit that I would prefer not to do so because changing it means less
cash for funding mozilla.corp
I hate to state the obvious but could the session restore button not
overlay any page?
Ben Eficent
that first restored session after upgrading. As Robert Kaiser points
out, even 5% is a lot of people, and 13% (Michael Lefevre) = several
countries'-worth!
Perhaps changing the home page of those users not using about:home is
all there's time for. But that's not without issue: if done at time
of upgrade they might think it a glitch (there's no button if there's
no session to restore) and revert. And some users share profiles.
But it would still represent a loss of functionality. How about this
dialog on exiting a saveable session:
[Title:] Quit Firefox
Do you want Firefox to restore your session automatically next time?
[Yes*] [No] [Cancel]
* default
On [Yes], Firefox would behave on startup as if the home page was set
to 'Show my windows and tabs from last time'. However, the preference
value wouldn't change, so if they either quit an empty session or
answered 'No' they'd get their normal home page. There's no loss of
session data, so the user can change their mind, and if they have
about:home they'll see the big button reminding them of this. And, of
course, if they quit accidentally they have the option to cancel.
I'm suggesting no 'Do not ask next time' option: less clutter, and the
existence of the dialog is helping the user. Plus, they have no easy
way of getting it back. (Advanced users could change the about:config
value, in which case Firefox would assume 'No' re restoring session.)
This restores the session into the first window rather than opening a
new one - a much cleaner solution. It provides all users the
functionality they currently enjoy, but with an 'undo' if they change
their minds about wanting the session, plus, of course, the ability to
restore ad hoc. In future, the visual aspects of the dialog could be
smartened up.
On Feb 22, 8:27 am, Alex Faaborg <faab...@mozilla.com> wrote:
> I'll need to check the test pilot data again to be sure, but I believe that
> the majority of users haven't customized their home page. Depending on that
> figure we may need to re-evaluate if we want to introduce a mechanism for
> users who have customized their home page (info bar, continuing to ask on
> exit). Either way this problem eventually can be resolved with the Home Tab
> (basically a Firefox provided app tab that contains additional browser
> functionality).
>
> But it's more likely the user would sometimes like for their
>
> > session NOT to be saved, and that's not now possible
>
> The best approach for the user in this case is to use private browsing mode
> (since it will also not store a lot of things outside of the normal notion
> of session, including obscure stuff like the contents of the clipboard, and
> adding entries to the recent documents list in the start menu for downloads,
> etc.). Additionally we have Clear Recent History for the undo case.
>
> isn't actually telling them where there session has gone kinda important?
>
>
>
> yep it's very important,and we need to make sure we are making this decision
> with good data on the percent of home page customizations (but if it's at
> <5% then we can probably take the impact of the problem for a few months
> before we switch over to the faster release cycle). For instance, currently
> Safari doesn't provide a clear path to restore (menu only), and it works
> well enough for most of their users.
>
> -Alex
>
> > > >https://bugzilla.mozilla.org/show_bug.cgi?id=635231-but I'm not
> > dev-apps-fire...@lists.mozilla.org
> >https://lists.mozilla.org/listinfo/dev-apps-firefox
Jesper Kristensen
> On Mon, Feb 21, 2011 at 2:27 PM, Gijs Kruitbosch
> <gijskru...@gmail.com>wrote:
>
>> <selfish>
>> Can we auto-switch users (like myself) who've switched their homepage to
>> about:blank to this magical new home tab, which I presume is also local?
>> </selfish>
>>
>
> That's a very good point. When upgrading, we should probably do that, and
> let the ones that *really* want a blank page turn it back on. I was in the
> same boat myself; I had about:blank as my default since the remote page was
> slow, but now that it's local, I've kept about:home.
Isn't this one of those critical dataloss cases where this feature
should be backed out if it cannot be fixed before Firefox 4, and not
just for about:blank but for any home page?
By the way, most of the idea of getting rid of this question is kind of
lost, as Firefox now just asks a different question. Couldn't we make
Firefox not ask any question at all, as the nightlies did at some point?
Den 21-02-2011 22:40, Alexander Limi skrev:
> The core of the issue is
> that you don't know whether you want to restore your previous session *on
> the way out*, but you know if something is missing when you're starting the
> browser and think "wait, I want the stuff I was working on yesterday".
This is quite the opposite for me. When I reopen Firefox, I can never
remember if I should restore my previous session or not. I use session
restore quite rarely. Probably a few times a month. It only happens when
I was doing something I need to finish, and I have to close my computer.
When I start my computer hours later, I have forgot that I was in the
middle of something, and that I would have to restore my session. The
first time this happened to me after the change, I didn't remember to
restore my session, and I had to explain to a not quite happy guy why I
hadn't delivered what I had promised. I have now learned to deal with
the new behavior. In Firefox 3.6 I would choose to save the session when
I quit Firefox. In Firefox 4 I now write a yellow Post-IT note saying
"remember to restore your session", and put it on the middle of the
screen, so that I cannot miss it when I start my computer next time.
Mike Beltzner
> Isn't this one of those critical dataloss cases where this feature should be backed out if it cannot be fixed before Firefox 4, and not just forabout:blank but for any home page?
Not at all. There is no dataloss, in fact, there is extra data protection. Users can always restore their session from the History Menu, even if about:home isn't set as a default.
Fundamentally, and sadly, this newsgroup and the participants in this thread are not at all representative of our predominant user base. We all use many tabs and want our sessions to stay with us most of the time, but that's not really the case for most users. While I was initially unconvinced about the change, I have seen more users *discover* the Session Restore feature via the new design than those who discovered it through the Quit dialog.
The title of this thread is pejorative and assumes facts not in evidence. I'm pretty confident that this will play out quite well. Also, the ability to always restore one's session remains, as ever, in Options.
cheers,
mike
Christian Legnitto
That would work for all homepage values and the user can ignore and browse, dismiss, or take action.
Christian
Mike Beltzner
> That would work for all homepage values and the user can ignore and browse, dismiss, or take action.
And be seriously annoying for the majority of users who don't care to restore their session, and a late UI change. No, I don't think this is a good idea.
We're all working on this presumption that this is a feature which all users want: where's the evidence supporting that? I understand that it's something people who are participating in *this* thread want, and yes, I'm sure "all [your] friends" want it, too, but I believe that the audience skew there is relentlessly techie.
I wish I had better data supporting my position, of course, and others have suggested using Test Pilot. What I can offer you is that this issue doesn't really pop up all that much amongst our beta users:
cheers,
mike
Christian Legnitto
These people were also savvy enough to change their homepage (one is igoogle, one is TMZ, one is CNN, etc) so they wouldn't see the about:home button.
Not data, but hearing those opinions was the reason I was keeping an eye on the discussion.
I think anyone who is technical enough to post here is not representative of the majority of our users so personal preferences are generally not useful FWIW.
Christian
Christian Legnitto
Benjamin Smedberg
>
> Fundamentally, and sadly, this newsgroup and the participants in this thread are not at all representative of our predominant user base. We all use many tabs and want our sessions to stay with us most of the time, but that's not really the case for most users. While I was initially unconvinced about the change, I have seen more users *discover* the Session Restore feature via the new design than those who discovered it through the Quit dialog.
changed, both from the perspective of existing users as well as new
users? I personally checked the box a while ago to always restore my
session, and I haven't noticed any changes in behavior which affect me.
The few people who I've helped use Firefox haven't "changed their
homepage" as far as I know, but they also typically rely on Firefox to
restore their session automatically, because the only time they shut
down their browser is when they crash or need to restart for updates. If
suddenly their tabs didn't recover in *this* case, I think they would be
upset. I think several of them have clicked the "always restore" button
in that dialog which appears.
--BDS
Mike Beltzner
> On 2/22/2011 10:34 AM, Mike Beltzner wrote:
>>
>> Fundamentally, and sadly, this newsgroup and the participants in this thread are not at all representative of our predominant user base. We all use many tabs and want our sessions to stay with us most of the time, but that's not really the case for most users. While I was initially unconvinced about the change, I have seen more users *discover* the Session Restore feature via the new design than those who discovered it through the Quit dialog.
> This thread is confusing to me: can somebody describe what has actually changed, both from the perspective of existing users as well as new users? I personally checked the box a while ago to always restore my session, and I haven't noticed any changes in behavior which affect me.
Firefox 3.0 -> 3.6 : on quit, users with multiple windows & tabs would be asked if they wanted to "Save and Quit", "Quit" or "Cancel" with an option to never be asked again. Selecting "Save and Quit" would result in the next browser start to initiate a session restore.
Firefox 4: on quit, users with multiple windows & tabs are asked to confirm that they wish to close all windows and tabs. When the browser is next started up, the History Menu will contain an option to "Restore Previous Session". Additionally, about:home will look like this: http://grab.by/95BX with a button that can be clicked to restore the user's previous session.
The UX decision here, as Faaborg put it earlier, was to move the decision to the time when it would be most relevant (ie: "Oh, I want to get those tabs back from last time") as well as to speed up startup time by not invoking session restore automatically. The option to always restore the session remains in preferences: http://grab.by/95C1
> The few people who I've helped use Firefox haven't "changed their homepage" as far as I know, but they also typically rely on Firefox to restore their session automatically, because the only time they shut down their browser is when they crash or need to restart for updates. If suddenly their tabs didn't recover in *this* case, I think they would be upset. I think several of them have clicked the "always restore" button in that dialog which appears.
Shutdowns which are forced on the browser (in the case of an overnight Windows System Update or application crash) will always have their sessions restored, yes. Same with restarts (for add-on updates, etc)
cheers,
mike
Benjamin Smedberg
>
> Firefox 4: on quit, users with multiple windows& tabs are asked to confirm that they wish to close all windows and tabs. When the browser is next started up, the History Menu will contain an option to "Restore Previous Session". Additionally, about:home will look like this: http://grab.by/95BX with a button that can be clicked to restore the user's previous session.
On nightlies, I don't see any prompt on shutdown in a clean profile with
multiple windows and tabs open. It appears that Firefox saves my session
on exit without prompting, which in general seems like a good idea
(unless I have a half filled-in form, or a page which cannot be restored
from cache).
>
> The UX decision here, as Faaborg put it earlier, was to move the decision to the time when it would be most relevant (ie: "Oh, I want to get those tabs back from last time") as well as to speed up startup time by not invoking session restore automatically. The option to always restore the session remains in preferences: http://grab.by/95C1
restore my prior session is saved in preferences as "Startup | When
Minefield starts: Show my windows and tabs from last time"?
> Shutdowns which are forced on the browser (in the case of an overnight Windows System Update or application crash) will always have their sessions restored, yes. Same with restarts (for add-on updates, etc)
I think the behavior we have in Fx4 appears to be fine. I do wonder if
discouraging people from automatically restoring their session is the
best long-term plan, but until we have e10s and can make the session
restore not cause the browser to become unresponsive, it seems like the
clearly better plan for Fx4.
--BDS
Mike Beltzner
> On nightlies, I don't see any prompt on shutdown in a clean profile with multiple windows and tabs open. It appears that Firefox saves my session on exit without prompting, which in general seems like a good idea (unless I have a half filled-in form, or a page which cannot be restored from cache).
For you, that's because ...
> So the decision I made in Firefox 3.6 "always remember my decision" to restore my prior session is saved in preferences as "Startup | When Minefield starts: Show my windows and tabs from last time"?
... yes. Note that by default, users would be prompted to ensure they meant to exit Firefox if multiple tabs are open (unless they disabled the warning in their preferences, here: http://grab.by/95Di). Further, they would be offered session restore on startup via: about:home.
> I think the behavior we have in Fx4 appears to be fine. I do wonder if discouraging people from automatically restoring their session is the best long-term plan, but until we have e10s and can make the session restore not cause the browser to become unresponsive, it seems like the clearly better plan for Fx4.
That was the thinking, yes. Also, as Faaborg mentioned, in the future the idea is to move the home button into a permanent "App Tab" which could be used for browser-to-user notifications like session restore, application updates, etc.
cheers,
mike
Benjamin Smedberg
> On 2011-02-22, at 11:45 AM, Benjamin Smedberg wrote:
>
>> On nightlies, I don't see any prompt on shutdown in a clean profile with multiple windows and tabs open. It appears that Firefox saves my session on exit without prompting, which in general seems like a good idea (unless I have a half filled-in form, or a page which cannot be restored from cache).
> For you, that's because ...
>
>> So the decision I made in Firefox 3.6 "always remember my decision" to restore my prior session is saved in preferences as "Startup | When Minefield starts: Show my windows and tabs from last time"?
> ... yes. Note that by default, users would be prompted to ensure they meant to exit Firefox if multiple tabs are open (unless they disabled the warning in their preferences, here: http://grab.by/95Di). Further, they would be offered session restore on startup via: about:home.
mac with 2 windows/4 tabs. I'm pretty certain that we never prompt by
default on shutdown now.
--BDS
EE
> I'll need to check the test pilot data again to be sure, but I believe that
> the majority of users haven't customized their home page. Depending on that
> figure we may need to re-evaluate if we want to introduce a mechanism for
> users who have customized their home page (info bar, continuing to ask on
> exit). Either way this problem eventually can be resolved with the Home Tab
> (basically a Firefox provided app tab that contains additional browser
> functionality).
>
home page? I have not seen any other browser that did not let the user
change the home page.
> But it's more likely the user would sometimes like for their
>> session NOT to be saved, and that's not now possible
>
So it will not be possible to clear private data any more? What happens
if the cache gets full? Browsers slow down when that happens. It
should be possible at least to do that. And what about history? I
would not want to keep that forever either. With Windows, I would then
be forced to use something like CCleaner just to clean things up, or use
Purity on Mac OS?
> The best approach for the user in this case is to use private browsing mode
> (since it will also not store a lot of things outside of the normal notion
> of session, including obscure stuff like the contents of the clipboard, and
> adding entries to the recent documents list in the start menu for downloads,
> etc.). Additionally we have Clear Recent History for the undo case.
>
Does that mean that you could clear today's history but not yesterday's,
or you can clear history for one minute ago but not one hour ago?
EE
Why would you assume that most users want to restore their previous
sessions? If I want to restore a session later, I know how to set that,
but 99.9% of the time, I want to clear out a session completely, and
start with a clean slate the next time.
Paul O’Shannessy
> Firefox 3.0 -> 3.6 : on quit, users with multiple windows & tabs would be asked if they wanted to "Save and Quit", "Quit" or "Cancel" with an option to never be asked again. Selecting "Save and Quit" would result in the next browser start to initiate a session restore.
>
This isn't quite right... We distinguish between explicitly quitting
(cmd-q, file > exit) and closing the last window.
By default:
* Fire | Exit results in *no dialog*. Your session is not restored,
but it is saved and available in the history menu and on the default
home page
* Closing the last window (Windows & Linux) with multiple tabs results
in the "you're closing multiple tabs" warning (controlled via visible
pref in tabs > warn when closing multiple tabs)
You can restore 3.0 -> 3.6 behavior very easily: set
browser.showQuitWarning to true.
We carry over the behavior that if your session will be restored on
startup (via prefs > startup > show windows & tabs), then we will not
prompt when quitting, regardless of the other settings.
Boris Zbarsky
>> But it's more likely the user would sometimes like for their
>>> session NOT to be saved, and that's not now possible
>>
> So it will not be possible to clear private data any more?
Uh... it'll be quite possible. What made you think otherwise?
-Boris
Boris Zbarsky
> Why would you assume that most users want to restore their previous
> sessions?
The point is that most users will _sometimes_ (but not always) want to
restore their previous sessions.
Since we don't know what those times will be, the idea is to always make
it possible to do it, but not do it by default; this allows users to do
it as desired.
-Boris
EE
> Wouldn't it be better to consider as a percentage of users that might
> actually want it? From test pilot data, if I remember correctly, the
> huge majority of users only had 2-3 tabs open most of the time. If the
> minority of users with dozens of tabs happen to be the same minority
> that have changed their home page, then the UI may be rather useless for
> the majority of people that actually wanted it (although I guess you
> could say that not having the dialog getting in the way is a benefit to
> everyone else, who didn't want it anyway...).
>
> Michael
That second "if" is a big one, and probably not true. Why would one
assume that people who have changed the home page are the same ones that
use dozens of tabs all the time? I may have gotten up to two dozen tabs
occasionally, but I do not keep them very long. I have changed my home
page on every browser that I have ever used. I would hate to be stuck
with some home page that I dislike.
Ron Hunter
Some people may never change from whatever default they are provided
with when they load the software, but I would rather think that Firefox
users are a bit more independent than that. My homepage is Google, and
NOT the one filtered through Mozilla!
Ben Eficent
me. See comment by Paul O’Shannessy for confirmation that all
versions of Firefox 3 prompt to save your session by default and
restore it immediately on startup. And see comment by Michael Lefevre
for confirmation that about 13% of users have a home page other than
about:home.
And although a minority of users will be relying on having their
session restored, within the group not using about:home, that minority
could easily number into the millions. Something needs to be done to
tell them where their session has gone. Or, better still, a change
that will stop them even worrying about it, as my sollution below
does. It also means not having to interfere with their home page
setting.
Since users are expecting a prompt on exit, and since having one saves
accidental exit, I'm suggesting a similar prompt but one that asks
whether you want your session restored immediately on retart. The
session would always be saved. Fine for this to default to [No] if
the majority of users closing a multiple tab session don't want it
immediately restored on restart.
In click terms, this is an improvement even for users with about:home,
who would need one click of the big button to restore, two to switch
to the second window and one to close the first window at the end.
With my quit prompt there is one click on [Yes] to tell Firefox you
want your session back on restart and that's it.
Michael Lefevre
> We're all working on this presumption that this is a feature which
> all users want: where's the evidence supporting that? I understand
> that it's something people who are participating in *this* thread
> want, and yes, I'm sure "all [your] friends" want it, too, but I
> believe that the audience skew there is relentlessly techie.
I very much doubt all users want it, and based on that, the UI change is
reasonable. The problem is that some of the minority of people who do
want it may not be able to find it.
> I wish I had better data supporting my position, of course, and
> others have suggested using Test Pilot. What I can offer you is that
> this issue doesn't really pop up all that much amongst our beta
> users:
>
> https://input.mozilla.com/en-US/beta/search?q=restore&product=firefox&version=4.0b11&date_start=&date_end=&sentiment=sad
That's beta 11 feedback though. Complaints about it have also tailed off
over in the MozillaZine forums. I imagine the issue is those moving up
from 3.6 who don't get the prompt, don't have about:home and so go "WTF?
Firefox has lost my data!!". Hopefully they then either find the history
menu option themselves, or ask a friend, or do a web search, or ask SUMO
or MozillaZine or whatever, and find out how it works. After that they
hopefully either get used to doing it differently or they use the hidden
pref.
I guess that not many people read release notes, but if there's any way
of getting the information about the change to more people before they
are unpleasantly surprised by it, I think that would be good.
Michael
Michael Lefevre
> On 2011/02/21 15:33, Michael Lefevre wrote:
>> Wouldn't it be better to consider as a percentage of users that
>> might actually want it? From test pilot data, if I remember
>> correctly, the huge majority of users only had 2-3 tabs open most
>> of the time. If the minority of users with dozens of tabs happen to
>> be the same minority that have changed their home page, then the UI
>> may be rather useless for the majority of people that actually
>> wanted it (although I guess you could say that not having the
>> dialog getting in the way is a benefit to everyone else, who didn't
>> want it anyway...).
>
> assume that people who have changed the home page are the same ones
> that use dozens of tabs all the time? I may have gotten up to two
> dozen tabs occasionally, but I do not keep them very long.
Er... I think you're reading my post out of context. I wasn't
necessarily trying to suggest that "if" was actually the case,
only that the opposite assumption may not be the case either.
> I have changed my home page on every browser that I have ever used. I
> would hate to be stuck with some home page that I dislike.
I'm sure lots of people would hate that. But I don't think anyone
suggested that you should be stuck with any particular home page, so
there is no need to worry...
Michael
Mike Beltzner
to this conversation that will progress it further in either direction.
I'm sorry that you believe that this will be a "train wreck." I suppose
only time will tell and you'll have the ability to say that you warned
us, if it turns out that this ends up being a major issue for users. I
don't believe that it will be, and in fact, believe that more people
will discover the feature this way.
(Benjamin and Paul are having a productive sideline about dialogs that
appear in various application close scenarios - that should continue!)
cheers,
mike
Ben Eficent
not on the default home page don't matter. This has been a
frustrating process with, it seems, a depressing outcome. The 'train-
wreck' analogy is suitable because Firefox 4 beta is now a freight
train racing towards the break in the line.
Bad me for pointing out the issue in various forums and coming up with
a simple solution that would help all users. What was I thinking?!
Minor correction: while the click count (4) for restoring from the big
button was correct, the order is slightly different: one on the
button, two to switch back to the original window and one to close it
so it isn't restored next time. Not doing this, your window count
would increase every time you restored. For the quit on exit, the
click count (1) is correct.
Minor improvement to the quit prompt which seems highly unlikely to
happen ("your" > "this"):
[Title:] Quit Firefox
Do you want Firefox to restore this session automatically next time?
[Yes] [No*] [Cancel]
* default
The vast majority of people who lose data - for instance the previous
behaviour of Firefox not restoring the "Oops, this is embarrassing"
session restore tab under some circumstances - do not let you know
about it. They just carry on the best they can. I didn't lose data,
but saw how millions could. Shame that realisation doesn't bother
you. Or maybe you don't believe it's a problem because you still
misunderstand the default behaviour of the entire Firefox 3 line!
One more nougat that will doubtless be lost: the slow startup when
users have over 9 tabs to be restored could be easily addressed by
using the cached version where available (and putting (cached) in the
page's title).
EE
I think that is the ideal situation. Why change that?
EE
If you cannot clear your session, how can you clear your history?
You cannot restore your session with all the history cleared, right?
Boris Zbarsky
The setup I described is the new setup, not the old setup.... The old
setup requires users to always deal with a dialog that they usually
don't want to deal with. The new one allows them to restore the session
when desired but otherwise stays out of the way.
-Boris
Boris Zbarsky
> If you cannot clear your session, how can you clear your history?
You're confusing session history and global history. You can't clear
the former (the back/forward button stuff) manually right now; that's
not changing. You can clear the latter any time you feel like; that's
not changing either.
-Boris
Ben Eficent
I disagree. The old dialog was helpful for the user. If they clicked
[Save & Quit], the session would be restored on restart, where under
the new system they may forget they had a session waiting to be
restored, esp if they don't use about:home. (Someone mentioned
sticking a Post-It to their screen to remind them!) They also had the
option to [Cancel] if they'd accidentally closed the last window. And
closing a large session is significant - the restore can easily use
50MB of download(!), obviously taking ages, and what you get back
isn't going to be identical. Also, Fx has known problems restoring
large sessions.
Granted, a way of changing your mind is desirable, but with the new
version of the dialog I'm suggesting (see 5 comments back) they'd have
that, as all sessions would be saved. They'd also get a cleaner
session restore with 3 fewer clicks, as detailed in that comment.
Ben Eficent
> home page. We could use Test Pilot to verify this if needed, but it seems to
> be pretty uncommon for users to do this, anecdotally speaking. If you change
> your home page, that's also the exact same point in the preferences where
> you can tell Firefox to always restore the previous session on startup. And,
> there's an entry to restore your previous session in the History menu.
This has been done (see comment by Michael Lefevre) and the estimate
is 50,000,000 users who are using a custom home page. That's quite a
few. The trouble with 'Show my windows and tabs from last time' is
that it is the behaviour for every time you start Firefox. There's
been a lot of talk here about the users who sometimes want to restore
their old session and sometimes not. Well, when they don't they'll
want to see their home page.
My revised solution, where all sessions are saved and the user is
asked whether they want their session restored on restart (defaulting
to 'No') provides this functionality. Firefox behaves as if
browser.startup.page was set to 3 ('Show my windows and tabs from last
time') on restart if they clicked [Yes], or shows their home page if
they clicked [No]. And, of course, they retain the ability to
[Cancel] if they closed the last window by mistake - significant for
various reasons, not least being that there are known problems with
Firefox restoring large sessions. They can still change their mind
and restore ad hoc.
> If you're talking about
> privacy implications, Unless they clear their history, the previous session
> was implicitly "saved", just can't be easily restored with a single click of
> a button. If you want privacy for a session, use Private Browsing.
Privacy is not my chief concern, but there _is_ a bit of an issue
there. If user 2 had a nose of user 1's history items, that would be
evident, unless they went to extreme lengths. With one-click session
restore, user 2 could, whether on purpose or not, load up user 1's
entire session. They could then exit and it would be saved again,
giving user 1 no way of knowing. Like I said, it's not my major
concern, but I imagine there'll be calls for a way of killing the
session. (Obviously, anyone with serious privacy concerns should be
using Private Browsing and not sharing their profile.)
> Most people don't read dialog boxes at all, and just answer "yes" and "ok"
> to get to the next step. And no amount of wordsmithing and/or scary
> colors/icons can make them read instead of skim them.
I agree that that's an issue, and that the ability to undo is
desirable. But as you say, users are used to making Yes/No choices on
exit. My new solution just gives them a slightly different choice:
"Do you want Firefox to restore this session automatically next time?
[Yes] [No*] [Cancel] *default There's no 'Do not ask next time', but
the user could uncheck the 'Warn me when closing multiple tabs'
option, in which case Firefox would not restore the session
automatically (unless their startup page was set to 'Show my windows
and tabs from last time': I suggest the quit dialog always be shown if
'Warn me when closing multiple tabs' is checked).
As it now defaults to 'No', your path of least resistance point is
observed, and they won't get slowed down by those 6 NYT articles on
restart.
They also get a cleaner session restore by clicking 'Yes' on exit: 3
fewer clicks (depending on OS), because they don't have to switch back
to the parent window and close it. (If you click the big button on
about:home the saved session opens in a new window, as it does if you
click the History menu item, which adds a 4th click.)
So I think my new solution addresses all your points, while retaining
the functionality/behaviour users of default Fx3 enjoy. It would
appeal to in-a-hurry, click-happy users because it saves them from
accidental quits. And it would satisfy power users because they get a
better session restore experience (vs Fx4b11). It also, of course,
notifies ALL users where their session is, thus addressing this
thread's key concern.
And in terms of implementation, what could be more straightforward
than a Yes/No/Cancel box?
On Feb 22, 8:40 am, Alexander Limi <l...@mozilla.com> wrote:
> On Mon, Feb 21, 2011 at 12:22 PM, Ben Eficent <beneficent...@gmail.com>wrote:
>
> > The button you mention is only available to users who kept the out-of-
> > the-box home page. Do other users - I'm guessing the majority - not
> > count?
>
> Based on what we see in the field, the majority of users do not change their
> home page. We could use Test Pilot to verify this if needed, but it seems to
> be pretty uncommon for users to do this, anecdotally speaking. If you change
> your home page, that's also the exact same point in the preferences where
> you can tell Firefox to always restore the previous session on startup. And,
> there's an entry to restore your previous session in the History menu.
>
> > I can see there could occasionally be a time the user changed their
> > mind. But it's more likely the user would sometimes like for their
> > session NOT to be saved, and that's not now possible, unless they go
> > through their tabs one by one closing them all.
>
> They shouldn't have to think about this at all. If you're talking about
> privacy implications, Unless they clear their history, the previous session
> was implicitly "saved", just can't be easily restored with a single click of
> a button. If you want privacy for a session, use Private Browsing.
>
> > In most cases, the user is going to click [Save and Quit] if they have
> > lots of tabs open, [Quit] if they have a few unimportant tabs open or
> > [Cancel] if they closed the window by mistake (either by clicking the
> > X or [Ctrl+Q]ing). It would take a seriously careless user to click
> > [Quit] if they had valuable unsaved tabs.
>
> That's not the issue, and that's not how people read that dialog box.
>
> Most people don't read dialog boxes at all, and just answer "yes" and "ok"
> to get to the next step. And no amount of wordsmithing and/or scary
> colors/icons can make them read instead of skim them.
>
> We see users in the field that see the three-option dialog as "Are you sure
> you want to exit?", which is what they have been conditioned to do by other
> applications. If they actually read the dialog, they are unsure of what to
> do, and always go for the most defensive option ("Save & Quit"), which also
> happens to be the one that makes Firefox feel slow to start up, and then
> they get annoyed with the browser showing them the 6 NYTimes articles they
> read yesterday when they open it the day after. The core of the issue is
> that you don't know whether you want to restore your previous session *on
> the way out*, but you know if something is missing when you're starting the
> browser and think "wait, I want the stuff I was working on yesterday".
>
> In other words, the change makes it possible to not have to answer that
> question on the way out — when you're really trying to shut down your
> computer, work on something else, etc — and let's you defer that decision
> until later, when you know whether you actually want to use that session
> with what you're doing right now.
>
> --
> Alexander Limi · Firefox UX Team · @limi <http://twitter.com/limi> ·
> limi.net
EE
I am able to do that right now. If I want to keep my session so that I
can go back to it, I tell the browser to "Save and Quit" and do not
clear the private data. That will result in my previous session being
restored when I start the browser again.
Ben Eficent
(13%) could be a vast under-estimate for the following reasons:
1) Some Test Pilot studies are only available in English, so many
users who have changed their home page believing it's not giving them
a localised Google will not be represented.
2) A high proportion of pilots (users) are likely to have English as
their first language, and a high proportion of those are likely to
reside in North America, so these are less likely to have a problem
with the Google version.
3) Anecdotal accounts are unhelpful because, again, most will be from
US users.
4) A significant number of pilots will be using a separate copy of
Firefox for test purposes, especially the ones testing the more
frequent releases. These users are less likely to customise this test
copy.
Certainly, I was surprised the figure was so low. Well, globally it
could be MUCH higher (as if 52m users exposed to data loss wasn't
enough to warrant attention!)
I appreciate that it's now very late in the day, and Last String was
theoretically two weeks ago. I'm also aware of the serious security
issue not addressed in Fx3. But I would point out that my solution is
the only watertight one that's been suggested so far, and we're
talking 4 strings, of which 3 barely count. So I would still urge
this route. As you know, I'm not a fan of the homepage-changing
option, but at least it would cover the majority of users, so better
than nothing. Would appreciate an update on developments.
Ben Eficent
> with good data on the percent of home page customizations (but if it's at
> <5% then we can probably take the impact of the problem for a few months
> before we switch over to the faster release cycle).
No - the decision is NOT being made on good data: Michael Lefevre's
estimate of 13% affected is likely a HUGE underestimate due to much of
the Test Pilot data being in English only and weighted towards US
users who do not tend to have a problem with the Google box pointing
to google.com (or seeming to) rather than their own version. Full
details here: http://groups.google.com/group/mozilla.dev.apps.firefox/msg/b10164f573606818
Whatever's done needs to be done at 4.0, as it's that first session
that's most in danger of being lost. I have refined my solution to
take into account the desire to streamline, as well as the different
ways people want to restore and be notified:
https://bugzilla.mozilla.org/show_bug.cgi?id=636777#c2
Would just like to say what an improvement I think showing link
locations in the new transitory status area is. Even switches to the
right if the mouse would obscure it - very nice!
On Feb 22, 8:27 am, Alex Faaborg <faab...@mozilla.com> wrote:
> I'll need to check the test pilot data again to be sure, but I believe that
> the majority of users haven't customized their home page. Depending on that
> figure we may need to re-evaluate if we want to introduce a mechanism for
> users who have customized their home page (info bar, continuing to ask on
> exit). Either way this problem eventually can be resolved with the Home Tab
> (basically a Firefox provided app tab that contains additional browser
> functionality).
>
> But it's more likely the user would sometimes like for their
>
> > session NOT to be saved, and that's not now possible
>
> The best approach for the user in this case is to use private browsing mode
> (since it will also not store a lot of things outside of the normal notion
> of session, including obscure stuff like the contents of the clipboard, and
> adding entries to the recent documents list in the start menu for downloads,
> etc.). Additionally we have Clear Recent History for the undo case.
>
> isn't actually telling them where there session has gone kinda important?
>
>
>
> yep it's very important,and we need to make sure we are making this decision
> with good data on the percent of home page customizations (but if it's at
> <5% then we can probably take the impact of the problem for a few months
> before we switch over to the faster release cycle). For instance, currently
> Safari doesn't provide a clear path to restore (menu only), and it works
> well enough for most of their users.
>
> -Alex
>
> On Mon, Feb 21, 2011 at 12:22 PM, Ben Eficent <beneficent...@gmail.com>wrote:
>
>
>
>
>
>
>
> > The button you mention is only available to users who kept the out-of-
> > the-box home page. Do other users - I'm guessing the majority - not
> > count?
>
> > I can see there could occasionally be a time the user changed their
> > mind. But it's more likely the user would sometimes like for their
> > session NOT to be saved, and that's not now possible, unless they go
> > through their tabs one by one closing them all.
>
> > In most cases, the user is going to click [Save and Quit] if they have
> > lots of tabs open, [Quit] if they have a few unimportant tabs open or
> > [Cancel] if they closed the window by mistake (either by clicking the
> > X or [Ctrl+Q]ing). It would take a seriously careless user to click
> > [Quit] if they had valuable unsaved tabs.
>
> > And, again, if this is all about catering for careless users, isn't
> > actually telling them where there session has gone kinda important?
>
> > One option would be to give the user about:home on startup if they had
> > a saved session, (and make clicking the button open the session such
> > that window 1 of the session was the current window, rather than the
> > really untidy opening in a new window). This is not ideal, and still
> > leaves the issue of accidental quits (although advanced users can
> > change a value in about:config to get that back), but at least it
> > would actually tell all users where their session had gone. It would
> > also not require any additional translation for the non-English
> > versions, so would be much quicker to implement.
>
> > On Feb 22, 6:45 am, Alex Faaborg <faab...@mozilla.com> wrote:
> > > A massive button to restore your session is hardly a train wreck.
> > > Additionally:
>
> > > 1) users are saving literally hundreds of tabs in Firefox, and then
> > > switching to Chrome because it starts up faster. The start up time
> > > complaints aren't about 1 second versus 2 seconds, there are 1 second
> > versus
> > > 30 seconds. Users could simply not save several hundred tabs in Firefox,
> > > but we can't really depend on them always acting rationally.
> > > 2) asking users to decide if they need the session in the future doesn't
> > > support undo, what if they change their mind in the future? When the
> > future
> > > is now, they will actually know if they want their session back or not.
>
> > > -Alex
>
> > > On Mon, Feb 21, 2011 at 7:52 AM, Ben Eficent <beneficent...@gmail.com
> > >wrote:
>
> > > > For some inexplicable reason, the handy addition of 'Restore Previous
> > > > Session' on the History menu has led developers to decide they no
> > > > longer need to (a) prompt on exit or (b) restore on startup. Worse,
> > > > they are not telling most users that their session has been saved or
> > > > what to do to get it back!
>
> > > > The result, if Firefox 4 ships this way, will be huge numbers of users
> > > > who currently click [Save and Quit] on exit and get their session back
> > > > on restart LOSING THEIR FIRST SESSION after upgrading, as they won't
> > > > know that it has been saved or how to get it back. Others will waste
> > > > time looking for a solution.
>
> > > > A restore button has been added to the default home page, about:home,
> > > > but just about anyone who doesn't use this will be none the wiser.
>
> > > > The people responsible for this decision are aware of the
> > > > consequences, but seem perfectly fine with them!
>
> > > > Well I think it's appalling. I have come up with a solution which
> > > > could be easily implemented to address this - see
> > > >https://bugzilla.mozilla.org/show_bug.cgi?id=635231-but I'm not
> > > > holding my breath! This is because the recent changes detailed above
> > > > are part of a wider move towards form over function.
>
> > > >http://groups.google.com/group/mozilla.dev.apps.firefox/browse_thread.
> > ..
>
> > > > The more I look into this, the more it seems a case of modal dialogs
> > > > offending the sensibilities of the developers – and who cares if a
> > > > million users lose session data?
>
> > > > There is also a distorted view of the impact of closing a session
> > > > accidentally. Large sessions typically do not restore as they were
> > > > before – web pages update – or even get taken down, video caches get
> > > > lost, etc, etc.
>
> > > > The quit prompt alone is enough of an issue that people are talking
> > > > about making a plug-in just to provide that functionality. But
> > > > everyday users – the ones most likely to be affected by accidental
> > > > closings – are unlikely to know of such plug-ins should they be
> > > > created.
>
> > > > This situation is a scandal waiting to happen; it needs to be
> > > > addressed before the release of Firefox 4.0. Users should certainly
> > > > have the option of a prompt on exit. But far more importantly, they
> > > > need to know how to restore their session!!!
> > > > _______________________________________________
> > > > dev-apps-firefox mailing list
> > > > dev-apps-fire...@lists.mozilla.org
> > > >https://lists.mozilla.org/listinfo/dev-apps-firefox
>
> > _______________________________________________
> > dev-apps-firefox mailing list
> > dev-apps-fire...@lists.mozilla.org
> >https://lists.mozilla.org/listinfo/dev-apps-firefox
Michael Lefevre
> No - the decision is NOT being made on good data: Michael Lefevre's
> estimate of 13% affected is likely a HUGE underestimate due to much of
> the Test Pilot data being in English only and weighted towards US
> users who do not tend to have a problem with the Google box pointing
> to google.com (or seeming to) rather than their own version.
I have no idea of how accurate "my" estimate it is, but it has nothing
to do with test pilot. I just used Google and found an article which
mentioned that figure - I've got no idea how that data was collected,
but it was from 2004, so I doubt it was anything to do with Firefox. I'd
post a link, but I can't actually find it again immediately.
> Whatever's done needs to be done at 4.0
4.0 is a handful of small fixes away from being a release candidate - if
anything about this was going to happen for 4.0, it needed to have
happened earlier.
Michael
Ben Eficent
> to do with test pilot. I just used Google and found an article which
> mentioned that figure - I've got no idea how that data was collected,
> but it was from 2004, so I doubt it was anything to do with Firefox. I'd
> post a link, but I can't actually find it again immediately.
Oh, that's even worse - not Firefox at all. That basically means we
don't have a clue as to the numbers involved. But I think it's a
reasonable assumption that most non-US users are far more likely to
change their home page. Pretty much the first thing I did when I saw
the Mozilla home page was rest the mouse over Advanced Search, see
that it pointed to google.com and change my home page.
I could conduct a straw poll on the Testpilot usenet group, but it
would only give a very rough indication - perhaps better than nothing.
> 4.0 is a handful of small fixes away from being a release candidate - if
> anything about this was going to happen for 4.0, it needed to have
> happened earlier.
It seems to me that one releases a new software version when one knows
it isn't going to lose maybe 25% of 50 to 150 million users data!
Given the global audience for Firefox, which is available in so many
more languages than other browsers, I don't think 100m affected users
is unrealistic. 25% failing to figure out where their first session
went is 25m users losing a session.
It would be the greatest of ironies if Mozilla chose not to act to
resolve a data loss bug that most affected non-English speakers
because of concern about burdening translators with converting more
strings for non-English speakers!
On Feb 27, 6:12 am, Michael Lefevre <mjl+n...@michaellefevre.com>
wrote:
Daniel Cater
I'll try and bring some new points to the discussion and not just to repeat everything that I said there.
The main point to consider is, as others have said, those users who do not have about:home as their homepage. I believe this to a be a considerable percentage of users. I do not have research to confirm this, but I have not seen any research to suggest otherwise, either. Here is a list of user groups who I think are likely to have a non-default homepage.
- Those at work where Firefox has been deployed. The homepage will often have been set to an intranet site, or to the company's homepage.
- Those who buy a computer with Firefox pre-installed. The homepage will often, again, have been customised.
- Users who manually change their homepage. As Christian said, there is evidence of many non-technical users doing this. I see http://www.bbc.co.uk/news/ fairly often as a homepage.
And, here's the biggie:
- Almost anyone who has ever switched from another browser, download and installed Firefox themselves. This is because, as Limi said, most users just click through any dialogs, pressing OK as they go. What does this mean? It means that for Windows users, everyone who does this gets their settings migrated over from Internet Explorer by the import wizard, including their homepage.
So, what about all of those users? I'm gonna be bold and suggest that they actually number in the hundred million range.
Mike has said that there is not a dataloss issue here. I would like to disagree, with the following scenario (it was also alluded to in a previous comment in this thread). Assume a non-default homepage.
Be in the middle of something (it's important, but not urgent), with a handful of tabs open. Close Firefox without closing the tabs individually. You'll come back to this later, you just have to do something. Get caught up in something else. You don't get back to your computer till the next day. Wake up, open Firefox, check the weather, check the news. Close Firefox, eat breakfast. Come back, remember that you were in the middle of something yesterday. Realise that you can't get your tabs back because your saved session got overwritten. Cry a little inside.
With 3.6, this wouldn't have happened because you'd be reminded about your previous session on startup. You could then just open a new window, or new tabs, to check the news and weather. Then when you closed Firefox the second time, the tabs would still be saved.
In one of the Bugzilla comments that I linked to above, I came up with a possible thought process for a user who has just upgraded from 3.6 to 4.0. I made it up, but I think it could be fairly representative.
I think there needs to be much more (if there already is any) real-world user studying for Firefox. The independent, observe-but-don't-assist type. With this scenario (prefixed by: install Firefox), I think you would see the flaws in the current design.
Limi posted a blog a while ago asking for nightly testers to use a fresh profile (http://limi.net/articles/firefox-preferences/). I think this highlights the trend seen in the last cycle, where Firefox design can tend towards ignoring users of older versions of Firefox who will be upgrading (and not installing from scratch), or who have imported their settings from another browser when installing. Things like CrossWeave and the dirty profiles for Talos have been a good step towards emulating real-world environments, but there needs to be more.
Remember that because of the size of the userbase and the effectiveness of the update system, many more people will get Fx4 by upgrading, rather than performing fresh installs.
Asa Dotzler
> It seems to me that one releases a new software version when one knows
> it isn't going to lose maybe 25% of 50 to 150 million users data!
Ben, no data is lost. Stop repeating this misleading statement.
Users can always Restore Previous Session, even if they've changed their
Firefox start page.
Please stop going on about this. You are not convincing anyone and
you're reaching the point where you're going to alienate people and
cause them to take you less seriously on other issues you might bring up
in the future.
- A
Ben Eficent
pages pulled across is an excellent one.
The main session loss danger is to users who kept the default 'Warn me
when closing multiple tabs' prompt throughout Firefox 3, got used to
it asking about saving their session on exit and restoring it on
restart and think it's been lost when they see neither happening in
Firefox 4.0 But yes, it's all too easy to lose a session other ways
too. I haven't even checked what happens if you click an attachment
in an email between sessions which opens in Firefox. There's also a
bug in which a user with a blank home page and a restorable session
loses it on exit, although that might get fixed before final release.
If Mozilla is keen to increase the number of partnerships delivering
pre-installed copies, you'd think they'd be very keen to demonstrate
that they cared about user data.
On Feb 27, 8:45 am, Daniel Cater <djca...@gmail.com> wrote:
> My initial comments on this change can be seen in bug 592822 (https://bugzilla.mozilla.org/show_bug.cgi?id=592822#c49), comments 49, 51 and 55, and bug 629485 (https://bugzilla.mozilla.org/show_bug.cgi?id=629485#c33) where I didn't get a response.
>
> I'll try and bring some new points to the discussion and not just to repeat everything that I said there.
>
> The main point to consider is, as others have said, those users who do not have about:home as their homepage. I believe this to a be a considerable percentage of users. I do not have research to confirm this, but I have not seen any research to suggest otherwise, either. Here is a list of user groups who I think are likely to have a non-default homepage.
>
> - Those at work where Firefox has been deployed. The homepage will often have been set to an intranet site, or to the company's homepage.
> - Those who buy a computer with Firefox pre-installed. The homepage will often, again, have been customised.
> - Users who manually change their homepage. As Christian said, there is evidence of many non-technical users doing this. I seehttp://www.bbc.co.uk/news/fairly often as a homepage.
Ben Eficent
> Ben, no data is lost. Stop repeating this misleading statement.
Lost means you can't find something. If the user can't find their
session before it's overwritten it's lost.
Dao
> On Feb 27, 9:44 am, Asa Dotzler<a...@mozilla.com> wrote:
>> Ben, no data is lost. Stop repeating this misleading statement.
>
> Lost means you can't find something. If the user can't find their
> session before it's overwritten it's lost.
True, we can't just blame users for not dealing with our UI.
I also share the concern that users may not remember whether their
previous session was worth keeping when starting Firefox again.
Robert Kaiser
> The main point to consider is, as others have said, those users who do not have about:home as their homepage. I believe this to a be a considerable percentage of users. I do not have research to confirm this, but I have not seen any research to suggest otherwise, either. Here is a list of user groups who I think are likely to have a non-default homepage.
I actually think that among the population of people using session
restore, the amount of people having configured their homepage to
something different might be quite high, as I'd connect both to the type
of customization made by advanced users.
And we should care about those people as they are often peers in our
marketing chain that try to spread the word about Firefox.
Robert Kaiser
--
Note that any statements of mine - no matter how passionate - are never
meant to be offensive but very often as food for thought or possible
arguments that we as a community needs answers to. And most of the time,
I even appreciate irony and fun! :)
Asa Dotzler
I have concerns to. Every change concerns me.
Are you calling for a halt to the release and another beta to try
alternative approaches? If so, then say so and we can address that
head-on. If not, then your semi-support for Ben's continued insistence
on an issue that's already been decided just encourages more kibitzing.
- A
Dao
> I have concerns to. Every change concerns me.
>
> Are you calling for a halt to the release and another beta to try
> alternative approaches? If so, then say so and we can address that
> head-on.
No, now's not the time to try alternative approaches. The only sane
alternative would be the 3.6 quit dialog at this point.
> If not, then your semi-support for Ben's continued insistence
> on an issue that's already been decided just encourages more kibitzing.
Changes generally cause concerns, a decision has been made. Fine. But
you're not saying that this should prevent me from supporting others
when they make valid points or otherwise expressing my opinion, are you?
Asa Dotzler
> On 27.02.2011 18:27, Asa Dotzler wrote:
>> I have concerns to. Every change concerns me.
>>
>> Are you calling for a halt to the release and another beta to try
>> alternative approaches? If so, then say so and we can address that
>> head-on.
>
> No, now's not the time to try alternative approaches. The only sane
> alternative would be the 3.6 quit dialog at this point.
Even that would probably demand another beta, in my opinion, something
we cannot afford today.
>> If not, then your semi-support for Ben's continued insistence
>> on an issue that's already been decided just encourages more kibitzing.
>
> Changes generally cause concerns, a decision has been made. Fine. But
> you're not saying that this should prevent me from supporting others
> when they make valid points or otherwise expressing my opinion, are you?
I'm saying that people who aren't paying close attention to, or don't
care about our ship schedule, believe this issue should block the
release and have proposed solutions that would absolutely block the
release and that by expressing even a little solidarity with them, you
encourage them in their quixotic quest.
- A
Ben Eficent
> care about our ship schedule, believe this issue should block the
> release and have proposed solutions that would absolutely block the
> release and that by expressing even a little solidarity with them, you
> encourage them in their quixotic quest.
Quixotic? What a great compliment!
But thinking about it, putting the 3.6 quit back is probably one
default pref change. It's still there, and it still restores
immediately on next startup. It's not incompatible with the changes
to about:home, etc, so nothing would have to be undone. Doesn't sound
exactly unrealistic to me.
I would be interested to know what your acceptable percentage of users
to be affected by this is. Given all the factors that have been
mentioned by various people on this thread, there's no way it's
anywhere near as low as 5%! And that's the only percentage anyone's
been prepared to put their name against as acceptable data loss
exposure.
Dao
>> No, now's not the time to try alternative approaches. The only sane
>> alternative would be the 3.6 quit dialog at this point.
>
> Even that would probably demand another beta, in my opinion, something
> we cannot afford today.
I don't see why it would. We've shipped it before. Anyway, I didn't
start replying to this thread to provoke a certain outcome. Whatever
this means for Firefox 4: Valid points have been made, they shouldn't be
played down.
> I'm saying that people who aren't paying close attention to, or don't
> care about our ship schedule, believe this issue should block the
> release and have proposed solutions that would absolutely block the
> release and that by expressing even a little solidarity with them, you
> encourage them in their quixotic quest.
I want to encourage people to unreservedly discuss problems they see. If
they propose solutions along with it, that's fine too. They may be
unrealistic. This doesn't make their feedback useless.
Asa Dotzler
> I want to encourage people to unreservedly discuss problems they see. If
> they propose solutions along with it, that's fine too. They may be
> unrealistic. This doesn't make their feedback useless.
And I didn't say that the feedback was useless. I think the feeback is
fine.
What I don't think helps is continued pushes to change Firefox 4 which
has been "feature complete" for quite some time and what I also don't
think helps is to encourage discussion on why Firefox 4 is going to be a
train-wreck.
If the discussion switches to "what can we do to improve Firefox 5" and
stops being "Firefox 4 is going to be a train wreck" then I'm all for
it. So far I don't see a lot of movement in that direction.
- A
Dao
> And I didn't say that the feedback was useless. I think the feeback is
> fine.
>
> What I don't think helps is continued pushes to change Firefox 4 which
> has been "feature complete" for quite some time
Well, no, it wasn't complete for quite some time actually. It's not
Ben's fault that his feedback is late.
> If the discussion switches to "what can we do to improve Firefox 5" and
> stops being "Firefox 4 is going to be a train wreck" then I'm all for
> it. So far I don't see a lot of movement in that direction.
Addressing dataloss issues in a later release when people would suffer
from them in Firefox 4 is suboptimal, so people are understandably not
pushing for that.
Asa Dotzler
> Addressing dataloss issues in a later release when people would suffer
> from them in Firefox 4 is suboptimal, so people are understandably not
> pushing for that.
I continue to think is an abuse of the word "dataloss". By this
definition, we can start calling everything in the new Firefox menu
dataloss since users have to find the new location to get at it.
"My bookmarks are all gone! My history is all gone! My Downloads are all
gone!"
No, they're not gone. They're now located in the Firefox menu. The same
is true for sessions. They're not gone, they're now located on the
Firefox Start page and in the Firefox History menu.
- A
Mike Beltzner
We recognized when we made the change that it would be a different experience for many people, and have been working with SUMO and others to try and prepare for the fact that some users will be upset and feel like they have lost data. Change is hard, but we feel this change is more for the better than the worse, and will ultimately improve our product for more people than it will unfortunately impact.
I'm the moderator of this group, and I'm strongly suggesting we consider this topic closed. It's becoming more of as distraction at this point. Instead, please file bugs for future versions of Firefox that can help with the perceived dataloss issue.
cheers,
mike
Asa Dotzler <a...@mozilla.com> wrote:
- A
_______________________________________________
dev-apps-firefox mailing list
dev-apps...@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-apps-firefox
Daniel Cater
> I continue to think is an abuse of the word "dataloss". By this
> definition, we can start calling everything in the new Firefox menu
> dataloss since users have to find the new location to get at it.
>
> "My bookmarks are all gone! My history is all gone! My Downloads are all
> gone!"
>
> No, they're not gone. They're now located in the Firefox menu. The same
> is true for sessions. They're not gone, they're now located on the
> Firefox Start page and in the Firefox History menu.
>
> - A
Asa, please see my earlier reply to this thread. It is now easier to irreversibly overwrite your saved session than it was previously. This is to some extent, a dataloss issue.
I'm not suggesting saving more than one session so that you can go back historically to a number of previous sessions (although it's an interesting idea that an extension could perhaps implement), but a way to remind users (who don't have about:home as their homepage) that when they last closed Firefox, they had more than one tab open.
Please (please) put yourself into the mind of Fx 3.6 user who is used to having the Save and Quit / Quit / Cancel dialog come up when they go to close a Firefox session and answer this question: what option should they choose when the following dialog now comes up?
Confirm close
You are about to close 5 tabs. Are you sure you want to continue?
Close Tabs / Cancel
It is much less obvious what is going on than in the 3.6 dialog. It used to say "Do you want Firefox to save your tabs for the next time it starts?" How on earth do they know that sessions are now saved by default? They could gamble, and assume that that is the case, but it's unlikely that they're going to do that if the session is important.
Daniel Cater
> Everyone: I appreciate the debate and the fact that people feel passionately about this. The fact of the matter is that the change has been in the product for months, and the feedback from our 2.5 million beta users has not given us strong concern that this will be a problematic change for the *majority* of our users.
I am interested in why you have highlighted the word majority there. Is it because you believe that a small percentage of users save their session? I'm not saying that that is or isn't true, I would just like to know. I've already posted the reasons why I believe the customised homepage percentage to be much higher than you're anticipating (Alex suggested that less than 5% would be OK to leave fixing it till the next cycle - I think it could be at least 5x that).
I think the feedback left here is non-negligible: http://input.mozilla.com/en-US/beta/search?q=save+tabs&product=firefox&version=--&date_start=&date_end=
You can clearly see when beta 10 was released. Yes, people testing betas are more likely to be saving sessions, but they are also more likely to go hunting around the menus for the "Restore Previous Session" button and yet they still post things like:
"it doesn't have option to save multiple tabs upon close."
"Not saving multiple tabs on quit."
"I'm not getting the option to save my tabs when I close the only Firefox window that was open, unlike before."
"Firefox 4 just gets worse. There is no way to save tabs anymore when you close the window. This is completely unacceptable!"
These are just from the last few hours.
Dao
> I continue to think is an abuse of the word "dataloss". By this
> definition, we can start calling everything in the new Firefox menu
> dataloss since users have to find the new location to get at it.
>
> "My bookmarks are all gone! My history is all gone! My Downloads are all
> gone!"
>
> No, they're not gone. They're now located in the Firefox menu. The same
> is true for sessions. They're not gone, they're now located on the
> Firefox Start page and in the Firefox History menu.
This comparison seems a bit flawed... Before finding the new location,
users may have restarted Firefox, causing the previous session to be
gone for good. After the new location has been discovered, the fact
remains that the UI is available at a different point in time. It
requires users to recall their previous session -- forgetting something
means losing it. This is a new, non-trivial burden.
It seems to me that a web browser should be smarter and more helpful
than that.
Dao
> The fact of the matter is that the change has been in the product for months
The critical part, the quit dialog removal, actually arrived only with
beta 10, a good month ago.
James May
regression, and worked around it by using the history menu. (I've never
seen anyone use the history menu before) a little annoying, but most of the
time I remember.
My mother will not be so lucky.
-- James
Sent from my phone, excuse brevity.
Michael Lefevre
> Everyone: I appreciate the debate and the fact that people feel
> passionately about this. The fact of the matter is that the change
> has been in the product for months
As already pointed out, it's more like one month... and in
response to some of the comments after the change was made, it was
pointed out that the whole design wasn't implemented yet, and some
follow ups happened less than 3 weeks ago, which just got released in
beta 12.
> We recognized when we made the change that it would be a different
> experience for many people, and have been working with SUMO and
> others to try and prepare for the fact that some users will be upset
> and feel like they have lost data.
That kind of work is what I was hoping for in my earlier post (I am not
advocating further delays to the release at this point), but I haven't
seen it. Maybe I'm not looking in the right place or with the right
search terms, but all I can find on SUMO is a forum question with the
answer providing a link to a MozillaZine knowledge base page (which
seems to have incomplete info given in a rather too technical manner).
There doesn't seem to be an actual SUMO article, anything in the release
notes or in the Firefox 4 features page (where there is stuff pointing
out the other changes that Asa compared this issue to in a different bit
of this thread).
The best information on this seems to be zpao's blog, which has a link
to the comment in nsBrowserGlue.js which documents how this works.
That's not an ideal place for people to be able to find it.
> Instead, please file bugs for future versions of Firefox that can
> help with the perceived dataloss issue.
I'm not sure further bugs are actually necessary - there are already at
least a couple open about this, with debates going on...
Michael
Ben Eficent
browser.showQuitWarning to true in Beta 12 would resolve the data loss
issues. The Fx3.6 save prompt comes up on exit (X or File > Exit)
with the tabs warning correctly suppressed, and if you click 'Yes' and
restart your session is restored immediately.
I say issues plural because there are a number of bugs relating to
sessions not immediately restored on startup being overwritten, and
probably more to be discovered. In https://bugzilla.mozilla.org/show_bug.cgi?id=637148
for instance, users with about:home set who open Firefox with a
session saved and immediately close it lose that session (a new bug in
Beta 12).
Setting the pref to default to true would give you time to work
through these bugs - and, ideally, add the facility to handle more
than one saved session - while the users who carefully save and
restore their sessions could continue to do so blissfully unaware of
such bugs. NOT to do so, KNOWING that a large number of users were
going to lose valuable large sessions - well, it's hard to get away
from terms like scandalous.
Asa, I'm NOT calling Firefox 4 a train-wreck. Firefox 4 is a next
generation browser with a lot of impressive developments. I'm saying
that not telling users where their session has gone is a train-wreck.
If the word 'train-wreck' is the thing that's making you and Mike see
red then that's unfortunate. If you could look in on yourselves from
outside you'd see that you're not putting the interests of the user
first if you don't address this, and data integrity and reputation are
more important than road-maps. (Plus, just how long does it take to
change one pref and check a few scenarios?) But what do you think
your user base will be calling Firefox 4 if they fall foul of these
traps? I imagine many will be calling it their ex-browser.
On Feb 22, 2:52 am, Ben Eficent <beneficent...@gmail.com> wrote:
> For some inexplicable reason, the handy addition of 'Restore Previous
> Session' on the History menu has led developers to decide they no
> longer need to (a) prompt on exit or (b) restore on startup. Worse,
> they are not telling most users that their session has been saved or
> what to do to get it back!
>
> The result, if Firefox 4 ships this way, will be huge numbers of users
> who currently click [Save and Quit] on exit and get their session back
> on restart LOSING THEIR FIRST SESSION after upgrading, as they won't
> know that it has been saved or how to get it back. Others will waste
> time looking for a solution.
>
> A restore button has been added to the default home page, about:home,
> but just about anyone who doesn't use this will be none the wiser.
>
> The people responsible for this decision are aware of the
> consequences, but seem perfectly fine with them!
>
> Well I think it's appalling. I have come up with a solution which
> could be easily implemented to address this - seehttps://bugzilla.mozilla.org/show_bug.cgi?id=635231- but I'm not
Mike Beltzner
The decision remains final, and rests with me. I rarely talk in such absolutes, but here we are!
cheers,
mike
Ben Eficent <benefi...@gmail.com> wrote:
On Feb 22, 2:52?am, Ben Eficent <beneficent...@gmail.com> wrote:
> For some inexplicable reason, the handy addition of 'Restore Previous
> Session' on the History menu has led developers to decide they no
> longer need to (a) prompt on exit or (b) restore on startup. ?Worse,
> they are not telling most users that their session has been saved or
> what to do to get it back!
>
> The result, if Firefox 4 ships this way, will be huge numbers of users
> who currently click [Save and Quit] on exit and get their session back
> on restart LOSING THEIR FIRST SESSION after upgrading, as they won't
> know that it has been saved or how to get it back. ?Others will waste
> time looking for a solution.
>
> A restore button has been added to the default home page, about:home,
> but just about anyone who doesn't use this will be none the wiser.
>
> The people responsible for this decision are aware of the
> consequences, but seem perfectly fine with them!
>
> Well I think it's appalling. ?I have come up with a solution which
> could be easily implemented to address this - seehttps://bugzilla.mozilla.org/show_bug.cgi?id=635231- but I'm not
> holding my breath! ?This is because the recent changes detailed above
> are part of a wider move towards form over function.http://groups.google.com/group/mozilla.dev.apps.firefox/browse_thread...
>
> The more I look into this, the more it seems a case of modal dialogs
> offending the sensibilities of the developers ? and who cares if a
> million users lose session data?
>
> There is also a distorted view of the impact of closing a session
> accidentally. ?Large sessions typically do not restore as they were
> before ? web pages update ? or even get taken down, video caches get
> lost, etc, etc.
>
> The quit prompt alone is enough of an issue that people are talking
> about making a plug-in just to provide that functionality. ?But
> everyday users ? the ones most likely to be affected by accidental
> closings ? are unlikely to know of such plug-ins should they be
> created.
>
> This situation is a scandal waiting to happen; it needs to be
> addressed before the release of Firefox 4.0. ?Users should certainly
> have the option of a prompt on exit. ?But far more importantly, they
> need to know how to restore their session!!!
Ben Eficent
Interesting. I seem to have missed your comment thanking me for
passionately arguing my position - you must have posted it to another
group and assumed I'd use Google Group search to locate it.
Actually, your developers seem to have been consistently contesting
your contentions throughout this post. Also, the lack of data on
numbers affected is staggering. It's a fundamental flaw in logic:
'We'll tell users with about:home where their data is but not bother
to set up a Test Pilot report to find out what the percentage actually
is.'
Ah, but you HAVE given a hint that numbers concerned are low: a link
to a Mozilla Input search for "restore". And it's fairly quiet.
Well, has it occurred to you that 'Restore Previous Session' is the
terminology for the new feature, and users used to having a prompt to
save tabs are going to complain about not being able to "save tabs",
and that that search is chockablock? And that the very fact of the
"restore" search being low indicates a problem finding the new
feature?
There is one issue I can think of that would override addressing an
issue exposing perhaps 25m users to session loss, and that is a
security issue so bad that it affects still more users. But given
that the 'change browser.showQuitWarning default to true' option could
probably be implemented in a day if a handful of developers were
tasked with scenario checking, that doesn't really pass muster.
I suspect that a number of dedicated Mozilla developers have just
discovered they're not working in the aspirational "We believe in..."
organisation they thought they were.
On Feb 28, 10:50 pm, Mike Beltzner <beltz...@mozilla.com> wrote:
> I'm not seeing red, Ben. I've explained my position patiently and repeatedly. I've apologized for not being able to adequately explain why I hold that position as that seems to be frustrating you, and thanked you for taking the time to argue passionately for your position.
>
> The decision remains final, and rests with me. I rarely talk in such absolutes, but here we are!
>
> cheers,
> mike
>
> Ben Eficent <beneficent...@gmail.com> wrote:
>
> I have just confirmed that yes, simply changing
> browser.showQuitWarning to true in Beta 12 would resolve the data loss
> issues. The Fx3.6 save prompt comes up on exit (X or File > Exit)
> with the tabs warning correctly suppressed, and if you click 'Yes' and
> restart your session is restored immediately.
>
> I say issues plural because there are a number of bugs relating to
> sessions not immediately restored on startup being overwritten, and
> probably more to be discovered. Inhttps://bugzilla.mozilla.org/show_bug.cgi?id=637148
> > could be easily implemented to address this - seehttps://bugzilla.mozilla.org/show_bug.cgi?id=635231-but I'm not
> > holding my breath! ?This is because the recent changes detailed above
> > are part of a wider move towards form over function.http://groups.google.com/group/mozilla.dev.apps.firefox/browse_thread...
>
> > The more I look into this, the more it seems a case of modal dialogs
> > offending the sensibilities of the developers ? and who cares if a
> > million users lose session data?
>
> > There is also a distorted view of the impact of closing a session
> > accidentally. ?Large sessions typically do not restore as they were
> > before ? web pages update ? or even get taken down, video caches get
> > lost, etc, etc.
>
> > The quit prompt alone is enough of an issue that people are talking
> > about making a plug-in just to provide that functionality. ?But
> > everyday users ? the ones most likely to be affected by accidental
> > closings ? are unlikely to know of such plug-ins should they be
> > created.
>
> > This situation is a scandal waiting to happen; it needs to be
> > addressed before the release of Firefox 4.0. ?Users should certainly
> > have the option of a prompt on exit. ?But far more importantly, they
> > need to know how to restore their session!!!
>
> _______________________________________________
> dev-apps-firefox mailing list
> dev-apps-fire...@lists.mozilla.orghttps://lists.mozilla.org/listinfo/dev-apps-firefox
Millwood
At some point even a poster who starts out with a legitimate issue can
become a troll. I fear we are there. So -
DON'T FEED THE TROLL.
Everything's been said, so lets all shut up.
Mike Beltzner
Believe what you will, Ben. I believe that your comments are motivated by the fact that the decision didn't go your way. Nothing I see in what you have written here changes that understanding.
cheers,
mike
Ben Eficent <benefi...@gmail.com> wrote:
> I'm not seeing red, Ben. I've explained my position patiently and repeatedly. I've apologized for not being able to adequately explain why I hold that position as that seems to be frustrating you, and thanked you for taking the time to argue passionately for your position.
Interesting. I seem to have missed your comment thanking me for
On Feb 28, 10:50?pm, Mike Beltzner <beltz...@mozilla.com> wrote:
> I'm not seeing red, Ben. I've explained my position patiently and repeatedly. I've apologized for not being able to adequately explain why I hold that position as that seems to be frustrating you, and thanked you for taking the time to argue passionately for your position.
>
> The decision remains final, and rests with me. I rarely talk in such absolutes, but here we are!
>
> cheers,
> mike
>
> Ben Eficent <beneficent...@gmail.com> wrote:
>
> I have just confirmed that yes, simply changing
> browser.showQuitWarning to true in Beta 12 would resolve the data loss
> issues. ?The Fx3.6 save prompt comes up on exit (X or File > Exit)
> with the tabs warning correctly suppressed, and if you click 'Yes' and
> restart your session is restored immediately.
>
> I say issues plural because there are a number of bugs relating to
> sessions not immediately restored on startup being overwritten, and
> probably more to be discovered. ?Inhttps://bugzilla.mozilla.org/show_bug.cgi?id=637148
> for instance, users with about:home set who open Firefox with a
> session saved and immediately close it lose that session (a new bug in
> Beta 12).
>
> Setting the pref to default to true would give you time to work
> through these bugs - and, ideally, add the facility to handle more
> than one saved session - while the users who carefully save and
> restore their sessions could continue to do so blissfully unaware of
> such bugs. ?NOT to do so, KNOWING that a large number of users were
> going to lose valuable large sessions - well, it's hard to get away
> from terms like scandalous.
>
> Asa, I'm NOT calling Firefox 4 a train-wreck. ?Firefox 4 is a next
> generation browser with a lot of impressive developments. ?I'm saying
> that not telling users where their session has gone is a train-wreck.
> If the word 'train-wreck' is the thing that's making you and Mike see
> red then that's unfortunate. ?If you could look in on yourselves from
> outside you'd see that you're not putting the interests of the user
> first if you don't address this, and data integrity and reputation are
> more important than road-maps. ?(Plus, just how long does it take to
> change one pref and check a few scenarios?) ?But what do you think
> your user base will be calling Firefox 4 if they fall foul of these
> traps? ?I imagine many will be calling it their ex-browser.
EE
If you are saying that there will for sure be a Firefox 5, are you not
saying that there will be something wrong with Firefox 4?
EE
Bookmarks have always been in the menu (and in the sidebar). If all the
history is now going to be in the menu, does that mean that you are
doing away with sidebars?
Is there a difference between the start page and the home page? Do you
mean that the sessions would become your home page? What if one wants
to clear out private data and get rid of the session on shutdown? Will
that still be available in the Tools menu?
Asa Dotzler
Yes, I am absolutely saying that 1) there will be a Firefox 5, and 2)
there will be things wrong with Firefox 4.
Anyone who disagrees with those statements is confused.
- A
Robert Kaiser
> If you are saying that there will for sure be a Firefox 5, are you not
> saying that there will be something wrong with Firefox 4?
There's "something wrong" (i.e. at least something that can be improved)
with any piece of software (or actually anything in life). There is also
always a next version - unless you get tired of making the future better
than the present. And you can be sure that Mozilla will try to improve
the future as long as this community and this web exists! :)
Ron Hunter
I don't disagree, but I am confused by this whole thread. I really
haven't seen any change in the action of my nightlies relative to the
home page, or any tendency for it to do anything different from what it
(Firefox) has done since Phoenix.
Dao
> I don't disagree, but I am confused by this whole thread. I really
> haven't seen any change in the action of my nightlies relative to the
> home page, or any tendency for it to do anything different from what it
> (Firefox) has done since Phoenix.
If you selected "When Minefield starts: Show my windows and tabs from
last time" in the options, you're unaffected.
Ron Hunter
No, mine starts with www.google.com. No changes needed, or noticed.
Daniel Cater
Some more feedback from the last few hours:
"firefox doesn't prompt me if i want to save my session like how it used to in previous versions."
"It is not possible to save the current session (open tabs)"
"Auto restore of last session is MISSING! A major feature for laptop users."
"The restore previous session wipes if you accidently close the newly opened browser!!"
"doesnt warn when closing multiple tabs. wont restore entire previous session. lost 5 hours of work here... you guys are starting to fail."
This is the kind of bug which causes a huge lack of confidence in the browser and a real reputation-damager. It is not the user's fault.
Asa Dotzler
> Asa, could you please reply to my earlier post, or at least state why you feel it is not worthy of a reply?
I won't reply to your earlier post or any other on this topic. I respect
the moderator of this group and he has asked quite nicely that we end
this discussion.
- A
On 2/27/2011 3:58 PM, Mike Beltzner wrote:
> I'm the moderator of this group, and I'm strongly suggesting we
> consider this topic closed. It's becoming more of as distraction at
> this point. Instead, please file bugs for future versions of Firefox
Ben Eficent
> definition, we can start calling everything in the new Firefox menu
> dataloss since users have to find the new location to get at it.
>
> "My bookmarks are all gone! My history is all gone! My Downloads are all
> gone!"
>
> No, they're not gone. They're now located in the Firefox menu. The same
> is true for sessions. They're not gone, they're now located on the
> Firefox Start page and in the Firefox History menu.
Asa, as I understand it, the Firefox menu is only displayed by default
in Windows Vista and 7. Everyone else gets a version of Firefox that
looks very similar to 3.6.
The user exits with tabs open, and is asked whether they want to close
them. "That's odd", they think, "Why isn't it asking me to save?"
Maybe they go back and check settings in Options, or check the File
menu to see if there's a save option - no clues there. So they
confirm close, hoping they'll then be prompted to save. Nope. Very
worried by this point. Restart Firefox. Session is not there. Saved
session has ALWAYS loaded up immediately on next start, so obviously
there's a bug in the new version and the saving has failed.
If they exit from the File menu - the more cautious approach which
they may have opted for specifically to avoid accidentally losing
their session - they get no prompt at all, and are immediately going
to be very worried if they have an important session. Again, with the
session not loaded at next start the obvious conclusion is a bug
caused by the upgrade.
Now the user has closed for a reason. They may well need to be
somewhere, so now close Firefox - and that's their session well and
truly gone, as their home page will overwrite the previous session.
Or maybe they try Help, where they'll be out of luck, and so again
will think it MUST be a bug.
The above is not supposition - it's borne out by Beta users several
times an hour: http://input.mozilla.com/en-US/beta/search?q=save+tabs&product=firefox&version=--&date_start=&date_end=
The user has done nothing wrong or unreasonable and cannot be
criticised in any way. So the blame for the situation lies with the
developers. It's serious data loss on a massive scale and it could be
avoided pretty much by the flip of a virtual switch.