(My idea from a while ago, taken from
https://wiki.ubuntu.com/IdeaPool to get some feedback)
I think there is far too much redundancy in the desktop at the moment. For example, icons on the desktop, in the menus, in a file manager, on a panel, a window on the desktop, tabs within windows, taskbar entries, notification area icons and panel applets can all represent the same thing. This is a complete waste, and it makes things confusing as to where certain options should be placed. At the moment it seems GNOME is following a put-it-anywhere-it-might-be-wanted philosophy (which results in such awful UI as the Microsoft Windows Control Centre, traversible only by remembering where developers ended up sticking the options you're after rather than by using common sense to find them) rather than a simple, logically layed out system.
For instance many applications can be closed via a widget on their window, an enty in a menu inside the application, right clicking on a taskbar entry, right clicking in a notification area icon, etc. but there is more relevant functionality which isn't as dispersed, for example setting an IM status or changing track in a music player, which can be accessed by notification area icons or panel applets but not by taskbar entry or desktop or application menu icon.
It would make a lot of sense to bring some of this stuff together, for instance the "desktop widget" idea for little snippets of information is a good one, but why not have that information in the application icons*? If a taskbar entry is flashing, does a notification area icon also need to flash? What is the difference between a panel applet and notification area icon (aside from the fact that panel applets can be moved around)? What is the difference between the taskbar showing the windows on a desktop and a bar of tabs showing the pages in a browser (most noticable for fullscreen windows)? What is the difference between a launcher icon and a taskbar entry (nothing, according to the Avant Window Navigator, and some apps even write their own hacks to do this, ie. switching to an open window when an icon is launched). What is the difference between a viewing application showing a document and the icon for that document? Should icons be static, so users can rely on them always being exactly the same, or should they include dynamic content, such as Evolution's icon showing the number of unread messages, or Pidgin's showing current status and new message notifications? Should icons exist at all now that we have the computing power and software to display large numbers of whole documents at any size? Should applications exist as standalone things or should their functionality be built into the file managers?
I think there are compelling reasons to stop bolting on yet more layers to the desktop system and take a step back to think of the fundamental goals of the system. The thinking behind GUIs has never really replaced anything, instead new methods of interaction are added *as well as* existing ways, which has resulted, after 20 years, in desktops where some tasks can be accomplished using a varity of methods, some only using a subset and others using a different subset, with no clear distinction why and with boundaries blurring and overlapping, which I feel makes a system which is confusing to those who aren't familiar with it.
This is a combination of a 'make all options available everywhere' idea and a 'make a clear way to carry out an action' idea, but also neither. As it stands there is no clear way to carry out some common actions (by clear I mean a possible solution is the first route tried, rather than the third), and options to do so are interspersed unevenly amongst various different metaphors and paradigms, requiring users to be familiar with all of them along with knowing which ones happen to have the option they want enabled.
* note that this idea predates KDE4's Plasma, which seems to be at least experimenting in this area :P
Tags:
(none)
|
Warbo wrote on the 20 Jun 08 at 03:31
|
noodlesgc, vexorian: Indeed it would cause outcry and confusion to change the current way of working, but it may be possible to unite the different forms of desktop interaction via incremental changes, each adding their own tweaks which users can pick up if they want or carry on as before.
The goal would be to advance our software so that in, say, 5 years time anyone from today would be familiar with the advanced desktops being used (maybe with a slight amount of "unlearning"), whilst any future users trying out our Ubuntu/Gnome/KDE of today would quickly get annoyed that they have to do things in roundabout ways, that their screen is cluttered with redundant icons and notifications, etc. and wonder how they ever put up with it. That would be a sign that we've progressed :)
To be clear, I'm not advocating we throw out the desktop and start again or anything like that. I'm saying that those making the infrastructure (KDE, Gnome, QT, GTK, etc.) should take a step back and, putting aside what they "know" about how to use the system they're making, look at how it "should" or "could" be used. This would be done through questions like "I can change my music by right clicking on the notification area icon. I try doing the same with the taskbar entry, since that's bigger and easier to click. Why doesn't it work?" or "There's a tiny picture in the corner which I can't drag around telling me how many unread news articles I've got. Why doesn't the bigger picture on the desktop, which I can drag around, tell me that?". Thought experiments like these can be added to a Wiki, their merit discussed along with reasons why they don't work at the moment so developers reading the Wiki can have fun challenges to attempt.
You can see a somewhat similar thing here (as in the process, not neccessarily the ideas, although I do like the "copying disc to a disc image" mockup) http://chabada.sk/better-desktop/
Sidney: Thanks, I suppose the idea is for opening up a debate around the issues rather than seeing everything from the current point of view (for instance saying "That would be cool to see!" rather than "You can't because taskbars are tied into the window manager."), but I did get a little carried away with examples :P
I'm by no means a usability expert, and wouldn't want to undermine or tread on the toes of those who are, but I do like thinking about UI and I'll do some Glade/GIMP mockups of ideas when I get some free time, save them to my blog and post links here :)
|
