667245 - the widget for switching folder pane views is heavy and ugly

Status: RESOLVED FIXED

(Thunderbird :: Theme, defect)

Description

[Asa Dotzler [:asa]]

The widget for cycling through the folder pane views is heavy and ugly. It's the larger and darker than any other icon in the interface. I don't believe this is the correct widget either. Ideally it'd be a select or a menu or something. It's not clear at all that the arrows are buttons or that they act on the title of the current view. At a minimum, they need to be styled more obvious and less heavy.

Comment 1

[Jim Porter (:squib)]

Could we just use the < and > icons for scrolling the tab bar here? This would also fix some related issues, like bug 469300 (I assume).

Comment 2

[Jim Porter (:squib)]

Attachment: Switch to a menulist instead of the arrows to cycle

In light of http://weblogs.mozillazine.org/asa/archives/2011/07/primary_thunderbird.html and the ill-fated patch from bug 535021, here's a patch to switch to a menulist.

This probably needs tests, but before I work on that, I want to make sure this is something we want (and that this is the way to do it). We probably also want to work on theming so this looks prettier (esp on aero), but getting the XUL fixed is the first step.

Comment 3

[Jim Porter (:squib)]

Attachment: What it looks like on Linux

Comment 4

[Andreas Nilsson (:andreasn)]

Even though this is probably a more correct widget to use, it might also give the whole area an even heavier feel to it. To some extent I feel that bug 535021 primary came into existence because a lot of people was very vocal about their dislike of unified folders being the default view (on 3.0 when it was). In that case it might be enough to make them small again, or something in between in size.

Taking a step back (and I hope this don't drift too far away from the bug topic), do we really need this view switch widget this prominent in the interface? 
Most other interfaces for changing the behavior of the 3-pane lives in menus and preferences dialogs. How is this different?

Comment 5

[Blake Winton (:bwinton) (:☕️)]

So, we were going to switch to a drop-down, as per https://bugzilla.mozilla.org/show_bug.cgi?id=535021#c34 (but didn't because there wasn't time to test it, and it would have been a large change in a .0.x release) and I believe that's the correct widget to use for this bit of UI, but…

Anecdotal evidence suggests that people use it exactly once, to switch to the view they want, and then never touch it, or only do by accident, and then get frustrated.  And so, if it's essentially never changing, then I think just removing it from the top of the folder list might be the better idea here.

Thanks,
Blake.

Comment 6

[Roland Tanglao needinfo please :rolandtanglao, :mohnkuchen, :adobo, :sinigang, :roland]

hi blake :

I believe data from GS will support your "anecdotal evidence"

if you can get me some keywords or regular expressions to search, I can get you hard data from Get Satisfaction using my GS regex search script of doom :-)

https://github.com/rtanglao/momogs/blob/master/bruteforceSearch.rb

Comment 7

[Asa Dotzler [:asa]]

(In reply to comment #5)
> So, we were going to switch to a drop-down, as per
> https://bugzilla.mozilla.org/show_bug.cgi?id=535021#c34 (but didn't because
> there wasn't time to test it, and it would have been a large change in a
> .0.x release) and I believe that's the correct widget to use for this bit of
> UI, but…
> 
> Anecdotal evidence suggests that people use it exactly once, to switch to
> the view they want, and then never touch it, or only do by accident, and
> then get frustrated.  And so, if it's essentially never changing, then I
> think just removing it from the top of the folder list might be the better
> idea here.


BTW, I'm totally cool with removing it. I would have proposed that but I figured it would get shot down so I went with "fix rather than delete" but I support removal. It might be nice to have a pref somewhere for preferred view.

Comment 8

[Richard Marti (:Paenglab)]

Just for record. In menu View > Folders you can also switch the view.

Comment 9

[Jim Porter (:squib)]

I think "Unread folders" and "Recent folders" are the only views that one would want to switch back and forth from on a regular basis. Of course, I have no idea how many people use those, or how frequently. Yet another reason we need Test Pilot...

Comment 10

[Blake Winton (:bwinton) (:☕️)]

Comment on attachment 544153 [details] [diff] [review]
Switch to a menulist instead of the arrows to cycle

So, it's pretty ugly on the Mac <http://dl.dropbox.com/u/2301433/Screenshots/MacFolderView.png>, so, I'm going to go with ui-r-.

And I think I'ld rather remove it entirely, so I'm going to go with feedback-, too.  :)

(All of the comments I've gotten so far are in favour of, or at least neutral towards, removing this UI, and using the menu items to switch.  Feel free to wait a while if you want to gather more data before removing it.)

Later,
Blake.

Comment 11

[Jim Porter (:squib)]

Attachment: Aero mockup without the header

Here's a mockup of what it would look like on Aero. I added 3px of vertical padding to the top of the folder tree and now both trees line up (woo, yay!). I'm sure this will be tough to get just right with non-default font sizes, but I'm not terribly worried about that; it's not like this would make it any worse for them.

This should also be doable on Mac, since the thread pane header is shorter than a regular row (so we just make the header the same height).

Linux is a bit harder, but I think we should wait until the theme refresh is done there before we get too worried about pixel-pushing here.

Comment 12

[Jim Porter (:squib)]

Attachment: Aero mockup without the header

Argh, this was the version I meant to attach.

Comment 13

[Jim Porter (:squib)]

The more I think about this, the more I think we should just ignore the vertical alignment issues for now. Sure, it'd be nice, but there's no good way to make the folder pane and the thread pane line up both when the quick filter bar is shown and when it's hidden. This would probably be a lot easier to handle further down the road after we've finished moving other things around.

Comment 14

[Blake Winton (:bwinton) (:☕️)]

Comment on attachment 544691 [details]
Aero mockup without the header

Ugh.  But I see your point.  But if we do push this off to later, I want some sort of guarantee that we actually get to it before the next release…

And since most of the feedback I've gotten has been positive, I'm going to say f=me.  ;)

(One of the suggestions was that we move the Folders menu's sub-items up to the View menu, to make it a little easier for users to get to them.  I don't know whether I'm entirely in favour of that, but it does seem like a nice thing to do for the people who are losing the quick switch…)

Thanks,
Blake.

Comment 15

[Jim Porter (:squib)]

(In reply to comment #14)
> Ugh.  But I see your point.  But if we do push this off to later, I want
> some sort of guarantee that we actually get to it before the next release…

Before I'd be willing to start trying to push pixels around, I'd want at minimum to know whether we should be optimizing for the QFB expanded or collapsed. I don't think it's likely that we'll be able to get the QFB's height to be an integer multiple of the treerow height, so no matter what, we're going to have vertical alignment problems in one case.

If, say, we went with Asa's proposed UI (or something like it), we could make the quick filter bar become a regular toolbar that goes across the folder pane and the thread pane. This would probably work pretty well with a tabs-on-top UI too, and then we'd only have to worry about the tree header, which is easy enough to account for. However, all that's really sketchy, and I don't think we could reasonably do all that in one release.

Comment 16

[Andreas Nilsson (:andreasn)]

Comment on attachment 544691 [details]
Aero mockup without the header

This looks good. The feedback I've seen so far is in favor of this change and it's good to get less chrome and more space for actual content. Lets go ahead with this approach!

Comment 17

[Jim Porter (:squib)]

While we're at it, should we do the same to the header in the address book? The "Address Books" label isn't particularly useful; in fact it's kind of confusing, since the whole window is the "Address Book", but it also *contains* other "Address Books".

This would have the happy side effect of neatly avoiding the problem mentioned in but 670639 comment 5.

Comment 18

[Andreas Nilsson (:andreasn)]

I'm possibly fine with losing that header too.

Comment 19

[Jim Porter (:squib)]

Attachment: Similar UI in Windows Explorer

For what it's worth, Windows Explorer does pretty much the same thing as I did in my proposal (i.e. adds some vertical padding to the folder pane). Of course, they get it wrong by one pixel, but they're close. :)

As mentioned in IRC, things do look a bit off when you have the very first folder pane item selected and account central is shown, but that's one case out of many, whereas other options are worse:

1) Don't add vertical padding to the folder pane => every folder looks off
2) Add vertical padding to the account central pane => every account but the
   first looks off (and anyone who has uses IMAP has at least 2 "accounts")

Comment 20

[Jim Porter (:squib)]

Attachment: Remove the headers from the folder pane and the address book

Here's a patch that hides the headers for the folder pane and address book. As mentioned above, I'm just hiding them for now, not removing them. I did add a comment telling future people to get rid of those elements, just in case we forget.

Comment 22

[Andreas Nilsson (:andreasn)]

Comment on attachment 546471 [details] [diff] [review]
Remove the headers from the folder pane and the address book

I think it would make sense with some spacing at the top, so if you want to add that, it would be a bonus. ui-r+ without that as well though.

Comment 23

[Roland Tanglao needinfo please :rolandtanglao, :mohnkuchen, :adobo, :sinigang, :roland]

(In reply to comment #6)
> hi blake :
> 
> I believe data from GS will support your "anecdotal evidence"
> 
> if you can get me some keywords or regular expressions to search, I can get
> you hard data from Get Satisfaction using my GS regex search script of doom
> :-)
> 
> https://github.com/rtanglao/momogs/blob/master/bruteforceSearch.rb

I did a brute force search and found 300 or so Get Satisfaction Thunderbird support topics that mention unified. If you look at these topics you will realize that the widget is misunderstood and confusing to users on Get Satisfaction:
https://gist.github.com/1090466

I think my work is done here :-) but if anybody wants to extra data from the Get Satisfaction Support Database of 46000 topics since July 2009 with respect to this bug or any other bugs, please ping me in IRC or email me directly instead of bugzilla!

Comment 24

[Blake Winton (:bwinton) (:☕️)]

Comment on attachment 546471 [details] [diff] [review]
Remove the headers from the folder pane and the address book

Review of attachment 546471 [details] [diff] [review]:
-----------------------------------------------------------------

I would prefer to file a followup bug instead of leaving comments in the code, so that we have a chance of actually doing it sometime.  ;)

And I've come to think that we should move the Folders menu's sub-items up to the View menu, to make it a little easier for users to get to them, since there were a few people who used the quick switch which we've taken away from them.

(But aside from that, I'm happy with the patch.  ;)

Comment 25

[Jim Porter (:squib)]

Attachment: Move the View -> Folder items up a level and clean up some code

Here's a patch that puts the View -> Folders items in the main View menu. For what it's worth, I don't really like this; I think it's too cluttered. I also cleaned up some of the menuitem-hiding code, since I had to modify it anyway. Hopefully it will be a little more manageable now.

Comment 26

[Jim Porter (:squib)]

Attachment: What the menu looks like

Here it is in action. I forgot to mention above that I changed some of the shortcut accelerators, and put "All Folders" first, since that's currently the default.

Comment 27

[Blake Winton (:bwinton) (:☕️)]

Ugh, that's more cluttered than I was hoping.  :(

Okay, let's revert that part, but I do like your idea of putting "All Folders" first, so we should keep that.

Thanks,
Blake.

Comment 28

[Blake Winton (:bwinton) (:☕️)]

(Oh, and I think you can make those changes without me checking over your shoulder, so r=me for that patch. ;)

Comment 29

[Jonathan Kamens]

I switch between different folder views using the arrows regularly and would be disappointed to have to go to a menu to switch them, especially given the new functionality I suggested in bug 674510.

I submit that using GetSatisfaction data to judge whether this would be a welcome change is a flawed methodology, considering that what you're going to find on GS is complaints from people who don't like / understand the current UI, not positive feedback from people who like it and/or regularly use the functionality you're talking about removing.

I suppose an alternative to the solution I proposed in bug 674510 would be to have the folder view automatically switch to all folders if it isn't in that mode and you hover over it for a while during a drag/drop, and then switch back automatically after you finish the drop. So if the change proposed here is indeed the right one (but as noted above, I'm not convinced the GS data proves that), there is still a way to solve the problem I described in bug 674510.

Comment 30

[Blake Winton (:bwinton) (:☕️)]

Jonathan,

I'm not basing my decision only on GS.  I've also taken into account comments in bugzilla, on the Thunderbird UI blog, and talks with a couple of UX professionals.  While the arrows are a feature some people use quite a bit, for the vast majority of people they are unnecessary and confusing, and so, in the interest of making life better for as many people as we can, we're removing them.

(In the future, we'll use Test Pilot to gather as much data as we can for this sort of thing, but for now, I feel I've gotten a reasonable cross-section of responses, and they're overwhelmingly for the removal.)

Having said that, we should figure out a way to fix the problem you report in bug 674510, but let's do that over there.)

Thanks,
Blake.

Comment 31

[Jonathan Kamens]

Blake, Is there a way to fix the problem that the current UI is heavy, ugly and confusing while preserving the ability to use button clicks to flip between views?

I'm thinking of the iOS ability to flip from screen to screen just by swiping. Now, obviously, TB can't rely on gestures for UI elements, but I've seen similar techniques used elsewhere. For example, perhaps the left and right borders of the folder view can be thickened slightly and "live" so that clicking on the border flips views?

Are there / will there be key bindings for flipping the view left or right? That, at least, would satisfy those of us who (a) flip views often, (b) don't want to have to use a slow menu sequence to do it, and (c) are willing to learn key bindings.

(And yes, I realize that if I want such a thing I can just use the keybindings add-on to set up the appropriate bindings myself, but it seems like it wouldn't hurt, if we're going to take away the arrows, to put the bindings in the core, if they aren't already there.)

Comment 32

[Blake Winton (:bwinton) (:☕️)]

Keybindings aren't a bad idea, but…
a) it's _really_ hard to find unused ones for other stuff we're removing,
b) we should probably reserve the un-taken key bindings for features used by more people, and
c) we don't actually want to make it too easy for people to switch, because that noticeably increased the amount of confusion and number of support requests we got.

On the plus side, the menu shouldn't be that slow, and indeed will involve fewer clicks than the arrows would have for most switches.  ;)  (And, if you're on Windows or Linux, the menus should be keyboard-accessible through the Alt key, so you will be able to use something like Alt-V-F-A to switch to the All Folders view.  Apparently Ctrl-F2 (or Ctrl-Fn-F2) on Macs will do something similar.)

And there's always the add-on route!  We've intentionally hidden (and not removed) the folder pane header so that someone could write a simple add-on that would show it again.  We'll be tracking the usage of those buttons, and looking at downloads of that add-on, and if it turns out that there are a lot of people using it, we'll definitely look into putting it back.

Thanks,
Blake.

Comment 33

[Jim Porter (:squib)]

Checked in: http://hg.mozilla.org/comm-central/rev/cb57fc284a41

Comment 34

[Jim Porter (:squib)]

Attachment: Patch for checkin

Here's the patch for check-in. Should it also go into comm-aurora? It's a pretty simple patch, but it doesn't fix anything that's "broken" exactly.

Comment 35

[Jonathan Kamens]

(In reply to comment #32)
> And there's always the add-on route!  We've intentionally hidden (and not
> removed) the folder pane header so that someone could write a simple add-on
> that would show it again.

Ooh, you just *knew* I wouldn't be able to resist that despite the fact that it would mean another late night of TB programming, didn't you? :-)

https://addons.mozilla.org/en-US/thunderbird/addon/folder-pane-view-switcher/

(This also addresses bug 674510.)

Comment 36

[Mark Banner (:standard8)]

Comment on attachment 548965 [details] [diff] [review]
Patch for checkin

Not for aurora. This is a UX change, not a bug fix and I think this should go through the cycle as normal.

Comment 37

[rsx11m]

(In reply to comment #35)
> Ooh, you just *knew* I wouldn't be able to resist that despite the fact that
> it would mean another late night of TB programming, didn't you? :-)

...and for people who don't like add-ons, put this into your userChrome.css:  ;-)

#folderPaneHeader, #abDirTreeHeader {
  display: -moz-box !important;
}

Comment 38

[Joe Sabash [:JoeS1]]

While the widget might have been "heavy and ugly" and also "error prone", it did have one more characteristic. It was informative. It showed the user what folder view was in effect. I fear we will see more support requests along the lines of a mistaken "view message body as..." selection.

Comment 39

[rsx11m]

> (comment #0) Ideally it'd be a select or a menu or something.

That was proposed in bug 535021 attachment 420102 [details] which eventually went a different path and just styled the arrows.

Comment 40

[Philipp Kewisch [:Fallen] [:📆][:🧩]]

Way too late, sorry I didn't see this bug earlier. I created the "advanced unread folders" extension [1] which adds a hierarchical unread view and I use this as my default view. It does happen more often that I need to see all folders, to access a folder that doesn't have unread messages. I guess I could get used to a keyboard shortcut, but that doesn't work out on mac. Too bad this went away, R.I.P.


[1] https://addons.mozilla.org/en-US/thunderbird/addon/advanced-unread-folders

Comment 41

[Blake Winton (:bwinton) (:☕️)]

Philipp,

if you propose a key that we could use to switch to the previously-selected folder view, I would be more than happy to review a patch to add it!  (Most of the people who used the folder view switcher (and there weren't many) used it mainly to go between a couple of views, so making that easier for them would satisfy about 90% of the people who are upset by this change)!

Thank you,
Blake.

Comment 42

[Philipp Kewisch [:Fallen] [:📆][:🧩]]

Yes, users of my addon would likely be switching between my folder pane view and the all folders view. I have though of just (optionally) reviving the header pane in my addon.

The problem with a key to cycle is not really finding a key (for example, Shift+F could work), but rather that you'd have to cycle through all views just to switch between two. Before, the lucky coincidence happened that the advanced unread view was just next to the all folders view, so users could use a single click on the < or > arrow to switch views.

Comment 43

[Blake Winton (:bwinton) (:☕️)]

I'm not a fan of non-Ctrl keys for things like this.  I've heard of problems with people typing into the wrong pane and having things magically disappear on them.  And I'm not sure what Ctrl-keys are available.

I'ld also prefer the key not to cycle between all the views, but to jump back and forth between the last two views you looked at, since I feel (and I've heard) that that's a much more common task.

Comment 44

[Blake Winton (:bwinton) (:☕️)]

Ooh, what about Ctrl-D, for "last folDer view"?

(Okay, fine, the mnemonic is terrible, but I'm pretty sure the key is unused, since I stole it for TbAir a while ago…)

Ludo, do you know of anything that uses Ctrl-D?

Comment 45

[Jim Porter (:squib)]

Lightning uses Ctrl+D.

In any case, I'm not convinced that a keyboard shortcut is the ideal way to handle this, though I'm sure it would be appreciated by people who switch folder views frequently. A context menu item in the folder pane might help, but I think this is the sort of thing where Test Pilot could help us (especially since keyboard shortcuts are a bit of a mess in Thunderbird).

Comment 46

[Marc Auslander]

I'm in the toggle back and forth camp, and suggest a built in or add-on mail toolbar widget as the easiest way out.  I'd put it right next to my toggle preview pane icon, also from an add-on.

Comment 47

[Philipp Kewisch [:Fallen] [:📆][:🧩]]

<p tone="sarcastic">What about a toolbar in the folder pane area?</p>

Comment 48

[Larry Gusaas]

I switch folder view several times a day.

Thank you for once again "Improving" Thunderbird by disabling this valuable feature.

I think that it is time to switch email clients.

Comment 49

[Marcel Berteler]

i concur with Larry... removing features is the last thing I was waiting for and definitely did not EXPECT this when upgrading. If your remove features yo should WARN the users before they apply the update and point them to alternative ways.

I understand that some features can be moved from the core to an extension, but INFORM people about this, nit just remove them and get people searching the net to find the bug that caused they folder switch to disappear...

https://addons.mozilla.org/en-US/thunderbird/addon/folder-pane-view-switcher/?src=search

Comment 50

[alex]

what is wrong with you people !?
do you ever do user testing !?
it's unbelievable someone removed such a feature because ONE developer found it "heavy and ugly".

you know, that's why we use a mail client and not webmail.. for these sorts of features one click away.

and what happened to if it's not broke don't fix it ?

bah...

Comment 51

[Andreas Nilsson (:andreasn)]

(In reply to alex from comment #50)
> what is wrong with you people !?
> do you ever do user testing !?
> it's unbelievable someone removed such a feature because ONE developer found
> it "heavy and ugly".

Yes we did user testing, and it was one of the things that people in that test accidentally triggered.

We also asked for feedback on Breakingtheegg [1] and it was clearly more than one developer who was for removing the chrome. In general we're having a big issue with having too much chrome and too little space for actual content so this is one of the reasons we decided to get rid of the folderpane switcher above the folders (it's still available from the menus of course).

1. http://breakingtheegg.tumblr.com/post/7353630782/anecdotal-evidence-suggests-that-most-people-use

Comment 52

[Andreas Nilsson (:andreasn)]

(In reply to Marcel Berteler from comment #49)
> I understand that some features can be moved from the core to an extension,
> but INFORM people about this, nit just remove them and get people searching
> the net to find the bug that caused they folder switch to disappear...

The way we would do this, and how we did it around the time of Thunderbird 3.0 (it might have been 3.1) is that we had a migration assistant. The problem with that was a) more code to maintain and b) really irritating to everyone upgrading. Doing it for one feature only is therefore in my opinion not worth that cost.
Do you have suggestions of other ways we could do this?

Comment 53

[Marcel Berteler]

Not sure if I have a solution for it, but maybe it is time to to start thinking about pre-installed plugins, or directly after upgrading point people to plugins that re-enable the features that where removed. Not a complete migration wizard, but a simple message the the below features have been removed and can be re-enabled by installing the following plugins.

I have no problem running a plugin for my compact headers and my folder pane switch, as long as I get informed about it in a visible way.

I know it was included in the 'whats new' but I fail to see why this is listed under 'New Features' and not under 'Removed Features'.  

Besides the location of the message, it is also hidden by default. I would say that removing features is so important that it should be visible by default as it is out of the norm. 

Fixed bugs, etc can be hidden by default as it is expected that bugs are fixed, and therefore it does not carry as much news worthiness credits as things that are not expected, like removed features.

well, thats my 5 cents.

Comment 54

[Jonathan Kamens]

(In reply to Andreas Nilsson (:andreasn) from comment #51)
> Yes we did user testing, and it was one of the things that people in that
> test accidentally triggered.

In my opinion, what you and Blake and everyone else claiming that you did "user testing" are missing is that posting stuff on the web and asking people for their opinions is not "user testing," and it certainly is not an effective way to take an effective cross-section of the user community.

1) It's a self-selecting sample, in that it only selects the kind of people who frequent the sites on which you ask the questions.

2) People who want the status quo to change are far more likely to speak up than people who want things to stay the same.

It's all well and good to say that in your opinion the feature was causing more harm than good, but it's just your opinion. Nothing I've seen provides anything like actual data supporting the claim that it does more harm than good.

Here's some real data: Three days ago, 20 people downloaded my add-on to put back the buttons. Two days ago, the day the beta was released with the change, 767 people downloaded it. Yesterday, 3,861 users downloaded it. So what were you saying about nobody wanting this feature?

You removed a useful feature and didn't replace it with anything better. "We're hoping to add keyboard shortcuts, but we can't figure out what shortcuts to use," and, "We're thinking about adding a toolbar button, but we haven't gotten around to it yet," are outrageous statements to be making if you're actually trying to be friendly to your users.

I understand that the tools for collecting real data about things like this are not as good as they could and should be. I understand the difficulty of collecting such data. I understand that sometimes you have to try something out and deal with the backlash if it turns out to be the wrong thing. What I am objecting here is the "We know better, so shut your trap" attitude that seems to permeate the comments of the people defending the change. You don't know better. Admit that it was an experiment and that maybe you were wrong, and people will be a lot less upset.

> We also asked for feedback on Breakingtheegg [1] and it was clearly more
> than one developer who was for removing the chrome.

See above. Note, especially, on the question of whether you're collecting representative data, that Blake explicitly stated in that Breakingtheegg thread that he was going to delete any "me too!" comments. So not only is what you're doing not going to gather good data for the reasons cited above, you're actually *changing* the data to make your decision look better.

Comment 55

[Jonathan Kamens]

(In reply to Jonathan Kamens from comment #54)
> Here's some real data: Three days ago, 20 people downloaded my add-on to put
> back the buttons. Two days ago, the day the beta was released with the
> change, 767 people downloaded it. Yesterday, 3,861 users downloaded it. So
> what were you saying about nobody wanting this feature?

Correction: It wasn't the beta that was released two days ago, it was the final 8.0 release. Sorry, didn't realize that.

Comment 56

[Jonathan Kamens]

(In reply to Andreas Nilsson (:andreasn) from comment #52)
> Do you have suggestions of other ways we could do this?

Well, for one thing, you could mention in the release notes how people could get the buttons back by installing my add-on, instead of just saying that the widget was removed. The add-on was released and fully reviewed well before 8.0 was released.

Comment 57

[Andreas Nilsson (:andreasn)]

(In reply to Jonathan Kamens from comment #54)
> (In reply to Andreas Nilsson (:andreasn) from comment #51)
> > Yes we did user testing, and it was one of the things that people in that
> > test accidentally triggered.
> 
> In my opinion, what you and Blake and everyone else claiming that you did
> "user testing" are missing is that posting stuff on the web and asking
> people for their opinions is not "user testing," and it certainly is not an
> effective way to take an effective cross-section of the user community.

Sorry if I was unclear. This was user testing at the Canonical Office in London with people from the street (so to say).
http://design.canonical.com/2011/02/thunderbird-in-the-usability-lab/

> Here's some real data: Three days ago, 20 people downloaded my add-on to put
> back the buttons. Two days ago, the day the beta was released with the
> change, 767 people downloaded it. Yesterday, 3,861 users downloaded it. So
> what were you saying about nobody wanting this feature?

It's really cool that the addon is popular, but at the same time we have the part of the userbase that don't use that feature and is actively hurt by it. There is also the part of the userbase that don't care at all. :)

> You removed a useful feature and didn't replace it with anything better.
> "We're hoping to add keyboard shortcuts, but we can't figure out what
> shortcuts to use," and, "We're thinking about adding a toolbar button, but
> we haven't gotten around to it yet," are outrageous statements to be making
> if you're actually trying to be friendly to your users.

Yes, we are indeed running very short on shortcuts, there are only that many letter on a keyboard :)
So alt+v+f+f gives you favorite folders and alt+v+f+a would give you all folders, do you think we need additional shortcuts apart from these?

> I understand that the tools for collecting real data about things like this
> are not as good as they could and should be. I understand the difficulty of
> collecting such data. I understand that sometimes you have to try something
> out and deal with the backlash if it turns out to be the wrong thing. What I
> am objecting here is the "We know better, so shut your trap" attitude that
> seems to permeate the comments of the people defending the change. You don't
> know better. Admit that it was an experiment and that maybe you were wrong,
> and people will be a lot less upset.

Collecting data is indeed really hard and the only data I have to go on right now is the user testing, the feedback in the post and the fact that we really want to remove chrome in order to get more space for content and that we wanted to fix alignment issues in the interface in general. I'm really sorry if I come of as arrogant, because that's not my intention at all. I just love Thunderbird and want to make it succeed.
The thing that makes me really happy though is that we have Testpilot now, so we can actually measure things like this on our userbase in the future, with proper numbers to back it up.
If we do a study that shows that the folderpane switching is really, really frequent on our userbase I'm more than happy to admit that I was wrong and will propagate heavily for adding it back.

Comment 58

[Andreas Nilsson (:andreasn)]

(In reply to Jonathan Kamens from comment #56)
> (In reply to Andreas Nilsson (:andreasn) from comment #52)
> > Do you have suggestions of other ways we could do this?
> 
> Well, for one thing, you could mention in the release notes how people could
> get the buttons back by installing my add-on, instead of just saying that
> the widget was removed. The add-on was released and fully reviewed well
> before 8.0 was released.

Great idea, I like it!
Could you post a message about this to tb-planning so we make sure not to miss that next time around we do a change to the ui like this?

Comment 59

[Marcel Berteler]

(In reply to Andreas Nilsson (:andreasn) from comment #58)
> (In reply to Jonathan Kamens from comment #56)
> > (In reply to Andreas Nilsson (:andreasn) from comment #52)
> > > Do you have suggestions of other ways we could do this?
> > 
> > Well, for one thing, you could mention in the release notes how people could
> > get the buttons back by installing my add-on, instead of just saying that
> > the widget was removed. The add-on was released and fully reviewed well
> > before 8.0 was released.
> 
> Great idea, I like it!
> Could you post a message about this to tb-planning so we make sure not to
> miss that next time around we do a change to the ui like this?

Well, this is rather weird. It is mentioned in the 'Whats New' page in TB itself, but not in the 'Release Notes' or the 'New in TB8' webpage. 

Are these various pages created by different people and is would it not be a good idea to have a section 'Removed Features' instead of clumping this under 'More New Features to Explore'?

Comment 60

[Jonathan Kamens]

(In reply to Andreas Nilsson (:andreasn) from comment #57)
> Sorry if I was unclear. This was user testing at the Canonical Office in
> London with people from the street (so to say).
> http://design.canonical.com/2011/02/thunderbird-in-the-usability-lab/

I read the notes from the user testing at that URL. There isn't a single word about the folder pane view switcher. So that particular issue could not have been a particularly significant one during the testing.

Not to mention the fact that the experiences of 11 new users trying out an application for one hour each are not indicative of how they will users will fare in the long-term.

If you design UIs that can be totally understood in the first hour of use, you will end up with simplistic, crippled UIs. It takes a while for users to figure out how to take advantage of every feature in a UI.

> Yes, we are indeed running very short on shortcuts, there are only that many
> letter on a keyboard :)

Yes, that's why buttons are useful.

> So alt+v+f+f gives you favorite folders and alt+v+f+a would give you all
> folders, do you think we need additional shortcuts apart from these?

Favorites on top and all on bottom like Outlook would be far superior to an obscure shortcut that no one can remember.

> Collecting data is indeed really hard and the only data I have to go on
> right now is the user testing, the feedback in the post and the fact that we
> really want to remove chrome in order to get more space for content

Space for content in the message list and preview pane are important. Space for content in the folder list is rather less so. You removed useful functionality to gain only one additional line in the folder list. I hardly think that's a good trade-off.

Comment 61

[Jonathan Kamens]

(In reply to Andreas Nilsson (:andreasn) from comment #58)
> Could you post a message about this to tb-planning so we make sure not to
> miss that next time around we do a change to the ui like this?

1. I have no idea what "tb-planning" is.

2. Don't wait until the next time around. Add a link to the release notes NOW, for TB8.

Yes, I realize that's a translation issue. So get it translated. Add it to each language's release notes as the translation for that language is complete. Just adding it to the English release notes, right now, will help the majority of Thunderbird users by itself. Do the best you can; don't let perfect be the enemy of good.

Comment 62

[rsx11m]

The "What's new" section of the release notes is virtually useless these days, and there is no hint about Jonathan's extension in the Mozilla KB either.

There is bug 701042 for a similar incident of missing documentation, and these things can be fixed right away and translated as localizers go (like help content, nothing breaks the application if the translations are a few days behind).

Btw: Substantial influx of complaints from users on GS over the last 24 hours.

Comment 63

[Jonathan Kamens]

(In reply to rsx11m from comment #62)
> ...there is no hint about Jonathan's extension in the Mozilla KB either.

I wrote an article about it in the KB yesterday. Still waiting for it to be reviewed and published.

Comment 64

[Mark Richards]

TB is wonderful, and I still love it, but glad I read the release notes: "Folder switching pane widget removed".  

Gulp.

As a principle, perhaps when something is taken away a good reason can be provided in return and - in this case - suggested alternatives to explore may be given.

Assuming that users do not use a feature and that it is ugly ("bloated") in the U/I are not particularly good reasons to dig into code (and likely break something else), especially when there are far more unglamorous bugs to be fixed.  This alone should have deep-6'd this at its infancy.

Reading the volumes of comments and work on this I have to wonder where the priorities are.

Less a rhetorical question: where are they?

Comment 65

[Jonathan Kamens]

I don't think this bug is an appropriate forum to discuss "where the priorities are."

Comment 66

[Chris Hecker]

Add me to the pile of people who upgraded and immediately said "WTF?" when finding the arrows missing.  I found this bug by finding the change that commented the folderpane header out, although changing it back in omni.jar doesn't seem to make it appear.  

I wish there was a way to subscribe to the security and core fixes to tbird without getting any of the rand() inspired UI changes that seem to land every time it's updated.  It's incredibly annoying.

Comment 67

[Blake Winton (:bwinton) (:☕️)]

(In reply to Chris Hecker from comment #66)
> Add me to the pile of people who upgraded and immediately said "WTF?" when
> finding the arrows missing.

Consider yourself added.

There are certainly more people complaining about the loss than I would have thought based on the feedback I had originally received.  I'm waiting for some numbers to come on Monday, but if they follow the recent trend, I would be in favour of adding a preference to re-add the switcher, along with a menu option in the "View » Folders" menu to toggle it on and off.  (I would even push to get it in Aurora, so that you don't have to wait as long to get it.)

> I found this bug by finding the change that
> commented the folderpane header out, although changing it back in omni.jar
> doesn't seem to make it appear.  

Does the add-on at https://addons.mozilla.org/en-US/thunderbird/addon/folder-pane-view-switcher/ work for you?

> I wish there was a way to subscribe to the security and core fixes to tbird
> without getting any of the rand() inspired UI changes that seem to land
> every time it's updated.  It's incredibly annoying.

As I understand it, Thunderbird will also have ESR releases sometime in the future, so you'll be able to get security and core fixes, and only need to deal with rand()*7 inspired UI changes every 42 weeks…  (I will be very interested to see how many people choose the ESR release of Thunderbird.)

Thanks,
Blake.

Comment 68

[Jim Porter (:squib)]

(In reply to Blake Winton (:bwinton - Thunderbird UX) from comment #67)
> There are certainly more people complaining about the loss than I would have
> thought based on the feedback I had originally received.  I'm waiting for
> some numbers to come on Monday, but if they follow the recent trend, I would
> be in favour of adding a preference to re-add the switcher, along with a
> menu option in the "View » Folders" menu to toggle it on and off.  (I would
> even push to get it in Aurora, so that you don't have to wait as long to get
> it.)

I think the best way of handling this "preference" would be an item in the mail toolbar, as I mentioned on IRC. This would also give us the opportunity to improve the UI a bit (i.e. not go back to the weird arrows).

Comment 69

[Chris Hecker]

(In reply to Blake Winton (:bwinton - Thunderbird UX) from comment #67)
> There are certainly more people complaining about the loss than I would have
> thought based on the feedback I had originally received.

I think if you'd taken out the arrows, but then added a hotkey for cycleMode that was right there in the menu, people would have said, "okay, at least they put some effort into it as opposed to just removing something".

> Does the add-on at
> https://addons.mozilla.org/en-US/thunderbird/addon/folder-pane-view-switcher/
> work for you?

Yes, although I don't understand why editing messegner.xul inside omni.jar didn't also fix it.

Chris

Comment 70

[Wayne Mery (:wsmwk)]

http://getsatisfaction.com/mozilla_messaging/topics/create_a_more_intuitive_default_ui_method_of_switching_between_folder_views_and_knowing_which_folder_view_you_are_in

Comment 71

[Alain Knaff]

Just one more datapoint:

I use the widget on a regular basis in order to toggle between the "Unified Folder", "Favorite Folders" and "Recent Folders" view. The first one is my "normal" view, whereas the latter 2 are very handy when sorting mails into folders. So, the statement "users use this exactly once" is NOT true. Getting rid of "Unified Folder" is not the only use case for the folder pane view switcher. And having the arrows there at the top of the folder view in incredibly more convenient to use than having to dig deep into the "View" menu whenever I want to toggle between a view having just my favourite folders, and one having all my folders...

And what's this business about "heavy" and "darker icons"? If the look of the widget needs improving, then just change the look, but don't remove it altogether. And apparently, the look has _already_ been fixed. Indeed on the version I'm using (11.0.1 on Ubuntu 10.04), it doesn't look any darker than the other UI elements. The grey background blends in perfectly with the grey of the toolbar above it, and with the quickfilter bar to the right of it.
Moreover, the arrow icons are not that large either, actually they are smaller than the icons in the toolbar just above it.

Ok, in the meantime, I've found the addon that restores this (https://addons.mozilla.org/en-US/thunderbird/addon/folder-pane-view-switcher/contribute/roadblock/?src=search&version=1.7), but its a pain when an upgrade removes features that people actually do use. Especially if this is done for such silly reasons as "oh, the icon's a tad too dark". If looks were more important to me than functionality, I'd use Windows and Outlook Express.
Just saying.