300198 reloaded: Still broken, still needs fixed
tyl...@gmail.com
id=300198) for a while. The short story is, the vast majority of users
think that loading a bunch of bookmarks in existing tabs results is
dataloss that needs to be fixed. 1 or 2 devs think that such
destructive behavior is desirable. Many arguments for fixing this bug
have apparently fallen on deaf ears, so since rational discussion
seems to go nowhere, I thought I'd do a count.
In a previous newsgroup discussion, (http://groups.google.com/group/
mozilla.dev.apps.firefox/browse_thread/thread/
3ab9fda96f0d4b27/4874d43ada944f9b?
lnk=gst&q=300198&rnum=1#4874d43ada944f9b), I count 10 posts in favor
of fixing the bug and 2 posts that don't really state an opinion.
Nobody clearly suggested that the bug remain unfixed.
In the Bugzilla entry, I count 14 posts in favor of fixing the
dataloss, 2 that don't really say, and only 3 posts that want to keep
the destructive behavior.
Furthermore, the Bugzilla entry has another 18 votes.
Clearly, most users think that 300198 is a problem that needs to be
fixed. If the devs are uncomfortable with the apparent
"inconsistency", then by all means change the wording from "Open in
Tabs" to "Open in New Tabs". Or, I'd even be fine with the current tab
being replaced and all other bookmarks being opened in new tabs. If a
tiny number of people really want all their existing tabs replaced,
then they can change their preferences -- default preferences should
reflect the preferences of the majority, and here I think that the
majority clearly want tabs appended, not replaced.
Whatever, I think that the people have spoken -- and they say to fix
this bug. I ask the devs to please reconsider their 'nofix' decision.
Steve Chapel
> Clearly, most users think that 300198 is a problem that needs to be
> fixed.
No, most users who bothered to comment in the places you looked think
it's a problem that needs to be fixed. Because users who are happy with
the current behavior have no motivation to seek out these places and
post their opinions, the "votes" you've tallied are seriously biased by
being skewed towards the opinion that the current behavior is a problem.
What you need to do to clearly demonstrate that most users think it's a
problem is to poll a random sample of users. Note that I'm not
disagreeing that it is a problem -- just pointing out a basic error in
your statistical methodology. See
<http://en.wikipedia.org/wiki/Sample_bias> for more details.
tyl...@gmail.com
> > fixed.
> No, most users who bothered to comment in the places you looked think
> it's a problem that needs to be fixed.
>
> What you need to do to clearly demonstrate that most users think it's a
> problem is to poll a random sample of users.
Yes, I agree that subscribers to the Bugzilla entry may present
somewhat of a bias. So I emailed some people I know, and so far the 8
of them that have responded have unanimously stated that replacing
existing tabs is undesired behavior. A few suggested changing the
wording to "Open in NEW Tabs" and a few suggested opening the tabs in
a new window to preserve the old tabs. One user reported that IE7
appends rather than replaces tabs -- I don't want FF to just "copy"
what IE does, but it's a strong argument that Microsoft, with all of
their resources and one ginormous userbase, did their homework and
found a compelling reason to append rather than replace.
(I should mention that I was very careful in the wording of my email
not to prejudice one behavior or the other. I also didn't tell them
about this FF bug until AFTER they had cast their vote. Some of the
people are IE users, some FF, some Opera, some I don't know. So these
results are about as clean as I can make them.)
If the devs really want to be "consistent", they can make left-
clicking a single Bookmark open in a new tab by default. That's much
better than destroying the users data as the current behavior does.
RQ
As for reasoning, I could just double what Gervase Markham said in the
previous discussion ( http://groups.google.com/group/
mozilla.dev.apps.firefox/browse_thread/thread/
3ab9fda96f0d4b27/0cfb577d655abe15?#0cfb577d655abe15 ) - quoting below.
Gervase Markham wrote:
> Mike Connor wrote:
> > I'd basically argue that there's no "obvious" expectation here. Is
> > overwriting the current tab dataloss?
>
> Overwriting existing non-current tabs is certainly dataloss;
> particularly if you were filling in forms in them.
>
> IMO, the most intuitive and consistent thing is to overwrite the current
> tab, and insert all the other pages in tabs between the current one and
> the next one over.
>
> People are used to a bookmark navigating them away from a page they are
> viewing - that's fine. They are also used to the act of choosing a
> bookmark _not_ affecting their other windows/pages.
>
> So, if we now postulate a form of bookmark which consists of multiple
> pages, the consistent behaviour is to overwrite the current page, but
> not any others.
RQ
Rimas Kudelis
Here are some quotes from mconnor's message, with my comments:
> Should that be true for bookmarks as well, since the argument is
> technically the same for 1 and > 1, just the impact is greater?
> Should we replace the current tab with a contiguous tabset, a la the
> Home behaviour in Fx2?
Yes, please!
> Really the main problem with your assertion is that if left click on
> a bookmark means "replace my current tab," why is left click on the
> bottom option not a replace action?
On the other hand: Why would opening one bookmark replace only ONE
tab? If you want consistency, then no matter how many tabs you're
opening from bookmarks (and one bookmark is one tab, while a bookmark
folder is just N tabs, right?), these should replace all available
tabs. Why is ONE an exception then?
The answer is obvious: because you are replacing only the current tab!
Same should happen in all cases. Plus, middle-clicking on a bookmark
folder should not touch the current tab at all, opening the whole set
of pages in new tabs.
RQ (r...@akl.lt in bugzilla)
Steve Chapel
> On Thu, 1 Feb 2007 18:55:05 UTC, Steve Chapel wrote:
>
>> What you need to do to clearly demonstrate that most users think it's a
>> problem is to poll a random sample of users.
>
> to use it (or what that at another site)?
The MozillaZine poll? Again, that's a self-selected poll, not a random
sample. Did you read the Wikipedia article I linked to?
Steve Chapel
>>> Clearly, most users think that 300198 is a problem that needs to be
>>> fixed.
>
>> No, most users who bothered to comment in the places you looked think
>> it's a problem that needs to be fixed.
>>
>> What you need to do to clearly demonstrate that most users think it's a
>> problem is to poll a random sample of users.
>
> Yes, I agree that subscribers to the Bugzilla entry may present
> somewhat of a bias. So I emailed some people I know, and so far the 8
> of them that have responded have unanimously stated that replacing
> existing tabs is undesired behavior.
"Some people you know" is very far from a random sample. ;-)
tyl...@gmail.com
> > of them that have responded have unanimously stated that replacing
> > existing tabs is undesired behavior.
> "Some people you know" is very far from a random sample. ;-)
Yeah, I figured you would call me on that. :)
You know, we could turn this around: challenge the other camp -- those
who think that replacing the tabs is correct -- to show their user
study or other documentation supporting their position. So far I have
seen nothing that supports that conclusion.
For the record, I do appreciate the effort that the devs have put into
FF. It's a great browser. I just can't understand why they don't want
it to be better.
Mike Shaver
> I just can't understand why they don't want
> it to be better.
This thread is not likely to be made more useful, so it would be great
if people could resist the (extremely strong, I know) temptation to
continue to rebut each other's points or offer statistical critique.
Or, at least, to copy the mailing list on it.
The original post's points are well-understood, and I believe that a
review of this behaviour (note: I AM NOT SAYING THAT IT WILL OR WILL
NOT BE CHANGED) is on mconnor's radar as part of Fx3 bookmarks work,
so:
- if you agree with a position already made in this group or in the
bug, rest assured that the previous incarnation of your position will
be considered; vote _silently_ (meaning "without additional comment)
for the bug if you would like; and
- if you have _genuinely_novel_ data or choices to propose, please do
so in the bug rather than in this thread. (But please, try to set the
bar for "genuinely novel" rather high!)
Thanks to everyone for their interest and enthusiasm!
Mike
Myk Melez
>>> I emailed some people I know, and so far the 8
>>> of them that have responded have unanimously stated that replacing
>>> existing tabs is undesired behavior.
>
>> "Some people you know" is very far from a random sample. ;-)
>
> Yeah, I figured you would call me on that. :)
Surveys aren't always the best way to figure out what's best for users,
but they might be helpful in this case. Maybe we should ask the user panel:
-myk
Mike Connor
Also self-selected to some extent, but I think the biases are known
there...
I personally prefer the replace behaviour, but I know I didn't like
it in the beginning (but that's over four years ago). Undo Close Tab
mitigates a lot of this, but more useful + unexpected < useful +
expected in UE design, so that argues for a change despite existing
behaviour. I'm curious why Safari's behaviour remains unchanged to
date, since both behaviours were implemented by Hyatt in a similar
timeframe.
I'm offline as I write, but the current plan I'm leaning toward (not
set in stone) is to follow the Home button behaviour I implemented
for Fx2. There will absolutely be a pref (likely hidden) to use the
old behaviour, because some people really prefer the old behaviour.
Key points:
* Middle-click is consistently open in new tab throughout the UI
(with the notable exception being closing tabs)
* Left-click replaces the current tab for every UI action that loads
a new URL
* Old behaviour needs to be available for old fogeys like me.
* Middle-click == Accel+click in all cases
Random notes:
* IIRC, middle click on the feed icon/menu doesn't preview/subscribe
(in the always use web reader case) in a new tab. We might have
fixed this late in Fx2, but I don't remember. If not, we should fix
that. Search has a bug on this that we should fix.
The Great and Irrefutable Draft Plan (please feel free to refute)
* Middle clicking either the Open All menuitem or the folder will
append the tabset to the current tabs
* Left clicking the Open All in Tabs menuitem will replace the
current tab with the selected tabset
* Context menu option will behave in line with left-click.
-- Mike
Mike Connor
On 1-Feb-07, at 5:30 PM, tyl...@gmail.com wrote:
>>> I emailed some people I know, and so far the 8
>>> of them that have responded have unanimously stated that replacing
>>> existing tabs is undesired behavior.
>
>> "Some people you know" is very far from a random sample. ;-)
>
> Yeah, I figured you would call me on that. :)
>
> You know, we could turn this around: challenge the other camp -- those
> who think that replacing the tabs is correct -- to show their user
> study or other documentation supporting their position. So far I have
> seen nothing that supports that conclusion.
a) Neither flip of the current pref is explicitly more correct, both
have drawbacks. My main focus here is to reduce complexity and
maintain UI consistency across the UI.
b) Current behaviour doesn't need justification, the change needs
justification. Like it or lump it, this is how you avoid UI churn.
Once we are satisfied that a new behaviour is _better_, we should
change it, but not before. That said, see my other post for my
proposed changes.
> For the record, I do appreciate the effort that the devs have put into
> FF. It's a great browser. I just can't understand why they don't want
> it to be better.
Of course we want it to be better, just because we disagree with one
particular conclusion doesn't mean we don't want the app to be
better. This is my least favourite rhetorical trick, which is
expanding a line of thinking out to an extreme conclusion that has no
basis in fact. What you're really saying is "I don't know why my
obviously superior opinion isn't being followed."
FWIW, this was on my radar for Fx2, as part of the Places work, and
when that pushed out, the tabs issue went with it.
-- Mike
tyl...@gmail.com
>
> current tab with the selected tabset
If I read that right, you are saying that only one existing tab (the
currently selected one) will be replaced -- all other existing tabs
remaining untouched -- and the rest of the new tabs will all be
appended? If so, then that's great. That eliminates the unexpected/
undesired dataloss. It's also more consistent, since opening one
bookmark has the same behavior as multiple bookmarks: replacing the
current tab and only the current tab.
Rimas Kudelis
> The Great and Irrefutable Draft Plan (please feel free to refute)
>
> * Middle clicking either the Open All menuitem or the folder will append
> the tabset to the current tabs
> * Left clicking the Open All in Tabs menuitem will replace the current
> tab with the selected tabset
> * Context menu option will behave in line with left-click.
I feel happy. Really. :)
RQ
Gervase Markham
>> The Great and Irrefutable Draft Plan (please feel free to refute)
>>
>> * Left clicking the Open All in Tabs menuitem will replace the
>> current tab with the selected tabset
>
> If I read that right, you are saying that only one existing tab (the
> currently selected one) will be replaced -- all other existing tabs
> remaining untouched -- and the rest of the new tabs will all be
> appended?
I believe so.
I entirely agree with this (unsurprisingly) but I note one issue: won't
this mean that if you have three tabs open (from left to right, A, B and
C), and you are viewing tab A, and you left click to open a tabset
consisting of D, E and F, you'll get:
D, B, C, E, F
i.e., D, E and F are not together? We may not consider this a problem,
of course. The other option is to rearrange the tabs at the same time,
so you do actually get:
B, C, D, E, F
Gerv
Myk Melez
> won't this mean that if you have three tabs open (from left to right,
> A, B and C), and you are viewing tab A, and you left click to open
> a tabset consisting of D, E and F, you'll get:
>
> D, B, C, E, F
>
> i.e., D, E and F are not together? We may not consider this a problem,
> of course. The other option is to rearrange the tabs at the same time,
> so you do actually get:
>
> B, C, D, E, F
A third option is to do:
D, E, F, B, C
-myk
tyl...@gmail.com
>
> D, E, F, B, C
This make the most sense to me.
Mike Shaver
Yeah, that's what mconnor said would happen, I think: "will replace
the current tab with the selected tabset".
Mike
Mike Connor
On 6-Feb-07, at 11:11 AM, Mike Shaver wrote:
> On 2/6/07, Myk Melez <m...@mozilla.org> wrote:
>> A third option is to do:
>>
>> D, E, F, B, C
>
> Yeah, that's what mconnor said would happen, I think: "will replace
> the current tab with the selected tabset".
That's what happens with the Home button currently. Its fairly
trivial to move the tabs to the right place, since I fixed
tabbrowser's loadTabs method to handle this for Firefox 2.
-- Mike
Gervase Markham
> A third option is to do:
>
> D, E, F, B, C
Yeah, that would make most sense, wouldn't it :-)
Gerv