LWN.net Logo

Day: A New Approach to GNOME Application Design

GNOME design team member Allan Day writes about ideas in GNOME 3 application design on his blog. In the article, he looks at the use of maximized windows, views, primary toolbars, and more. The design team is documenting these ideas in a new version of the GNOME Human Interface Guidelines (HIG). "There are many other application design patterns that we've been working on, including application menus, a new grid view for displaying collections of content, in-app notifications, new models for dialogs, nice full screen controls and a sidebar list pattern. Together, these provide the opportunity to create applications that efficient, modern, elegant, and a pleasure to use."
(Log in to post comments)

Day: A New Approach to GNOME Application Design

Posted Feb 11, 2012 10:31 UTC (Sat) by krakensden (subscriber, #72039) [Link]

> Windows that aren’t maximised also create additional tasks for people. Often you need to adjust their size, or you have to move them around.

See, if you hadn't broken alt-tab and virtual desktops, it wouldn't be a problem...

Day: A New Approach to GNOME Application Design

Posted Feb 11, 2012 13:58 UTC (Sat) by drag (subscriber, #31333) [Link]

Compared to other desktops the Gnome shell is the only one with a "alt-tab" that isn't broken. And the virtual desktops are very good. It's called 'progress'.

Day: A New Approach to GNOME Application Design

Posted Feb 11, 2012 16:36 UTC (Sat) by Pawlerson (guest, #74136) [Link]

I beg to differ. It's not broken in KDE and Windows and I don't know about Gnome.

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 4:22 UTC (Sun) by drag (subscriber, #31333) [Link]

> I beg to differ. It's not broken in KDE and Windows and I don't know about Gnome.

Well then you should try to know something about it then. It's not difficult.

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 10:12 UTC (Sun) by Pawlerson (guest, #74136) [Link]

I played with Gnome3.0 for a while, but it was hard to use, so I'm sticking with KDE.

Day: A New Approach to GNOME Application Design

Posted Feb 11, 2012 16:59 UTC (Sat) by tjc (guest, #137) [Link]

> It's called 'progress'.

Not for me; it's a regression.

I've been using OS X at work for nearly 2 years now (ever since my SPARCstation was decommissioned), so it's not like I haven't given the document-centric paradigm (which GNOME aped from Apple) a chance. Two years later and it's still an impediment to my workflow.

Day: A New Approach to GNOME Application Design

Posted Feb 11, 2012 20:48 UTC (Sat) by krakensden (subscriber, #72039) [Link]

Only if you define "progress" as "more like OS X". Besides, it kind of ignores what people use Linux for:

1) a cheap, manageable replacement for Windows in corporate environments
2) nerd shit

GNOME Shell totally hoses the first demographic, it's too different and they can't/won't change.

GNOME Shell *partially* hoses the second, because they're more willing to learn, and they already know about the OS X model.

Besides, it doesn't make any sense to me to embrace the OS X document-centric model when the only application in wide use that benefits from it is the GIMP. They broke terminals for the GIMP.

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 4:21 UTC (Sun) by drag (subscriber, #31333) [Link]

> Only if you define "progress" as "more like OS X".

No. I define 'progress' as 'getting better'. Which is what is happening.

> Besides, it kind of ignores what people use Linux for:

No it doesn't.

> 1) a cheap, manageable replacement for Windows in corporate environments

It's not 'cheap', nor is it 'manageable', nor it is a replacement for Windows. At least for desktops. Centralized management and configuration is still difficult, full of pitfalls, and expensive in practice.

If Dconf gets integrated support from something like FreeIPA we will finally start getting close to what Windows offers to corporations. Then instead of being 20 years behind we will only be 10.

> 2) nerd shit

Oh, I am sorry if a fully scriptable window manager, with built in extension/plugin support and integrated debugger is not nerdy enough for you.

> GNOME Shell *partially* hoses the second, because they're more willing to learn, and they already know about the OS X model.

Nothing you said there makes any sense.

> They broke terminals for the GIMP.

Funny. My terminals work just fine. And that still doesn't make any sense.

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 20:47 UTC (Sun) by xan (subscriber, #58606) [Link]

> Only if you define "progress" as "more like OS X". Besides, it kind of ignores what people use Linux for:
> 1) a cheap, manageable replacement for Windows in corporate environments
> 2) nerd shit

So now anyone doing UIs for Linux must strictly target those demographics or get lost? So much for "Linux is about choice", not to mention that is a self-fulfilling prophecy the size of Jupiter.

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 22:20 UTC (Sun) by daniels (subscriber, #16193) [Link]

BUT LINUX IS ABOUT CHOICE

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 23:02 UTC (Sun) by josh (subscriber, #17465) [Link]

For anyone who doesn't get the reference: *please* read https://www.redhat.com/archives/rhl-devel-list/2008-Janua...

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 0:13 UTC (Mon) by ebassi (subscriber, #54855) [Link]

or, in conveniently pre-packaged URL: http://islinuxaboutchoice.com

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 0:21 UTC (Mon) by josh (subscriber, #17465) [Link]

Thanks! I remembered seeing that version a while ago, but I'd never managed to find it again.

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 16:59 UTC (Sun) by coriordan (guest, #7544) [Link]

Not being able to Alt-Tab between two terminals, means Alt-Tab is broken.

There are lots of nice things about GNOME3, and I'm sticking with it, but there are also some mistakes that I'm hoping get fixed with less of a delay than spacial Nautilus took.

(A partial fix to the broken Alt-Tab: I installed three different terminal applications. Of course, it's annoying when you want a fourth terminal, and this solution is too clumsy for most computer users to accept.)

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 17:17 UTC (Sun) by drag (subscriber, #31333) [Link]

you can 'alt-tab' between terminals with 'alt-~'

However I can't alt-tab between different applications in any other desktop if I happen to choose the wrong icon in the alt-tab display the first time...

With Gnome-shell I not only get to choose which window I alt-tab too, I get previews and grouping. And I still get the ability to alt-~ between terminal windows.

What is so horrible about this? What, exactly, is broken?

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 19:14 UTC (Sun) by drag (subscriber, #31333) [Link]

Since people have trouble coming to grips with changes to the Microsoft Windows style alt-tab behavior I made a quick screencast demonstrating Gnome-shell's behavior:

Gnome Shell alt-tab demonstration
http://www.youtube.com/watch?v=kH8F4oWLR9M

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 20:21 UTC (Sun) by Kit (guest, #55925) [Link]

The combination of alt-tab and alt-` made it work a lot nicer than I would have expected (since I hadn't heard about alt-`).

But, even still, the grouping to me is illogical. To me, different windows for a single application rarely have anything to do with each other. I'm a bit frustrated by this behavior on OS X (I generally just use 3-finger swipe for expose, instead), since I'll often have, say, a text editor, a browser window, and a terminal that I switch between, all part of one 'task' (i.e. writing code).

On the occasions that I use cmd+tab, I'll often find myself going to the /wrong/ browser window (that's what tends to hit me the most, or the file manager), on the wrong virtual desktop. I, personally, prefer the traditional Linux/Windows way of simply enumerating the windows, which has never been an issue for me.

I'd imagine if you got into the habit of using both alt+tab and alt+` together that it wouldn't really be an issue in practice, but it still seems to be the wrong grouping (grouping by virtual desktop would make more sense based on how I use apps).

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 20:45 UTC (Sun) by rahulsundaram (subscriber, #21946) [Link]

If you prefer the old alt+tab behaviour, use this extension

https://extensions.gnome.org/extension/15/alternatetab/

In general, if your workflow is different from the standard GNOME 3 one, extensions do address that well. I suggest browsing through the available ones. A few minutes there would be worth investing.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 16:08 UTC (Mon) by dashesy (subscriber, #74652) [Link]

I appreciate all the good effort you make for FOSS, I really hate to complain, but couldn't you leave Alt+Tab alone and invent new UI using let's say Win+Tab.
It says Windows so it might be used for something Window related.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 16:15 UTC (Mon) by rahulsundaram (subscriber, #21946) [Link]

I am not a GNOME developer. Just a user explaining how to get what some other users might want.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 14:39 UTC (Mon) by hmh (subscriber, #3838) [Link]

Well, you should probably investigate the Activity concept of KDE4, then. You can fast-switch between activities using the keyboard (or icons, etc), and it works sort of like virtual desktops with extra functionality.

It is orthogonal to the other fast-focus-switch axis, and groups things by, well, activities. I am quite sure the concept has not been explored to its best yet, I need to test the latest KDE to see how much it improved lately.

Maybe they can add activities or something like that to the gnome-shell, it is a far superior way to group a set of *windows* (the same application can have windows that belong to different activities, after all. In fact, it is almost the rule for the browser, text terminals and text editors).

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 17:22 UTC (Mon) by bjartur (guest, #67801) [Link]

If you ever happen to have more than three top-level windows open, you might want to consider tagging.[1]

The Windows way simply does not scale.

1: http://dwm.suckless.org/

Will other new users find this intuitive?

Posted Feb 12, 2012 23:40 UTC (Sun) by coriordan (guest, #7544) [Link]

Thanks for the video.

That improves GNOME3 significantly for me. I'm not sure if this window-switching system is an improvement on GNOME2, but I'll get used to it.

But the bigger question is: Will most other new users experience the same frustration as I did? Or are these key combos already well known to users of the major desktops?

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 12:40 UTC (Mon) by niner (subscriber, #26151) [Link]

Brilliant. On a German keyboard layout that means I'd have to press Alt with the left hand and Alt Gr + ~ on the right hand. For sure a usability improvement over just two buttons on one hand.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 12:45 UTC (Mon) by mjg59 (subscriber, #23239) [Link]

No, just press alt and the key above tab. It's based on scancode, not actually the ` mark.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 19:38 UTC (Mon) by nix (subscriber, #2304) [Link]

The only way you can identify 'the key above tab' is via the geometry file. So this is now depending on an up-to-date geometry file?!

If my keyboard had one (it doesn't), that would be tab and, uh, ctrl. Alt-Ctrl? I don't think so. Since it doesn't have one, it's tab and `, a key halfway across the keyboard.

A great user interface innovation this is not. It might work for people with standard keyboards. People with ergonomic keyboards can just sod off into their little boxes of pain I suppose.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 19:44 UTC (Mon) by mjg59 (subscriber, #23239) [Link]

Then remap it yourself. At the point where you're using non-standard input devices you're going to have to take some responsibility for handling the consequences. It's utterly impractical to expect developers to cater for every weirdness that exists.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 21:35 UTC (Mon) by dgm (subscriber, #49227) [Link]

Please, can you express that in a more civilized way?

For instance, you could say: "I'm terribly sorry for you, but the amount of resources we have is not enough to please every one, even if we would be happy to do so". See? Much better that saying "It's your keyboard, do not bother me." And it doesn't make you seem a moron.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 22:57 UTC (Mon) by nix (subscriber, #2304) [Link]

Alternatively I can simply say 'GNOME doesn't care about me, I don't have to care about it or use it'. Which is what I've done.

IMNSHO, anyone using XKb geometry data for functional effect is a madman. Keyboard geometry differs in vastly many ways: the geometry data does not capture any but the most popular. Nor will it ever, *because* next to nothing uses it and it's almost undocumented and horrible to work with.

Day: A New Approach to GNOME Application Design

Posted Feb 14, 2012 0:26 UTC (Tue) by mjg59 (subscriber, #23239) [Link]

It doesn't use the geometry data - you're right that it would be of approximately no use. The default binding is to whatever key generates X keycode 49. There's a couple of more mainstream exceptions to that being the key just above tab - older Toshiba laptops (where it's the key next to alt) and some Sun type 6s. But otherwise, it's in a natural location that matches user expectations.

Once you start moving to non-traditional keyboards you're inevitably going to find cases where designer expectations aren't borne out by reality. There's only really two choices there - refuse to standardise on any new shortcuts or accept that you're always going to fall short of immediate satisfaction on more niche hardware and assume that the user will be able to cope.

Day: A New Approach to GNOME Application Design

Posted Feb 14, 2012 1:36 UTC (Tue) by nix (subscriber, #2304) [Link]

Boggle. They hardwire the keycode number?! (or even softwire it? normal users can't be expected to dig out keycodes or even know what they are in order to fix such things...)

That is utterly appalling (though it is true that there is probably no better way: we probably need one: for starters a simple binding of keyboard name/nationality to expected keycode number would do the trick, but that's a horrible special-purpose hack).

Keycode 49, hm, for me that is indeed `, at the upper right of the keyboard and basically nowhere near tab, but this is probably because Maltron (and Kinesis) keyboards specifically fake the keycodes of a normal QWERTY keyboard for compatibility's sake.

Looking at a bunch of national keyboard layouts for vaguely Latinate scripts on Wikipedia, the assumption that ` is above Tab will fail on French Canadian, Danish, Dutch, Estonian, Faroese, Finnish, Italian, Norwegian, Farsi (stretching a bit to call it Latinate), Portguese (non-Brazilian), Spanish (including Latin American), Swedish, Turkish, Hungarian, German, Swiss, and Belgian standard keyboard layouts, and I probably missed a few and Wikipedia is doubtless missing some anyway.

Now maybe some of the keyboards have defined 'the key above tab' to still be keycode 49, and maybe they have not. But it seems... unlikely in the extreme that *all* of them have. That's probably several hundred million people being mis-served here by one silly user interface decision. It's not a few pathetic RSI sufferers like me who can just be left to rot like the blind users. It's whole countries (heck, if the Spanish keyboards get it wrong, most of a continent). Unless you're lucky and all QWERTYish keyboards bind keycodes entirely independently of the keycap and nationality and based only on which US keyboard key they are closest to if you overlay the two keyboard layouts. Maybe you're lucky. But I doubt it.

Day: A New Approach to GNOME Application Design

Posted Feb 14, 2012 2:44 UTC (Tue) by mjg59 (subscriber, #23239) [Link]

Yeah, the biggest problem with this is the "Oh, use alt-backtick" - which is obviously inaccurate for a huge number of locales. But "Key above tab is keycode 49" is common to the vast majority of keyboards, independent of locale. Vendors will change keycaps to suit local markets, but it's way more expensive to change the keyboard logic.

As you say, there are various niche keyboards which break this assumption. But they really are at the niche end of things, and once we're into heavy ergonomics we're really talking more about accessibility support than we are about UI design. Accessibility isn't a one size fits all problem. The keycode used for this is easily reconfigured using the standard keyboard shortcut interface, but it probably is worthwhile considering it in terms of integration with other accessibility functionality - there may be other shortcuts that are also non-optimal in common ergonomic scenarios.

Day: A New Approach to GNOME Application Design

Posted Feb 14, 2012 11:49 UTC (Tue) by nix (subscriber, #2304) [Link]

But "Key above tab is keycode 49" is common to the vast majority of keyboards
In a good few of those countries, alt-` (or altgr-`) is necessary to produce characters essential for writing text in the local language...

Day: A New Approach to GNOME Application Design

Posted Feb 14, 2012 13:07 UTC (Tue) by michaeljt (subscriber, #39183) [Link]

> But "Key above tab is keycode 49" is common to the vast majority of keyboards, independent of locale.

Quibble: on many (all?) Macbooks it is 94. For some reason they chose to swap those two keys. Don't know about other Intel Macs.

Day: A New Approach to GNOME Application Design

Posted Feb 14, 2012 16:51 UTC (Tue) by dashesy (subscriber, #74652) [Link]

Sorry to repost this here, why not leaving Alt+Tab alone (old expected functionality), and use Win+Tab for the all-shiny new UI? It even says Windows, it would be great to use it for handling Windows, it is usually closer to the Tab and ` keys, and the best is that people do not feel forced using a foreign UI.

Day: A New Approach to GNOME Application Design

Posted Feb 14, 2012 12:57 UTC (Tue) by michaeljt (subscriber, #39183) [Link]

> That is utterly appalling (though it is true that there is probably no better way: we probably need one: for starters a simple binding of keyboard name/nationality to expected keycode number would do the trick, but that's a horrible special-purpose hack).

If you want some amusement (and have a good stomach - the code started off as a cut-and-paste from Wine and has mutated out of recognition since then), take a look at the X11 keyboard handling code in VirtualBox [1]. We check geometry. We check keycode numbers. We check key symbols and compare them with known keyboard layouts. Then we try to put it together to attach a PC scan code to a key (which may not be on a PC keyboard).

[1] https://www.virtualbox.org/browser/vbox/trunk/src/VBox/Fr...

Day: A New Approach to GNOME Application Design

Posted Feb 14, 2012 23:50 UTC (Tue) by nix (subscriber, #2304) [Link]

Yeah, virtualization is an area where we *do* have a need to drive X backwards and get back to a key, and then somehow manage to get this to the system inside the virtual machine and repeat the keystroke in some 'equivalent' fashion. SPICE, xvnc, and QEMU's SDL layer do it with hardwired key lists. This is... not at all ideal.

What VirtualBox is doing looks a lot better, but is still a most appalling kludge. I have no immediate idea how to do it properly :( you can do it with heuristics and big tables, and I'm sure that VirtualBox is either an argument for or against that. I'm not sure which. :)

Day: A New Approach to GNOME Application Design

Posted Feb 14, 2012 0:00 UTC (Tue) by bronson (subscriber, #4806) [Link]

If it was an obscure chording keyboard then you might have a point. But, since it works in Gnome 2 and other DEs, I don't think you can so carelessly cast the blame entirely on the user. Where do you draw the line?

I can see it now... requirements for Gnome 4 will be a 1366x768 screen, US layout Microsoft 105-key keyboard, and some sort of touchpad. Anyone running anything else will, as you say, be on their own.

Day: A New Approach to GNOME Application Design

Posted Feb 14, 2012 1:42 UTC (Tue) by nix (subscriber, #2304) [Link]

Well, a Maltron *is* about as obscure a keyboard as you can find: in thirty years they've made about ten thousand of them. But Kinesis use a very similar layout and have sold a lot more (because they're cheaper and don't look quite so much like a tank).

It works in every DE you can find, because it fakes the keycodes of a UK 105-key keyboard precisely. Other national keyboard layouts almost certainly won't go to that trouble (because they don't need to: they're not minority devices used by those who have no other choice). The only thing it won't handle is environments that assume a certain mapping between keycode and keyboard geometry. To date, I have heard of none but GNOME 3.

(It's not so good for certain games that assume a particular geometric layout of keys, but where using a normal 105-key keyboard for the odd game is acceptable, using a 105-key keyboard for *task switching* would not be. Now I'm an Emacs user: I can cope with random weird key assignments half a keyboard apart to cycle in different directions. But normal users who just happen to be RSI-afflicted and thus forced into using a Kinesis or Maltron keyboard? They'll throw GNOME 3, and any desktop that makes similar assumptions about keyboard geometry, to one side with great force.)

Day: A New Approach to GNOME Application Design

Posted Feb 14, 2012 0:45 UTC (Tue) by mathstuf (subscriber, #69389) [Link]

Out of curiosity, how does this work when two keyboards with different layouts are available (e.g., a native keyboard and also US keyboard)? Does each keyboard have a different keystroke to trigger the action? Can one keyboard trigger the action with the other keyboard's geometric binding?

Day: A New Approach to GNOME Application Design

Posted Feb 14, 2012 2:39 UTC (Tue) by mjg59 (subscriber, #23239) [Link]

It's bound to the keycode rather than the keysym, so assuming that both keyboards are pc105-style, it'll be the same physical key even if both keyboards have different keymaps.

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 19:07 UTC (Sun) by quotemstr (subscriber, #45331) [Link]

For me, different terminals often represent entirely different applications. I don't want my window manager to start guessing which top-level windows represent different tasks in the same application and which represent different applications.

I want it to just friggin switch between top-level windows.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 0:01 UTC (Mon) by drag (subscriber, #31333) [Link]

> For me, different terminals often represent entirely different applications.

Well Gnome-shell already has the idea about 'web apps'. That is you can use Epiphany/Gnome-shell integration to launch web apps into their own 'top level' application window. So that your web mail can be treated by Gnome-shell as a separate application from your regular web browsing activities.

It makes all the sense in the world that you should be able to treat the terminal in the same way. So you should be able to put 'mutt' into it's own application or 'emacs' or 'vim' (I actually prefer a terminal version of both apps over GTK)

Sounds like a sensible thing to file a feature request bug over. Probably can be a nice enhancement.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 17:57 UTC (Mon) by sorpigal (subscriber, #36106) [Link]

It's so sensible it should be the default. An application should be required to hint to the WM whether its windows should be treated as separate apps or windows within an app, and when they don't hint the default should be "behave the way people expect" which is the legacy "everything is its own top level window" behavior.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 21:49 UTC (Mon) by oak (guest, #2786) [Link]

Isn't the WM_CLASS property intended for this?

Day: A New Approach to GNOME Application Design

Posted Feb 11, 2012 11:21 UTC (Sat) by ssam (subscriber, #46587) [Link]

I guess gnome is due for a huge redesign. its has not had one of thoses for several months now.

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 4:36 UTC (Sun) by bojan (subscriber, #14302) [Link]

And I have a working design for them: the Ctrl + Alt + Fn.

You get fully maximised "windows" all the time. There are no distractions on the screen apart from the app you are running.

Never mind cut'n'paste from window to window, drag and drop and all the other stupid things. You don't need that mouse of yours. You don't need to be looking at a spreadsheet while doing your tax return. You don't need to see several shells at one. All that is for losers.

;-)

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 0:21 UTC (Mon) by tzafrir (subscriber, #11501) [Link]

Copy&Paste does work (between text-mode virtual consoles) assuming you use gpm. Never tried cut&paste, though.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 2:19 UTC (Mon) by jmalcolm (subscriber, #8876) [Link]

He means that it is much less convenient to work with multiple windows at once if the metaphor is assumed to be "full screen all the time". Many people find it quite functional to have multiple windows open and visible at once both for moving information via mechanisms like cut-and-paste and for moving information through the eyes of the user and out to another window via the keyboard and mouse.

Maximized by default == FAIL

Posted Feb 11, 2012 14:16 UTC (Sat) by dskoll (subscriber, #1630) [Link]

"Displaying multiple windows at the same time means that screen space isn’t used efficiently, and it means that you don’t get a focused view of what it is that you are interested in."

There's a big fail. Because we know how easy it is to drag-n-drop something from one completely-maximized window to another... We also know how no-one ever needs to run something in one window and copy/paste it into another.

Maximized by default == FAIL

Posted Feb 11, 2012 14:59 UTC (Sat) by aleXXX (subscriber, #2742) [Link]

...or see multiple windows at once, now that we have wide displays.
Like an editor, a web or man page, and a terminal.

I never use maximized windows.
That's surely different on a tablet and maybe on netbooks.

Alex

Maximized by default == FAIL

Posted Feb 11, 2012 16:25 UTC (Sat) by drag (subscriber, #31333) [Link]

> There's a big fail. Because we know how easy it is to drag-n-drop something from one completely-maximized window to another... We also know how no-one ever needs to run something in one window and copy/paste it into another.

I don't see why it's a big fail. Designing applications so that they are able to take proper advantage of being full screen is not a terrible thing.

There is no reason why it's safe to assume that he is talking about only allowing one window to be open at a time or being displayed at one time; if that is what you are doing.

Maximized by default == FAIL

Posted Feb 11, 2012 19:46 UTC (Sat) by luya (subscriber, #50741) [Link]

Sugar Interface anyone?

Maximized by default == FAIL

Posted Feb 12, 2012 3:24 UTC (Sun) by csigler (subscriber, #1224) [Link]

> "Displaying multiple windows at the same time means that
> screen space isn’t used efficiently, and it means that
> you don’t get a focused view of what it is that you are
> interested in. Windows that aren’t maximised also create
> additional tasks for people. Often you need to adjust
> their size, or you have to move them around."

As a *workstation* interface, then, GNOME's new target audience, is, what, 8 to 12 year-old developers? Having two windows open at once, the need to occasionally move them around or resize them, is too complicated for computer users? Having two windows open at once may require a user to perform two tasks in tandem and that's more than anyone should be capable of?

I hope I can be forgiven for saying it: "GNOME is for babies."

Clemmitt

Maximized by default == FAIL

Posted Feb 12, 2012 4:25 UTC (Sun) by drag (subscriber, #31333) [Link]

> As a *workstation* interface, then, GNOME's new target audience, is, what, 8 to 12 year-old developers?

If those developers have a desire to manage their desktop workflows using custom scriptable, extendable, composted desktop environment... then yes.

Maximized by default == FAIL

Posted Feb 12, 2012 18:08 UTC (Sun) by dskoll (subscriber, #1630) [Link]

...composted desktop...

Now there's a Freudian slip if ever there was one.

Maximized by default == FAIL

Posted Feb 12, 2012 21:56 UTC (Sun) by droundy (subscriber, #4559) [Link]

Yes, desktop environments should *always* be composted, rather than sent to the landfill!

Screen size is the problem

Posted Feb 11, 2012 15:32 UTC (Sat) by bredelings (subscriber, #53082) [Link]

Having all applications maximized makes sense when your screen is small.
  • What does it mean to "maximize" an application on a dual-monitor setup?
  • Even for big monitors... do you really only want to let one thing be visible at a time?
I can see the appeal for cell-phones and tablets, though.

Screen size is the problem

Posted Feb 11, 2012 16:20 UTC (Sat) by epa (subscriber, #39769) [Link]

I have a very large monitor and I still run everything maximized. Partly this is because, well, what more than Emacs do you need? But partly it's because of the horrible click-to-raise behaviour where any interaction with an application window brings it to the front. This makes working with overlapping windows very painful - and nor do I have time to carefully tile the windows so they don't overlap.

If non-maximized, overlapping windows were a bit more usable (e.g. bring to front only with a click on the titlebar) then there would be less need to keep everything fullscreen all the time. It's probably too late, though; nontechnical users have never really learned about multiple-window interfaces and now they're probably never going to.

Screen size is the problem

Posted Feb 11, 2012 19:57 UTC (Sat) by sfeam (subscriber, #2841) [Link]

What you describe is not a problem with click-to-raise. It is a problem with not selecting focus-follows-mouse. FFM is the one true paradigm [TM] precisely because it allows you to type into an overlapped window without bringing it to the front.

Screen size is the problem

Posted Feb 11, 2012 20:12 UTC (Sat) by mpr22 (subscriber, #60784) [Link]

All non-telepathic focus and raise policies (click-to-focus, focus-follows-mouse, focus-by-hotkey; raise-on-focus, raise-on-click, raise-on-title-click) of which I am aware are broken, including the ones I like. Anyone who says one of them isn't merely isn't impacted by the breakage case for that model.

Focus-follows-brain

Posted Feb 12, 2012 12:41 UTC (Sun) by tialaramex (subscriber, #21167) [Link]

Indeed, focus-follows-brain is the only correct model. But while we wait, sloppy-focus turns out to be a pretty close substitute for me, I curse and shake my fist only once a week or so (whereas on a newly installed GNOME 3 machine it took only minutes before I was fumbling around trying to find sloppy-focus and yelling expletives).

The continued existence, indeed popularity, of click-to-focus bemuses me. Mostly because when you watch ordinary users‡ through a one-way mirror they're just as baffled by the consequences of click-to-focus as I would be, their experience is basically "sometimes it doesn't work and you have to click twice".

‡ ie not people who know what "focus" is, let alone being capable of a discussion about the desirability of focus-follows-brain.

Screen size is the problem

Posted Feb 11, 2012 22:08 UTC (Sat) by epa (subscriber, #39769) [Link]

Focus-follows-mouse is fine if all you do is type, but as soon as you click in a window to interact with it using the mouse, bam! it jumps to the front. I would rather that just interacting with a window (whether that be using mouse or keyboard) did not always force it to the front whether you like it or not.

Screen size is the problem

Posted Feb 11, 2012 23:52 UTC (Sat) by speedster1 (subscriber, #8143) [Link]

> Focus-follows-mouse is fine if all you do is type, but as soon as you click in a window to interact with it using the mouse, bam! it jumps to the front. I would rather that just interacting with a window (whether that be using mouse or keyboard) did not always force it to the front whether you like it or not.

Window-raise policy is orthogonal to focus-follows-mouse. My favorite setup has focus-follows-mouse on, and window-raise when clicking window frame or alt-click anywhere in window.

I think gnome 2 supported both policies, using focus_mode=sloppy and raise_on_click=false; not sure how one can achieve this with gnome 3 (maybe with a gnome-shell extension?)

Screen size is the problem

Posted Feb 12, 2012 14:05 UTC (Sun) by aleXXX (subscriber, #2742) [Link]

kwin also supports all these focus and raise modes.
I think it should be possible to use kwin outside a KDE desktop.

Alex

Screen size is the problem

Posted Feb 13, 2012 18:01 UTC (Mon) by sorpigal (subscriber, #36106) [Link]

If you want this it's entirely possible to have it and, in fact, until recently it was the norm to have this. Configuring raise only on explicit titlebar click is trivial in almost any WM released in the 90s, for example. Maybe you really want to be running FVWM?

Screen size is the problem

Posted Feb 11, 2012 20:12 UTC (Sat) by bjartur (guest, #67801) [Link]

The problem you describe is not with stacking but with overlapping. Have you considered using a tiling window manager?

Overlapping windows usually means that content is obscured and screen space is left unused *and* you have to spend time moving windows around and resizing.

Screen size is the problem

Posted Feb 11, 2012 22:11 UTC (Sat) by epa (subscriber, #39769) [Link]

The problem is that with current window manager behaviour, overlapping is a pain to use. This has led people to avoid it by adopting strategies such as maximizing everything, or tiling the windows (which a tiling WM helps with).

My point is that with sane window-raising behaviour, overlapping windows are more usable and become a viable choice along with tiling etc.

Tiling WM

Posted Feb 12, 2012 21:18 UTC (Sun) by Yenya (subscriber, #52846) [Link]

I, for example, have tried a tiling WM briefly, but found it unusable for me. The main reason is that for two of my most important applications (mutt for e-mail and vim for code editing) having windows with exactly 80 characters width is strictly necessary.

Anyway, window overlapping is not a problem, as long as it is possible for any window (including partially obscured one) to receive a keyboard input, without being raised at the same time.

Recently I have added "switch focus to window on the right/left/up/bottom of the currently focused window" hotkeys to my WM setup, and it greatly improved my workflow.

Disclaimer: I am long time user of GNOME, but with GNOME 3 I have switched to XFCE with Sawfish as my WM.

Tiling WM

Posted Feb 13, 2012 2:40 UTC (Mon) by jmalcolm (subscriber, #8876) [Link]

Off-topic but do you mind me asking why you settled on Sawfish? At one time, it was the default window manager for GNOME but that is an archeological topic at this point.

Do you use any Sawfish specific features? Do you using rep (LISP) at all?

Tiling WM

Posted Feb 13, 2012 11:30 UTC (Mon) by Yenya (subscriber, #52846) [Link]

I stayed with sawfish when GNOME moved to Metacity and at that time did not allow me to configure things the way I wanted them to be (vertical maximization for terminals, automatic placement of some windows to particular virtual desktops, etc.). After moving to XFCE I have stayed with sawfish mostly because I did not want to spend more time configuring yet another WM.

Sawfish is still being developed, BTW.

In my configuration, there is no LISP code I have written (except the
configuration file, which can be modified even without any LISP knowledge).
I have some plugins in LISP, which I have downloaded - Focus-by-direction.jl, for example, is in ~/.sawfish on all of my machines.

Apart from that, I appreciate the extreme configurability - smart window matching, hot-keys, etc. Sawfish-ui is really easy to use, so I don't have to manually look for window names/properties to match, etc.

And I have found a theme which looks neat, but does not stand out enough to distract me. And the focused window is marked by changing the whole frame - not only the title bar, for example - so it can be found even when partially obscured.

Screen size is the problem

Posted Feb 13, 2012 13:49 UTC (Mon) by dbnichol (subscriber, #39622) [Link]

There are actually is one tiling features in gnome-shell that I'm aware of. You can drag a window to the right or left edge of the screen and it will tile the window to that half of the window. Look for tiling in the page below.

https://live.gnome.org/GnomeShell/CheatSheet

Screen size is the problem

Posted Feb 11, 2012 17:04 UTC (Sat) by Kit (guest, #55925) [Link]

Screen size wasn't an issue with netbooks several years ago, and those had quite low resolutions. By the end of the year, all the high end tablets will have resolutions rivaling even the highest end monster of a laptops (the Transformer Prime will have a 1920x1200 screen, the iPad 3 is rumored to be even higher than that).

When you have high enough DPI, on a still reasonably sized physical screen, limiting applications to full-screen only mode makes less and less sense. Sure, design a nice interface _IF_ you go full screen, but, don't remove functionality that it becoming more useful in more markets.

It's pretty easy to receive benefit of having multiple windows open and viewable simultaneously. You don't have to have them both completely uncovered, just having the important, content areas can be more than enough in most situations.

All the time I'll have both a text editor and a web browser sitting side by side. Generally it'll be because whatever resource in the browser is useful for whatever code I'm working on, but there's been many times where I just throw a video up in the browser to semi-pay attention to, to keep me a bit amused while I work. That's very simple, and very straight forward. I don't have to constantly mess with the positions of both windows (like some people are acting like you do), and it only took me about to mouse clicks to set it up (click and drag one window, click and drag the other- both were already decently sized enough to work without issue).

If you're OCD about the overlapping... well, I don't see that as a reason to force the entire desktop paradigm to change for everyone else. Add a full screen mode to GNOME if you really want (I thought they already added that? I've yet to try GNOME 3 on a machine that it didn't simply take me to the failsafe desktop, so I've not gotten to experience it first hand), but not at the expense of the far more common, and far more useful situation.

Hell, I don't keep my browser window maximized simply because of the few sites that don't constrain the width of the text (reading a normal sized font on a wide display is fairly tedious, the lines become long enough that it becomes easy to jump between them without realizing). I've seen many other people express the same complaint about wide screen displays. On the other hand, I've yet to see someone use a computer and have major issues with the whole idea of using multiple windows simultaneously. I've seen people that make me simply cringe with how bewilderedly they seemed to operate the machine, and yet when in situations where they would benefit from multiple windows visible even they've not had issues with deciding for them self to resize and position the windows for simultaneous viewing.

Screen size is the problem

Posted Feb 11, 2012 19:38 UTC (Sat) by hechacker1 (subscriber, #82466) [Link]

Agreed completely.

The one application per screen paradigm developers are pushing us towards just doesn't make sense if you are a power user in any sense of the phrase.

Windows 8 will bring us 1.5 applications per screen in certain circumstances.

I see this as a huge regress after having decent multi window management in Windows 7 and previously Gnome.

Windows 7 had it right. It was easy to drag the windows to one half, or to drag it to maximize. It made working with multiple windows easily.

Now we have touchscreen, small-screen focused window managers. Sure they do a great job there, but I don't really think they make sense to power users.

So, I guess I'll just have to use Mint's cinnamon, which is a sane compromise of Gnome 3 and window management. Or XFCE which looks a lot like gnome 2.

Screen size is the problem

Posted Feb 11, 2012 20:42 UTC (Sat) by elanthis (guest, #6227) [Link]

It's Important to note in Windows' case that the "traditional" Windows 7 desktop is not going away. The Metro overlay being the default UX can even be turned off. Microsoft is also betting hard on tablets and net books, but they're also smart enough to know that the desktop PC is not going anywhere. At "worst" they'll be replaced with tablet docks; businesses, creative users, and workstations and power users are not giving up keyboards and mce and big screens any time soon, because they literally -can't- do without those particilar features.

While GNOME is not forcing maximized windows (people need to RTFA), it is still failing to take advantage of the improvements from the proprietary OSes that focus on those power users. Bitching about how hard it is to tile windows, for instance, just means that the developers have no idea that certain well-known Vista/Win7 features exist (which I won't do without now).

Screen size is the problem

Posted Feb 11, 2012 22:23 UTC (Sat) by luya (subscriber, #50741) [Link]

On Gnome 3,an extension called gtile allow to set different windows by tile.
https://extensions.gnome.org/extension/28/gtile/

Screen size is the problem

Posted Feb 12, 2012 9:04 UTC (Sun) by hechacker1 (subscriber, #82466) [Link]

Well looking at the Windows 8 demos, indeed old Windows 7 behavior for window management is still there. As long as you use legacy applications.

But what worries me is that any Metro apps (using the new WinRT API) are forced into the metro style windows management. With metro you only get 1 full screen view, or one side bar view.

So this allows either full screen, or .33 off an app, and .66 of another showing on a screen. Otherwise, you have to flip through the virtual screens to see different applications. Metro is a very different windowing model that puts focus on a single applications, with a sidebar to perhaps interact with a second task.

I wonder just how many developers are going to be going with WinRT. Microsoft is giving incentives in doing so, considering you get into the App store with it, you also get to easily port it to ARM and future Windows 8 phones.

Otherwise you are locked into older x86 stuff.

Screen size is the problem

Posted Feb 11, 2012 20:30 UTC (Sat) by bjartur (guest, #67801) [Link]

Hell, I don't keep my browser window maximized simply because of the few sites that don't constrain the width of the text (reading a normal sized font on a wide display is fairly tedious, the lines become long enough that it becomes easy to jump between them without realizing)

Yeah, I've got a max-width on <p>s for this very reason. On my 17.2" widescreen laptop I've got a test build of Opera that can columnate, for reading longer text without scrolling.

I think maximizing makes sense most of the time anyway; the frequency of scrolling is inversely proportional to the amount of text you can fit on the screen at a time. It's just that Dreamwoven websites are not reflowable.

If you're OCD about the overlapping...

Guilty as charged - I'm using HashTWM on _Windows_ right now.

I don't have to constantly mess with the positions of both windows (like some people are acting like you do), and it only took me about to mouse clicks to set it up (click and drag one window, click and drag the other- both were already decently sized enough to work without issue).

Two Alt-presses, two clicks and two drags, and it sort-of works and only because the windows were small to begin with. I do think switching between Maximum (monocle) and tiled mode is quicker.

Screen size is the problem

Posted Feb 11, 2012 20:58 UTC (Sat) by dlang (✭ supporter ✭, #313) [Link]

if you do have content that is constrained on width (many websites, any paper-size related stuff) using the full screen width is even more wasteful

Styling is the problem

Posted Feb 12, 2012 0:28 UTC (Sun) by bjartur (guest, #67801) [Link]

When I don't columnate text, I zoom a lot and tend to disable CSS positioning. I resist flattening out tables, but table layouts don't scale so there is not much choice. As for PDF, I'm considering pdftotext|fmt.

Oh, if only branding and graphical design were not the most important issues to webmasters.

Screen size is the problem

Posted Feb 11, 2012 23:42 UTC (Sat) by bredelings (subscriber, #53082) [Link]

The authors added an update/edit at the end of the article responding to comments, and noting (among other things) that
  • "Second, although these applications will maximise by default, it will still be possible to unmaximise them."
  • "Everyone involved is well aware of the need to work well with large screens!"

Facebook Interface

Posted Feb 11, 2012 22:13 UTC (Sat) by zzxtty (guest, #45175) [Link]

GNOME developers have changed their target audience, its with in their right, but they really should acknowlege that one size does not fit all. It used to be very good for power users, I used to use it at work, 8 virtual desktops, 3-5 windows per desktop. It was very good for organising workloads. Their target audience is now the single application user, personally i think this is a mythical beast, someone who spends their entire time in facebook? Do people really do that?

I do have a non-power user usage, I have a laptop for banking and gnucash, I tried gnome3 on that, awful. I need to access two windows at once, bank website in web browser and gnucash. Cant be done, you end up having to scroll a cm here, resize a window there, move one out of the way, awful. I'd say this is a more normal user usage pattern and it still doesn't work.

Maybe it is just a facebook interface, thankfully we have other window managers.

Facebook Interface

Posted Feb 12, 2012 9:15 UTC (Sun) by drago01 (subscriber, #50715) [Link]

What the hell is a "power user" ? ... it is just a random buzz word without any meaning.

Does it simply mean "someone stuck to old technologies and unable / unwilling to adapt to changes" ?

And you can and will be able to use multiple windows in gnome3 just fine. I am working with multiple windows and workspaces and I would not want to go back to the old style win95 interface.

So no the target audience is not "single application user" also the the blog is about applications maximizing themselves by default. You can unmaximize them, open other windows, resize them, move them around, tile them etc etc.

Facebook Interface

Posted Feb 12, 2012 9:53 UTC (Sun) by dskoll (subscriber, #1630) [Link]

What the hell is a "power user" ?

Someone who is royally fed up with GNOME 3.

Facebook Interface

Posted Feb 12, 2012 10:13 UTC (Sun) by khim (guest, #9252) [Link]

Power is user is user who knows enough to tangle his (or her) system enough to make it unusable, but who don't know enough to make it usable again (without full reinstall).

It rarely makes sense to develop anything for power user.

Facebook Interface

Posted Feb 12, 2012 14:40 UTC (Sun) by nim-nim (subscriber, #34454) [Link]

> What the hell is a "power user" ?

I'd say anyone clueful enough to locate gnome3 authors and give them un-requested feedback

Facebook Interface

Posted Feb 12, 2012 16:54 UTC (Sun) by drago01 (subscriber, #50715) [Link]

Odd I though I have read about the mighty "power users" way before gnome3 existed.

Facebook Interface

Posted Feb 12, 2012 17:41 UTC (Sun) by drag (subscriber, #31333) [Link]

Definition of power user:

pow*er us*er (noun) \ˈpau̇(-ə)r\ \ˈyü-zər\

1. (obsolete): A user who is not only is knowledgeable enough to be aware of advanced features in computers, but has the desire and ability to take advantage of advanced features.

1. (modern): A user who loudly complains about the lack of a GUI button or slider to enable a niche feature, but is too lazy to be bothered to edit configuration, install plugins, or search Google to get it/enable it.

Power User definition

Posted Feb 12, 2012 18:13 UTC (Sun) by dskoll (subscriber, #1630) [Link]

No, your definitions are off.

1. (modern): A user who has, over the years, developed a workflow which is profoundly efficient for himself or herself only to have said workflow disrupted for no good reason by über-chic GUI designers who think they know better than anyone else what the One True Workflow should be.

Power User definition

Posted Feb 12, 2012 19:39 UTC (Sun) by bkor (guest, #27950) [Link]

glass half empty people :P

Power User definition

Posted Feb 12, 2012 19:43 UTC (Sun) by drag (subscriber, #31333) [Link]

No.

That is called a 'normal user'.

People don't like change. When people see change they don't like it. It doesn't matter if it's a improvement or not. They will bitch, always. ALWAYS.

_ALWAYS_

_EVERY_SINGLE_TIME_.

Doesn't matter if they are 'power users' or not.

Justification doesn't enter into it. It's a emotional response, not a rational one. Rationalization happens later when somebody has to attempt to defend their statements. This is why people just throw out terms like 'workflow' and 'power user' which have no real definition and thus are difficult to discuss rationally.

Then when a person points out that they can still have their sloppy focus (or that it's a planned feature), or set number of virtual desktops, or tiling interface, or scripted interface, or arrange things how they like it etc etc.. just like they had it before they just ignore it completely and will complain about the same exact things later on. Because it's not really the problem they are having with it.

A smart person behaving rationally is going to be smart enough to simply say:

"I don't care if it's better or not. They changed things and I don't like it."

Which is, actually, very easy to understand. I run into this constantly with all sorts of things far beyond computer desktops. Cars, color of houses, the way streets are laid out, the way their diet coke tastes... I is one of the very significant reasons why Linux has never caught up on the desktop.

To which the obvious rational response is:

"That is fine. Just use CentOS, Redhat, (or Debian Stable) for your Gnome 2 desktop. You can continue to avoid any change until 2021 or so, at which time Redhat will stop supplying security updates"

Everybody can have their cake and eat it, too. Problem solved, or at least deferred for 7-8 years. This is one of the reasons we have things like CentOS or Debian Stable.

Just go and look back at what people are saying: Personal attacks against developers. Calling users 'babies'. Claiming that they are making interfaces that are for 'idiots' and such things when it's obvious that the newer system is significantly more flexible and powerful... It is just people lashing out on a emotional level. Like they are hurt or something.

On a slightly side note:

What I find amazing is that more people don't realize that being incredibly anally retentive about the positioning and arranging of windows among their desktops is a trained behavior (on a almost unconscious level) that is the result a significant design flaw in their desktop/window manager.

Power User definition

Posted Feb 12, 2012 22:34 UTC (Sun) by bojan (subscriber, #14302) [Link]

> People don't like change.

And people like stupid change even less.

Power User definition

Posted Feb 13, 2012 2:35 UTC (Mon) by dskoll (subscriber, #1630) [Link]

drag, you are full of ****.

I don't hate change. I hate bad change.

I loved changing from MS-DOS to UNIX because that was an improvement. I loved changing from twm to KDE 1 (or 2? can't remember) because that was an improvement.

KDE 2 to KDE 3 was also an improvment. KDE3 to 4, not so much.

twm to GNOME 1 was an improvement. GNOME 1 to 2 was OK. And any of the above to XFCE4 was an improvement.

I have changed my "desktop environment" at least 4 times since I started using Linux and at least 8 times since I started with UNIX.

So don't dismiss my comments with "Ah, ignore the whinger because he hates change."

You are wrong. I like good change.

I hate bad change.

Power User definition

Posted Feb 13, 2012 7:41 UTC (Mon) by bronson (subscriber, #4806) [Link]

> People don't like change. When people see change they don't like it. It doesn't matter if it's a improvement or not. They will bitch, always. ALWAYS.
> _ALWAYS_
> _EVERY_SINGLE_TIME_.

I call your bluff. You sound rather sure of yourself, and go on for great length, but your premise is flawed.

My evidence: people gravitating to hybrid cars, iPads, smart phones, etc. Heck, trade magazines and blogs are basically long homages to change. And people love it.

So, I guess I'm saying: [Citation Needed].

Power User definition

Posted Feb 13, 2012 8:43 UTC (Mon) by neilbrown (subscriber, #359) [Link]

Good call.

I think the thing the people really don't like is the feeling that they aren't in control.

When things don't change, people can feel like they are in control because they don't won't it to change, and it doesn't.

When people make a deliberate decision to initiate or accept a change, then again they feel like they are in control (they may have been manipulated by clever advertising, but that isn't relevant - they *feel* like they are in control). So hybrid cars etc are easy to explain.

When someone decides to accept a software update and there are little changes, they might not like them but they can probably accept it because they chose to perform the update after all.

But when you accept an update and it causes big changes, then the effect seems out of proportion with the decision and you get a strong feeling that you are out of control. *That* is what people don't like. Ever.

So I recommend small continuous changes which are - where possible - selectable ("Would you like to try the new Frobnicator widget once(1), every time (Y) or not now (N)?[1/Y/N]").

I call this "Evolution, not revolution" - I bet no-one has heard of that idea before...

(sorry - no citation).

Power User definition

Posted Feb 13, 2012 11:23 UTC (Mon) by drago01 (subscriber, #50715) [Link]

> people gravitating to hybrid cars

Well they are pretty much the same as "normal" cars from a drivers pov. i.e No need to learn new stuff; no need to adapt to changes which seems to be one of the primary reasons people hate changes so much.

> iPads

Tables are not any "changes" they are new toys / tools. People don't replace anything with tablets they just buy it as an additional device. And no comparing sales to traditional computers does not prove anything. Some people are claiming "tablets are replacing PCs" ... but most PCs are "good enough" for what they are used for so people rather spend their money on new shiny stuff rather then buying a new PC.

> smart phones

That's the only point that is true here but the marketing and media hype seems to be stronger then peoples resistance to changes.

Power User definition

Posted Feb 13, 2012 12:36 UTC (Mon) by dskoll (subscriber, #1630) [Link]

Again, I say full of ****.

People don't replace anything with tablets they just buy it as an additional device

I know several people who have replaced their laptops with tablets. Could you provide citations to back up your points?

I don't know if you are a GNOME developer, but you certainly act like one: When people criticize the direction GNOME is taking, you insult them by saying they're just whiners who hate change instead of actually seeing if they have any valid points.

Power User definition

Posted Feb 13, 2012 12:55 UTC (Mon) by drago01 (subscriber, #50715) [Link]

> I know several people who have replaced their laptops with tablets. Could you provide citations to back up your points?

I know several people who buy them as additional toys. So? The argument "people are buying more tablets" => "people are replacing PCs with tablets" is just flawed. Where is your citation to back *your* point?

Anyway you want citations there you go: http://www.heise.de/ct/artikel/Touch-oder-Tasten-1430422.... (its German though). Basically the paragraph "Kein Notebook-Ersatz" ("No laptop replacement") is worth reading.

> I don't know if you are a GNOME developer

I am.

> you insult them by saying they're just whiners who hate change instead of actually seeing if they have any valid points.

So you consider "It is bad change" valid feedback to work with?

Power User definition

Posted Feb 13, 2012 14:41 UTC (Mon) by dskoll (subscriber, #1630) [Link]

So you consider "It is bad change" valid feedback to work with?

Yes, absolutely. If GNOME were not Free Software (ie, if you had to satisfy paying customers) you'd pay attention to such feedback. My job is creating software for paying customers and we certainly pay attention to our customers when they don't like a change we propose.

This is one of the few weaknesses of the Free Software development model, alas: Developers can happily ignore feedback without much risk. They get stuck in confirmation bias without any cold dose of reality to shake them out of it.

Power User definition

Posted Feb 13, 2012 16:39 UTC (Mon) by drago01 (subscriber, #50715) [Link]

> Yes, absolutely. If GNOME were not Free Software (ie, if you had to satisfy paying customers) you'd pay attention to such feedback.

No, when that happens (I have to do that but not with GNOME) I just do ask the customers *what* is bad. "It is bad change something" is just worthless noise IMO.

> This is one of the few weaknesses of the Free Software development model, alas: Developers can happily ignore feedback without much risk.

Not implementing everything every single user ask for is not ignoring feedback. That's not how it works. Fortunately many do know how to provide feedback which results into improvements being done (compare GNOME 3.0 to 3.2 and the now in development 3.4 version).

Power User definition

Posted Feb 13, 2012 17:04 UTC (Mon) by dskoll (subscriber, #1630) [Link]

No, when that happens (I have to do that but not with GNOME) I just do ask the customers *what* is bad. "It is bad change something" is just worthless noise IMO.

Well, of course it's "worthless noise" IYO. That's what I mean by confirmation bias.

Power User definition

Posted Feb 13, 2012 17:06 UTC (Mon) by drago01 (subscriber, #50715) [Link]

OK I give up you seem to be just trolling.

Power User definition

Posted Feb 13, 2012 17:31 UTC (Mon) by dskoll (subscriber, #1630) [Link]

I'm not trolling. GNOME developers made a bunch of changes to the way users interact with the UI. Many GNOME users said "Those changes are very bad."

I really don't see how feedback can be any clearer. You did X and users replied "Don't do X!"

Power User definition

Posted Feb 13, 2012 18:05 UTC (Mon) by mpr22 (subscriber, #60784) [Link]

I'm quite bad at parsing implicit communication, but it still seems to me that "this change is bad" is likely to imply "this change is for the worse".

Power User definition

Posted Feb 13, 2012 15:09 UTC (Mon) by cmm (guest, #81305) [Link]

Well, GUI-related arguments are necessarily subjective.

The fact is that I had a setup I was reasonably happy with (Gnome 2) for several years, and had it yanked from under me and replaced by something I neither like out of the box (in fact, couldn't even /figure out/ out of the box -- Gnome 3 gets a huge FAIL for discoverability) nor can configure into something resembling my old setup (unless I want to dig through an alphabetically-sorted pile of "extensions" with mostly unobvious names and unknown quality -- why thank you very much!). I like change when the change solves a problem. I had no problem with Gnome 2.

Examples? How about this waste of space on the bottom of my wide^Wshort screen? Why is it there? How do I get rid of it? How about not grabbing "Alt-Tab" (which I actually use in Emacs -- I prefer window-management shortcuts to use the otherwise-useless "Windows" modifier key)? Can I make the top thing (I understand it's not called "panel" anymore) be vertical? Or at least make it auto-hide? That's just off the top of my head, and do realize that in the presence of other options (which is not something you Gnome people can change, thank god) those problems are deal-breakers.

I'm not saying Gnome 3 is unworkable for me -- it may well be workable, but the fact that I'd have to spend significant effort to make it usable is irritating, and the additional fact that there are other options means that you a) lost a user and b) that lost user has moved on with whatever opinion of your work he had formed on the initial encounter.

I'd shut up if I thought I'm some sort of OCD-afflicted weirdo, but I know for a fact that I'm not: there are a lot of us, see. And we'll stay with those initial opinions. Deal.

You made your own bed, guys, now please stop whinging about your users "not liking change".

(Also, keep coming up with New Revolutionary Visions every 2 months and the CADT impression of you will only get stronger.)

Power User definition

Posted Feb 13, 2012 20:33 UTC (Mon) by bojan (subscriber, #14302) [Link]

> Can I make the top thing (I understand it's not called "panel" anymore) be vertical? Or at least make it auto-hide?

Don't ask for things like that - you'll be told to go and write an extension.

Power User definition

Posted Feb 13, 2012 15:36 UTC (Mon) by bronson (subscriber, #4806) [Link]

> Tablets are not any "changes" they are new toys / tools.

I must not understand... This sentence appears to be contradictory. New tools ARE change, are they not?

> People don't replace anything with tablets

Flight manuals: http://www.businessweek.com/news/2012-02-13/u-s-air-force...
Cash Registers: http://allthingsd.com/20110823/apples-ipad-already-replac...
Text Books: http://www.inquisitr.com/138739/ipad-replacing-many-textb...
Medical Devices: http://www.padgadget.com/2012/02/08/medical-industry-fast...
etc. etc. etc.

> That's the only point that is true here but the marketing and media hype seems to be stronger then peoples resistance to changes.

That's an amazingly cynical attitude. Marketing is NOT the reason the smart phone succeeded.

Power User definition

Posted Feb 13, 2012 19:43 UTC (Mon) by nix (subscriber, #2304) [Link]

I'm confused. I thought GNOME 3's entire design was predicated on the assumption that people would be using tablets for it all the time rather than PCs? (I can't see any sense to most of its design decisions except in light of that.)

Yet now we get 'people don't replace anything with tablets'. Hmm.

Power User definition

Posted Feb 13, 2012 23:03 UTC (Mon) by raven667 (subscriber, #5198) [Link]

I don't think they believe there will only be tablet but there will be a lot of them and it's a new open market that Apple and Microsoft haven't completely locked up yet. Keeping the UI flexible so that it can be used both on a tablet and a traditional desktop, without doubling the work, seems to make logical engineering sense. The two big vendors, Apple and Microsoft have independently come to the same conclusion as they are merging a bunch of their smartphone/tablet UI onto their desktop PC systems. I'm guessing that using a mouse on a UI that was designed with touch in mind is much easier than going the other way. Have you ever tried a traditional Windows tablet in the last 20 years? Awful, all sorts of mouse-dependent behavior that doesn't work with a touchscreen.

They may be right or they may be wrong but they came to their conclusions logically and reasonably AFAICT. We'll see in 5 years if these design decisions bear fruit.

Power User definition

Posted Feb 18, 2012 0:37 UTC (Sat) by cmccabe (subscriber, #60281) [Link]

The UI you need for a tablet is very different from the UI you need for a desktop. Apple recognizes that and hasn't tried to make the user interfaces the same between the two.

> I'm guessing that using a mouse on a UI that was designed with
> touch in mind is much easier than going the other way.

Bad guess. There's a lot of touch operations that have no mouse equivalent, like anything using multitouch.

The mouse is a precision instrument, whereas touch is very imprecise.

Users get annoyed when they have to move their mouse a long distance, whereas tapping a different part of the touch screen is easy.

> Have you ever tried a traditional Windows tablet in the last 20
> years? Awful, all sorts of mouse-dependent behavior that doesn't
> work with a touchscreen.

Wow-- it's almost like the two products need different user interfaces. Maybe someone should make a note of that. A lot of companies and open source organizations have sunk a huge amount of time and money into the foolish assumption that tablet UIs are relevant on the desktop, and vice versa.

Power User definition

Posted Feb 18, 2012 6:15 UTC (Sat) by raven667 (subscriber, #5198) [Link]

> The UI you need for a tablet is very different from the UI you need for a desktop. Apple recognizes that and hasn't tried to make the user interfaces the same between the two.

That's just nonsense. Most of the user-visible changes in Lion and the announce features for Mountain Lion are the transfer of iOS UI elements to the desktop (such as the app launcher or Notification Center) and even whole applications (like Notes, Messaging)

http://arstechnica.com/apple/news/2012/02/apple-unleashes...

Power User definition

Posted Feb 18, 2012 10:14 UTC (Sat) by cmccabe (subscriber, #60281) [Link]

That article says that they added "messages, notes, and reminders" to "Mac OS Mountain Lion." I don't see anything in there about making the user interface the same between the desktop and portable devices.

Power User definition

Posted Feb 18, 2012 18:11 UTC (Sat) by raven667 (subscriber, #5198) [Link]

If you didn't see that it was the same app and same ui then you must have not been looking, I don't know what else to say.

Power User definition

Posted Feb 13, 2012 13:34 UTC (Mon) by aliguori (subscriber, #30636) [Link]

>People don't like change. When people see change they don't like it. It doesn't matter if it's a improvement or not. They will bitch, always. ALWAYS.

That's because there's a cost associated with adopting to a change. If there isn't a perceived benefit to change that outweighs the cost, then people are going to be unhappy about it.

Think of it like maintaining API compatibility in libc. At some point, someone realized that read() could be improved by adding an offset to eliminate the need for two syscalls (lseek and read). If they would have just changed read(), that would have broke a lot of applications. So instead they introduced pread().

Yes, having 2 (or 3, or 4) different calls for read() is "crufty". We'd all rather work in code bases that never maintained these sorts of things.

But it's unrealistic to build a library and pretend like there's zero-cost in porting applications if the library breaks API compatibility.

Likewise, it's unrealistic to build a GUI environment and pretend like there's zero-cost in learning new workflows/options. Imagine how upset people would get if glibc started breaking ABI compatibility every couple years in core POSIX calls. That's effectively what Gnome is doing.

Power User definition

Posted Feb 13, 2012 21:30 UTC (Mon) by elanthis (guest, #6227) [Link]

That's a nice way of phrasing that. I hope you don't mind me stealing that metaphor. :)

Power User definition

Posted Feb 13, 2012 18:19 UTC (Mon) by sorpigal (subscriber, #36106) [Link]

> Justification doesn't enter into it. It's a emotional response, not a rational one

This is where I think you're wrong: Sometimes it's both. Dismissing all criticism as an irrational and emotional reaction to change is unhelpful. I understand that it gets frustrating to be constantly on the receiving end of bitching from people like me; try not to let it get to you.

> What I find amazing is that more people don't realize that being incredibly anally retentive about the positioning and arranging of windows among their desktops is a trained behavior (on a almost unconscious level) that is the result a significant design flaw in their desktop/window manager.

Curiously enough I have been running with mostly-maximized as my model for many years and I abuse virtual desktops heavily to make this work. I observe the following behaviors as a result:

* No matter what my "application" is for the current desktop I will always, sooner or later, also open an xterm that mostly sits behind the "main" window and alt-tab to it as needed.
* On some desktops my "application" is really a task involving multiple applications and in those cases they're never maximized.
* Max-width and especially max-height are usually all the window tweaking you need when not running maximized.
* Alt tab needs to be single-desktop, otherwise the list is crazily long.
* More desktops==better task isolation==better productivity.
* I almost never spend time tweaking window size or position.
* In the GIMP I'm always tweaking window size and position.

For some reason I also notice that settings-type windows seem hostile when they use the whole screen. I'm not sure why.

Power User definition

Posted Feb 13, 2012 19:03 UTC (Mon) by cmorgan (subscriber, #71980) [Link]

+1

People have several alternatives. They can use another version of the distribution, another distribution or another OS. They can get involved and help shape the future. Complaining is a LOT easier though.

Power User definition

Posted Feb 13, 2012 21:44 UTC (Mon) by elanthis (guest, #6227) [Link]

Most people have significantly better things to do than switch the OS every six months because some hobbyist UX designer decided to swap things around for no real reason.

I know the FOSS community has trouble understanding this, but most people do not use a computer just to use a computer; they are not computer geeks. They don't think that installing 20 different alternatives to try them all out is fun. They don't think that seeing new technology is fun. They don't think that shell scripting to work around a GUI flaw is fun. All of these things take time, and that time could be "better" spent doing something that most people do find fun. That may be some other computer-related task (socializing online, playing a game, etc.) or something completely unrelated to computers (going to the bar, hanging out with friends or family, hitting the beach, playing a sport, reading a book, etc.).

So this is the problem. Either (a) you admit that the Linux desktops like GNOME are by geeks for geeks and nobody else and quit eschewing the geeky power users by saying you're aiming for more casual users, or (b) you realize that non-geeks aren't like geeks and stop the endless churn and massive redesigns and instead focus on polishing the little details of the things that already worked just fine and your supposed 10's of millions of users are already comfortable and happy with.

As it is in GNOME land, both groups of users are frequently getting knocked over. I have absolutely zero doubt that the GNOME designers and developers like what they're making (or they wouldn't be making it), but making what you like and what other people like are often different tasks. As a game dev, for instance, I still think MUDs are the pinnacle of online interactions, but you don't see me trying to convince people that GemStone IV is going to take over World of Warcraft by just replacing the text commands with touchscreen gestures, do you. It would be silly to believe something like that.

Power User definition

Posted Feb 13, 2012 23:53 UTC (Mon) by bkor (guest, #27950) [Link]

You could say the same about most of the Microsoft Office versions. Loads of change between versions.

Power User definition

Posted Feb 14, 2012 0:17 UTC (Tue) by bronson (subscriber, #4806) [Link]

Really? Do they completely redo the entire UI with no ability to go back to the previous interface? Do they routinely jettison unpopular features? Do they break existing settings and workflow?

Business users would never put up with having to re-learn Word's interface all the time.

Yes, there are loads of changes. But they're done with a keen eye toward keeping existing users comfortable. I agree, it's a great example.

Power User definition

Posted Feb 14, 2012 0:19 UTC (Tue) by dlang (✭ supporter ✭, #313) [Link]

look at the ribbon menu approach that microsoft changed to, and look at the number of people who complain about it (at least they give you the option to switch back, at least for now)

Power User definition

Posted Feb 14, 2012 0:26 UTC (Tue) by bronson (subscriber, #4806) [Link]

> at least they give you the option to switch back, at least for now

This is the critical difference. As far as I can tell, Microsoft spends a lot of time ensuring existing users have an easy path forward.

Power User definition

Posted Feb 17, 2012 12:33 UTC (Fri) by job (subscriber, #670) [Link]

Change is only an annoyance. What's important is what comes after, whether the users cries out in terror or delight when they learn their new workflow.

You might want to skip on the condescending attitude. It prohibits your understanding of the issues involved.

Power User definition

Posted Feb 12, 2012 19:50 UTC (Sun) by SiB (subscriber, #4048) [Link]

A power user either does not use a desktop at all, but just a window
manager. Or he/she can adapt both the workflow and the desktop tools
when those evolve with time.

My fvwm2rc.m4 configuration file works almost unchanged for 12 years
now. But once in a while it gets me too. Recently, the Gnome people
decided it is a good idea not to offer the cwd as default location in
the gtk file chooser. Lots of programs now offer me an almost useless
file chooser. I did not figure out how to fix that.

When I start a printed circuit board editor in a shell from the
project directory, and I need to reload a netlist from this same
project directory, then I do not expect that I need to click all the
way into that project directory again, which is particularly annoying
if said project directory is nfs-mounted through a VPN over a ADSL
link.

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 3:57 UTC (Sun) by slashdot (guest, #22014) [Link]

I think the very vast majority of users simply maximizes all windows, and has done that since the beginning of time.

Especially when 800x600 or 1024x768 monitors were common, not having maximized windows was definitely not an option for general use.

Most people generally spend most of their time looking at a web browser/document viewer, a game, an office application, an IDE or an artist tool, and all these need as much screen space as possible, so nobody sane uses them non-minimized outside special cases.

Not really sure what's new there.

Also, the only reasonable choice for a default size for a new window is maximized, since any other choice would be arbitrary.

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 5:52 UTC (Sun) by dlang (✭ supporter ✭, #313) [Link]

office workers don't look at a office app window or e-mail window all day long, they have their IM app up in a corner somewhere at the very least.

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 11:44 UTC (Sun) by daniels (subscriber, #16193) [Link]

office workers don't look at a office app window or e-mail window all day long, they have their IM app up in a corner somewhere at the very least.

Their ... what?

Do you mean Facebook Chat in the web UI?

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 18:40 UTC (Sun) by dlang (✭ supporter ✭, #313) [Link]

whatever instant messaging client they use, in large companies there is usually an internal server so that the chat traffic within the company doesn't go out over the Internet.

in small companies it may be something like a facebook or twitter window or app.

in many large companies, running something like this all the time is close to mandatory.

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 19:59 UTC (Sun) by drag (subscriber, #31333) [Link]

Yes.

Chat is now a integrated feature into Microsoft's groupware. "Office Communicator". It's commonly used now.

Before that I noticed it was very common to use external chat services. For technical users it is typical to setup a IRC server or set up internal Jabber server. I've seen this is just about any place I ever worked in or at least sat down in. In places without internal IM I see people using text messages or using rouge services like 'facebook chat'.

(Business with IT goons that try to block any sort of ability to use IM is really doing everybody in the company a big disservice. It's very reasonable to keep things internal-only, of course.)

Even when not using full screens people will always focus on what they are doing. Especially if they are reading text they will mentally blot out everything else happening on the screen. So it's a always a challenge to be able to design a desktop so that people don't miss high priority notifications. Especially since what is a 'priority message' is so flexible and changes depending on tasks and particular users.

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 7:43 UTC (Sun) by Kit (guest, #55925) [Link]

> I think the very vast majority of users simply maximizes all windows,
> and has done that since the beginning of time.

Not users that I've looked over the shoulders of. The vast majority of them have windows maybe filling up to 90% of the screen, but rarely do they maximize stuff (the only exception I've seen as a common case are people actively using a projector as part of presentations).

I'd really wager that most things _don't_ benefit from being absolutely full screen, especially with how common widescreen monitors are, especially since you lose the easy ability to glance at any other app (whether it's an IM app like dlang said, or some resource they're also using).

> Also, the only reasonable choice for a default size for a new window
> is maximized, since any other choice would be arbitrary.
For an application never-before opened it would be one thing (it would still be an arbitrary, and in a way, rude, choice, as it presumes that THIS app is the most important thing you're doing). For apps you've previously opened? They should default to the last size they were, THAT is truly the only non-arbitrary option.

If you start shoving maximized-by-default mode on people, though, new users quite likely _won't_realize_ that you can tile (even if only manually) windows, unless someone shows them (or they do it on accident). IMO, Desktop Environments shouldn't do anything to _exacerbate_ the dumbing down of the "average" computer user.

Quite frankly, I don't really think the "average" computer user is getting dumber. The vast majority of people out there did NOT grow up with computers, and the ones that did, didn't grow up with ones that had an interface all that similar to the current paradigms.

Hopefully one day I'll have a spare system powerful enough that GNOME 3 won't simply fail on it, so I can actually try it first hand to see if the various horrible descriptions of it are accurate... Personally, I've never been a fan of GNOME, and found the 'gnomeification' of various GTK apps I used to always be a major regression (I like my option windows, damnit!), so I don't have much confidence GNOME 3 was a reversal of anything.

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 11:27 UTC (Sun) by mpr22 (subscriber, #60784) [Link]

Given that computer presentation of text generally uses horizontal lines, and that extremely long horizontal lines are awkward to read, maximizing doesn't really make that much sense on a 1920x1080 monitor unless you're using the kind of app that has lots of auxiliary controls to chew up horizontal screen space with. (1080x1920 would be another matter. There, it probably would make sense.)

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 17:31 UTC (Sun) by drag (subscriber, #31333) [Link]

What does matter is visual distraction.

A 80 character by itself floating in the middle of a 1080x1920 blank background is vastly easier and more pleasurable to read then:
A) 240 character wide column of text
or
B) 3 80 wide columns of text stuffed next to one another.
or
C) 80 character wide column of text surrounded by pulsating buttons and videos flipping around.

Depending on what you are trying to do it is more advantagous to stuff the maximum amount of information into the smallest amount of space... this is something that tiling window managers do a very good job of.

But it is certainly not desirable every time or even most of the time.

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 18:09 UTC (Sun) by mpr22 (subscriber, #60784) [Link]

I reject both option (@) and options (A) through (C), preferring (D) a desktop whose applications don't look like they belong in downtown Las Vegas, so that having one of them visible around the edges of another is not significantly distracting. Any application that can't do anything useful with the full width of a 1920-wide display has no business being maximized on one except at the explicit choice of its user.

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 19:12 UTC (Sun) by drag (subscriber, #31333) [Link]

> Any application that can't do anything useful with the full width of a 1920-wide display has no business being maximized on one except at the explicit choice of its user.

Which is _EXACTLY_ what the Gnome developer said should happen in his blog.

Unless the application developers requests the application to open maximized by default (because that is what he is intended by the application), then it won't.

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 12:35 UTC (Sun) by ballombe (subscriber, #9523) [Link]

> I think the very vast majority of users simply maximizes all windows, and has done that since the beginning of time.

I really doubt that. Most basic desktop users I know are completely lost if they cannot easily access the desktop (background), because they put everything on the desktop, so they always keep a part of it uncovered for drag-and-drop operation (they are unable to use a file selector).

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 18:32 UTC (Mon) by sorpigal (subscriber, #36106) [Link]

This is one of the raging failures of the modern desktop which ought to be corrected: Desktops and desktop icons. It's not a desktop, the metaphor is decreasingly useful on a conceptual level and users should be taught some kind of storage principle that doesn't involve "on the screen in front of me."

Even if you replace the "desktop" with a big "FILES" button that's always visible in a corner of the screen and that shows the former "desktop" folder, and do nothing else, that would be an improvement. Users constantly "lose" things behind windows and mostly do not understand stacking.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 2:56 UTC (Mon) by jmalcolm (subscriber, #8876) [Link]

Reading this thread, I have been wanting to respond that I hardly ever maximize windows. Then I realized that I am reading this article in a maximized browser window. I am reading it on my laptop which is somewhat low resolution (1280 horizontal pixels and about 10 inches).

On my main desktop, I would never maximize the browser because it is 1920 pixels across and a 20" monitor. Typically, my browser is about the same size and dimensions as a A4/letter sheet of paper. This about half the width of the screen.

Thinking about it a little more, if a given window has to be 75% or more of the width of my screen to be usable then I might just maximize it. All this proves is the "maximized makes sense for small screens" argument (for me). I am a little surprised though about how big "small screens" turns out to be (for me at least).

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 18:24 UTC (Mon) by sorpigal (subscriber, #36106) [Link]

> I think the very vast majority of users simply maximizes all windows, and has done that since the beginning of time.

In my experience most "real" users don't maximize windows at all. Instead they carefully drag the expando until the window is big enough, which is sometimes approximately the same size as maximized, and then carefully position it in the center or top-left of the screen.

"Why don't you click the maximuze button instead?"

"The what?"

This happens all the time.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 21:50 UTC (Mon) by elanthis (guest, #6227) [Link]

> Also, the only reasonable choice for a default size for a new window is maximized, since any other choice would be arbitrary.

I strongly disagree. Many apps have very natural sizes. GTK+ in particular was one of the first toolkits to have good natural auto-sizing in widget hierarchies (I think, could be wrong).

Some windows have a bit resizable content area, and the size of that area is indeed "arbitrary." Things like the main windows of a word processor, web browser, image editor, etc. Other windows have extremely specific natural sizes based on their contents; there is nothing arbitrary at all about the window size chosen by the application when mapping the windows.

Even the apps with resizable content areas often have a decent "good guess" as to a natural default size that may not map to fullscreen. Web browsers are best off targeting a specific view area "resolution" for instance based on how Web sites are actually designed. e.g., they are generally made to fit something between 800 to 1024 horizontal virtual pixels, with any extra space "wasted" by whitespace; other sites just assume the resolution will not be much higher and so will gladly fill the extra horizontal space by making the main content very hard to read by stretching out text lines far beyond comfortable lengths.

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 15:41 UTC (Sun) by mordae (subscriber, #54701) [Link]

OK, how do you tell these people that they are morons and should seriously stop modelling user interface in a way that makes it 10% easier for idiots and 100% harder for productive people to use?

Why do they keep breaking terminals?

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 18:15 UTC (Sun) by dskoll (subscriber, #1630) [Link]

OK, how do you tell these people that they are morons

Don't bother. Just don't use GNOME. XFCE4 is a great alternative.

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 19:34 UTC (Sun) by bkor (guest, #27950) [Link]

It doesn't seem you are not bothered judging by all your replies here. It is quite enlightening to read the replies on the blog post itself. Very positive and not needlessly negative.

Some work on maximize is being done. On the blog post it is appreciated and politely questioned if large screens aren't forgotten.

I suggest to actually read the blog post and the answers given in the comments. That, or really do not care.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 8:02 UTC (Mon) by bronson (subscriber, #4806) [Link]

It's pretty clear that "Don't bother" meant "Don't bother running Gnome 3", not "I'm not bothered by Gnome 3."

> I suggest to actually read the blog post and the answers given in the comments.

Do you suppose the "David F. Skoll" commenter there might be dskoll? Just a guess.

I read through a lot of the comments... People expressed some concerns about big desktops and multiple monitors, an area where Gnome 3 still suffers. The concerns were hand-waved away with "we're working on it" (paraphrased). Not very encouraging.

We'll see! Should be fun to watch.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 10:04 UTC (Mon) by bkor (guest, #27950) [Link]

I find it awesomely funny that you summarize "we're working on it" as hand-waving. To me it means it is acknowledged to be a problem and they want to fix it... but you could also of course just complain on LWN. :P

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 15:55 UTC (Mon) by bronson (subscriber, #4806) [Link]

It is hand-waving until you actually hint about how you're going to solve it. I didn't intend the term in the pejorative sense, just the "unsubstantiated claim" sense.

(Unless I missed a reply where Allan Day explained how these features, in conflict with large or irregularly-sized desktops, can be reconciled...? I admit, there were a LOT of comments...)

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 23:51 UTC (Mon) by bkor (guest, #27950) [Link]

It is not hand-waving. He said at FOSDEM that it is a problem, they want to focus first on one thing, then work on the next thing. He explained it as iterative design.
What he explained that design is not perfect from the start. It is tested and changed until it works better. So when he says something is not perfect, and work is done by iteration, it conflicts with expecting that everything is explained up front. That is not how they do things.

If he said "will get to it", but didn't mean it (or never making time to get to it). Yes, that is hand-waving. But that is not the case here.

I hope you at least attend a conference so you can hear this in person.

Day: A New Approach to GNOME Application Design

Posted Feb 14, 2012 0:50 UTC (Tue) by bronson (subscriber, #4806) [Link]

That sounds great to me as long as they get far enough before releasing it. Gnome 3.2 still doesn't work on my multiple-monitor setup (bugs filed by others) so I would not be surprised if this too ships a little early... Somewhat uncommon desktop setups (like multiple monitors at different resolutions) could be left in the cold again.

We've both made our points, now we'll see! At least this time, should it ship a little early, I'll be a little more prepared for it. :)

I can barely keep up with my EE conference schedule but I'll keep an eye out. If Gnome 3 comes to the Bay Area, I'll make time.

Day: A New Approach to GNOME Application Design

Posted Feb 14, 2012 9:54 UTC (Tue) by drago01 (subscriber, #50715) [Link]

> Gnome 3.2 still doesn't work on my multiple-monitor setup (bugs filed by others) so I would not be surprised if this too ships a little early... Somewhat uncommon desktop setups (like multiple monitors at different resolutions) could be left in the cold again.

This is supposed to work ... where is said bug report?

Day: A New Approach to GNOME Application Design

Posted Feb 17, 2012 3:31 UTC (Fri) by bronson (subscriber, #4806) [Link]

I think this is it...

https://bugzilla.redhat.com/show_bug.cgi?id=748843

I'm running Radeon but not Intel so this might not be related to the problem I'm seeing. That said, I'm not too interested in this bug... I couldn't get Fedora to recognize my wireless adapter so I'm running XFCE on Mint. Worked perfectly out of the box, super stable.

I'll try Fedora again in a release or two.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 13:52 UTC (Mon) by ssam (subscriber, #46587) [Link]

the trouble is that XFCE is not as good as GNOME2 in some respects (it is really great in most).

For example Gnome seems to be the only desktop that when I hot plug a monitor, it remembers where I like it to be (above my laptop screen). Other DEs either, hotplug it but forget where I put it last time or dont hot plug at all so i need to press a button. So will leave windows in inaccessible places after removing the screen. Worst is KDE which pops up a dialog telling you that you plugged in a monitor (duh), and has a but to open a massive control panel so i can configure the monitor back to where I want it.

However GNOME3 enforces that the top (external) monitor is the main monitor, other wise i cant drag windows above the menu bar. And then only does virtual desktops on one monitor (yes there is a gconf option, but it is very buggy).

Luckily GNOME2 did not magically stop working. the GNOME devs attempts to kill it with name clashes are solved by MATE desktop.

Don't tell me about extensions. The extension website managed to make the whole of GNOME3 crash.

GNOME3 is like BTRFS, exciting and full of new features. But no sane distro would make it default, remove ext4 and then say if you dont like it just use FAT32.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 15:29 UTC (Mon) by bkor (guest, #27950) [Link]

Did you file a bugreport for the GNOME 3 crash with extensions? Somewhat surprised as I didn't notice seeing any bugreports about it on bugzilla.gnome.org. Usually at least someone would file a bug.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 22:04 UTC (Mon) by elanthis (guest, #6227) [Link]

Something we do during the alpha/beta testing of games is to put a super easy "Report Bug" button right in the UI.

If GNOME is still this heavily under UX design, it probably needs something like that.

Filing a bug in Bugzilla is a giant, mega pain in the ass. You first have to Google how to file the bug, because every project has a different way. Then you end up on a site with a truly horrific user-unfriendly uber-technical UI thanks to Bugzilla being designed by Perl developers. Then you have to make Yet Another ****ing Web Account and go through that whole process. Then you have to wait for the account to be verified and an access code emailed to you. Then you can finally log in. Then you can wade through 10,000 component options that you are probably not expert enough to select from. Then you are asked to fill in an arbitrary text field with a lot of TL;DR text about formatting with complicated-sounding tests, because the user is also expected to be a trained QA agent. Then there's an overwhelming chance that you're just going to be asked to submit a crash dump or run some technical test or otherwise do a bunch more work because the owner of the bug is optimizing for his time instead of the user's and made zero attempt to repo or investigate the issue. Finally, the bug may be fixed, and the user will eventually get that fix the next time they reinstall their OS ~6 months later because the fixes only show up in the next major version of the desktop and the Linux distros still don't do rolling releases despite being on a ridiculously fast (dev-friendly, user-hostile) release cycle.

Of course you end up with a ton of crashes and broken/missing features that never get filed as bugs. Your process for filing them is hell.

Make a bugs.gnome.org page. It has a few fields for name/email (optional) and a couple guided text fields to describe the bug. Add a submit button. That's it. If the user want to create an account, it's _optional_. If the user doesn't want to give contact info, then just throw away bug reports that need more information (after at least making a passable attempt at investigating/repo'ign the issue, if possible). Make it stupid easy for people to give you feedback. And then just put up with the garbage feedback you're going to get by spending the few minutes to filter them out, because that's the QA person's job and not the users' job.

Then put a nice button on the GNOME3 bar for reporting bugs that just opens up that page in a chrome-free browser window. Now you've got a feedback mechanism that actually works.

If you're worried about spam in the form, btw, Akismet works wonders.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 23:46 UTC (Mon) by bkor (guest, #27950) [Link]

I've been the GNOME bugmaster for various years. If that is your suggestion for Bugzilla, then I am happy to say that we have better ideas. Your idea is actually still far more complicated than it needs to be for a crasher bug.
So in short: Totally acknowledge Bugzilla is far from ideal. But it is what we have at the moment. So if we have that *at the moment*, then if you want your crasher to be resolved, then I rather focus on having the crasher at least be known to the developers.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 21:49 UTC (Mon) by dgm (subscriber, #49227) [Link]

Hear, hear. Xfce is great and working better with each iteration. Also, GNOME developers have shown they are not interested in that kind of feedback, no matter how sincere. Even worse, some become picky and "fight back", so its better to leave them alone.

Day: A New Approach to GNOME Application Design

Posted Feb 12, 2012 23:55 UTC (Sun) by drago01 (subscriber, #50715) [Link]

My terminals work fine.

How does having apps that maximize themselves by default break your terminal?

Instead of throwing random insults and flames you could better spend your time giving constructive feedback (as long as it is anything more useful then a reworded "I hate change").

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 0:02 UTC (Mon) by dlang (✭ supporter ✭, #313) [Link]

if the GNOME standard says that a good app always maximizes itself, because doing otherwise wastes screen space on window decorations and borders, then it's not unreasonable to believe that future versions of the terminal will follow this standard, and people are arguing that doing so is a bad thing.

and the argument that people only want to have one thing on a screen is false in so many ways it's not funny, even for 'normal people', let alone for power users.

small screens may be the exception, but as Android is now starting to discover, even a 7" or 10" screen can be "large" if the app was designed for a 3" screen and simply blowing up the app to fill the entire screen doesn't work very well. right now the android standards have morphed to talk about displaying multiple panels of the app on larger screens, but I don't think it's going to be long before there's a version of android that supports putting apps in windows and displaying more than one app at a time on larger screens (probably fixed sized windows initially, but that will change too)

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 0:27 UTC (Mon) by drago01 (subscriber, #50715) [Link]

> If the GNOME standard says that a good app always maximizes itself, because doing otherwise wastes screen space on window decorations and borders, then it's not unreasonable to believe that future versions of the terminal will follow this standard, and people are arguing that doing so is a bad thing.

Let me quote from the blog
"smaller utility applications will not be affected by this" and
"First, not all applications will use this behaviour – only those that have been designed to do so."

Terminals will stay the way they are and maximizing them by default would not equal to "breaking them" (I do sometimes use maximized terminals, so its not like they are useless in this state).

> and the argument that people only want to have one thing on a screen is false in so many ways it's not funny, even for 'normal people', let alone for power users.

That's not what the blog says nor does anyone plan to remove the ability to have multiple windows. And throwing in meaningless buzzwords like "power users" does not add any value so just stop that. See discussion above I was quite serious while asking "what the hell is a power user" because it does not make sense at all. You are just a user. You might be using you computer for writing documents, developing software, browsing the web, editing photos, cutting movies, ... or doing multiple of this tasks. What of them makes you a "power user"? nothing really.

Day: A New Approach to GNOME Application Design

Posted Feb 13, 2012 13:30 UTC (Mon) by daniels (subscriber, #16193) [Link]

What of them makes you a "power user"? nothing really.

AFAICT, the primary difference is the willingness to call developers idiots in LWN comments.

Why the outcry?

Posted Feb 14, 2012 1:40 UTC (Tue) by mstefani (subscriber, #31644) [Link]

The comments seriously drifted off...

The guidelines are for Gnome3 applications that would benefit from running full screen and not for the little tools around the desktop. This isn't something like the breakage of the GTK file selector or the swapping of the OK/Cancel buttons position that affected people outside the Gnome world.

So why is there such an outcry? Especially for the people that moved off of Gnome Shell the impact should be be negligible. Due to F15 I had to move myself from Gnome to XFCE and about the only Gnome apps I'm still using are gnote and evince. gnote is a small utility so it's unlikely that it will be affected. Evince on the other hand has a less then optimal default of "Fit page width" so I end up changing that as well as the window size and hit "Save current settings as default" which takes care of the problem. And until recently I used gthumb but replaced it with gwenview.

Now am I missing out on any must have / killer Gnome application?
While I did prefer Gnome2 to KDE as the less cluttered and more polished desktop environment I do think that KDE had the better applications.

Copyright © 2012, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds