unity-design team mailing list archive
-
unity-design team
-
Mailing list archive
-
Message #05328
Re: why global menubar/application menu isn't such a great idea
There are several large, glaring, problems with the global menu as I see it:
Menu's are outdated. It's clear from the last few years that menus as we
know it are playing a more diminished role. Large amounts of applications
simply don't use them as they are largely a crutch for poor UI design. Not
that they should go entirely as programs like Photoshop pretty much couldn't
do without them but for the bulk of end-user applications such as browsers
(and even word processors) they are not exactly needed and the functionality
can be exposed in alternative, superior, ways.
What the global menu does is take this decision away from the application
developer and basically says 'you are getting a menu, tough'. It wouldn't
be too bad but do not forget that this decision will have ramifications for
Ubuntu for at least a decade. Once it's in it'll define the UI and
application development and changing it will probably cause a lot of
breakages. The reason OSX still has a global menu imo is probably because
of historical inertia they can't change it as it will break too much.
It's purposefully throwing a roadblock in the way of evolution once it is
in it can't get 'evolved' out as it is too ingrained.
The second problem is one of scope and the user model. Tabs-on-top has been
adopted as the default UI mechanism in browsers because the address bar
belongs to the page, not the browser, and each page has it's own address
bar. If you click on a tab it should only change things within the domain
of the tab contents and the old style caused things outside the scope of the
tab to change (address bar). Changing things outside of the defined window
border as a result of actions inside the window border destroys the concept
of having a program as a self contained unit and introduces uncertainty.
You can't simply say 'the program is in this box' if various other parts of
the OS UI change depending on the current application.
The whole Fitt's Law argument is also largely invalid. The difference in
targeting time for edge items and central items is not really significant.
It is important to consider but should not be used to justify anything as
just about every other argument carries more weight. I think it is just
unfortunate it is one of the only usability 'rules' that has any form of
empirical backing so gets given emphasis in every decision. Without the
global menu tabs and the window decorations then gain this much vaunted
space and although I do not have any empirical evidence I would hazard
people use the decorations and tabs more than they use the menu.
The last argument for the global menu 'We have all this space, lets jam
something in it' is quite frankly disturbing. If there is lots of wasted
slack space in a UI then it's an argument for removing the space, not
filling it with rubbish. Looking at the top bar on stock Gnome, only about
15% is actually used for anything useful and commonly accessed. Putting the
global menu at the top does not save space if the space does not need to be
used. It's an argument for rethinking the layout, not for the global menu.
Why not just condense it all in to one bar?
To be honest there are few good reasons for the global menu and plenty of
problems (touch, large monitors, multi monitors) that will seriously degrade
the users experience.
From: Ian Santopietro <isantop@xxxxxxxxx>
Date: Mon, 4 Apr 2011 14:57:21 -0600
To: giff g <giffgilll@xxxxxxxxxxx>
Cc: <ayatana@xxxxxxxxxxxxxxxxxxx>
Subject: Re: [Ayatana] why global menubar/application menu isn't such a
great idea
There is no definitive fact that says that Google knows best. They have
their preferences about UX, and Canonical has their's. Just because these
two entities don't agree doesn't make one or the other right or wrong.
Most of Canonical's usability testing seems to indicate that it's easier to
hit the Gobal menu. It's at the edge of the screen, so you only need to aim
along one dimension. Plus, the first (Typically File) menu is in the exact
same place every single time, even between a maximized vs. restored window.
I've been using Unity since Alpha 3, and while the global menu isn't
perfect, it is better than what we had before.
Chrome and Firefox do it wrong, IMO. I use the global-menu firefox
extension, and wish I could do that with Chrome. Cramming all of that menu
into a single button is not ergonomic.
The top panel displays a lot of information, including the menu, BFB, and
indicators. most windows still have titlebars (Including Firefox 4). That
won't be changing. Putting the menu there saves space because you don't need
a menu bar or menu button anywhere else; it's all up there. The Show on
hover is not great, usability wise, but there aren't a whole lot of viable
alternatives. There have been some good exceptions, but with Unity at it's
current state, I don't think it's realistic to try an reimplement that much
code in such a short time.
Your "Menus are outdated" arguement is invalid. There are lots of outdated
items in the current Desktop Metaphor that are outdated, and revolution
isn't the way to go there. Evolution keeps users much happier.
2011/4/4 giff g <giffgilll@xxxxxxxxxxx>
>> > From: conscioususer@xxxxxxx
>> > To: ayatana@xxxxxxxxxxxxxxxxxxx
>> > Date: Mon, 4 Apr 2011 14:53:45 -0300
>> > Subject: Re: [Ayatana] why global menubar/application menu isn't such a
>> great idea
>
>> >
>
>> > While I disagree with Mitja's tone (as usual), I agree with
>> > his main point. Most of giff's points were based on general
>> > assumptions backed up by little more than anecdotal evidence.
>> > And anecdotal evidence is easily countered: less than a week
>> > ago a user in this list mentioned how he had no problems with
>> > using OSX in a large HD monitor, for example.
>
> I said as much and hoped I made it pretty clear that the
> points I rise are only opinions that matter to me and how I use
> computer interfaces.
>
> I also hinted at criticising how everything the Unity team is doing is
> based on personal experiences, anectotes, preference etc. instead of
> how it should be done: scientifically, with hard data and a large
> data set. Mozilla and Google know this and that's why I emphasized
> their browser UI and the testing that went into that.
>
> So in turn I'd expect that my opinions are accepted for what they are
> and well, pot and kettle (@Mitja)
>
>> > The existence of things like DejaMenu is hardly convincing
>> > evidence either, specially in the Linux ecosystem where there
>> > are hacks for anything and everything.
>
> I agree but I also provided a reason why I'm bothering about
> this particular "hack": The defaults should be the best possible
> compromise, I provided some reasons why the defaults now aren't
> the best default - for large screens.
>
>> > Also, giff mistakenly uses an old post about the original
>> > Unity as an argument, ignoring the fact that netbooks are not
>> > the primary target anymore, effectively invalidating some of
>> > his points from the very beginning.
>
> I disagree. First of all I know that Unity is intended for all sorts of
> devices and form factors (hence the point 2 about Desktops).
> Secondly, that change in my eyes only aids my arguments:
> On small screens conflating titlebar and menubar and decreasing window
> hight is a worth trade-offs like multitasking and the eye focus problem
> doesn't exist. The fact that the global menu is seen as the best possible
> choice for 30 inch displays is what I argued against.
>
> Does that change invalidate the articles quoted?
> Netbooks are still an important target and so are other form-factors
> like tablets. More important than desktops simply because of market
> share and growth rate.
>
> Anyway the "net" centric computing will only increase, no matter
> what device you use.
>
>> > The rest of the text is mostly questionable, with some apparent
>> > contradictions, both internal (ex: emphasizing how unnecessary
>> > the menu is, while complaining about the global menu making it
>> > slow)
>
> Some parts of my post are probably a bit unstructured and could need
> some editing to clear up some points...
>
> In response to that apparent contradiction:
> When I talk about how the text menu is becoming obsolete I have
> native OS X applications in mind, modern "apps" written in Cocoa.
> Not "legacy" gtk2 programs that haven't seen an interface change in
> years or I have Windows 7 in mind, where IE, Office, built in programs
> like Wordpad and Paint switched to a menubar-less interface.
> KDE is toying with going into that direction as well I heard. Then
> Firefox and Chrome. Point 1) is my primary concern and that's why I
> filled a bug against it.
>
> Now, for those programs that really do need the menu and the
> menu has to be accessed frequently it's a different matter. In full
> screen application the menu should be at the top like Unity does.
>
> Generally, the valuable screen estate at the screen edges should be
> reserved for the most frequently accessed interface elements and
> not wasted with a large title bar for example (The office ribbon
> doesn't get this right).
>
> The lower screen edge is equally important and with Unity freeing it,
> it's up to the application developers to make clever use of it instead
> of for example waisting it with a statusbar nobody needs (because
> you always put a statusbar there, right? The statusbar is a good
> example. Up until Chrome ALL browsers had one, now they are
> all replacing it with temporary url previews. Just because it's old and
> tried doesn't mean it's "the best possible solution". I extend that
> to the concept of global menubars.)
>
> The problem with the global menu is that it's static, so even for those
> apps that don't need one it's there. It's taking up space and wasting
> the preciouse screen edge area. For those apps it is "unnecessary".
>
> The problem with "slowness" really only comes into play when
> talking about multi-tasking.
>
>> > and external (ex: complaining how prominent it is, while
>> > a lot of people are complaining about not being prominent
>> > enough due to the show-on-hover).
>
> I didn't go into that and I think it's a separate issue altogether.
> But I don't see how that contradicts anything I said.
>
> Just take look at my Chromium and Firefox example:
> You don't need the window title and you don't need the menu.
> In either cases this is lost screen estate and tabs are harder to
> access than on Windows, KDE and Chrome OS.
> Hover or not changes nothing.
>
> But if you want my opinion on that specific issue:
> show-on-hover is a bad design, period. Most HIGs will agree with
> my opinion here...
> It's a noisy interface, not exactly discoverable and the main
> advantage of the menubar (items are in predictable places, use
> muscle memory) is lost. For full screen applications the window title
> isn't very necessary and for tabbed applications (which are numerous and
> probably growing, browsers, text editors, photo editors, file managers)
> and windows with an address/location bar it's not needed at all.
>
>> > Overall, the points are not clear from a realistic point of
>> > view. At the end of the day, it seems the main point of the
>> > text is "menus will die someday, so let's pretend this day has
>> > already arrived and move from there", which kinda... doesn't
>> > work in real life. :)
>
> My main point is bug #749335
> Firefox and Chrome are not some day, they are now.
> For me the browser is the most frequently used applications
> and I deeply care about getting the best possible user experience there.
>
> Google who best know about hard data as opposed to opinions
> is writing a whole OS without a menubar. They do get UX.
> The way Ubuntu is positioning itself today and the way I predict
> how the OS landscape is going to be in the coming year I think
> Chrome OS is going to be replace a certain someone in bug #1.
>
>
>
>
> _______________________________________________
> Mailing list: https://launchpad.net/~ayatana
> <https://launchpad.net/%7Eayatana>
> Post to : ayatana@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~ayatana
> <https://launchpad.net/%7Eayatana>
> More help : https://help.launchpad.net/ListHelp
>
--
Ian Santopietro
"Eala Earendel enlga beorohtast
Ofer middangeard monnum sended"
Pa gur yv y porthaur?
Public GPG key (RSA):
http://keyserver.ubuntu.com:11371/pks/lookup?op=get&search=0x412F52DB1BBF123
4
_______________________________________________ Mailing list:
https://launchpad.net/~ayatana Post to : ayatana@xxxxxxxxxxxxxxxxxxx
Unsubscribe : https://launchpad.net/~ayatana More help :
https://help.launchpad.net/ListHelp
Follow ups
References