I’ve been thinking about window management systems recently. Let’s consider Apple’s hodgepodge solution for OS X — It started with the Dock and the Application Switcher (technically a part of Dock.app, but that’s irrelevant). Not too complicated, and it worked well — but had problems when you had more than a certain quantity of open windows (dependent on the physical size of your screen).
So, in Panther, we saw the release of Exposé — heralded as the fix for this problem, Exposé allowed a user to open many more windows than they were previously capable of before the onset of chaos and confusion. Users being what they are, though, this wasn’t enough — with the increase of the soft limit, came an increase in the need for a way to manage even more open windows. So, in Leopard we saw the introduction of yet another window management system component — Spaces.
However, Spaces saw quite a bit of controversy when Leopard was released, as people found it difficult to use — things simply didn’t map well to the way the human mind things. I myself hardly ever find myself using Spaces, except as an alternative to minimizing a ton of windows — if I have a single quick task to get done, but I need to have multiple windows open to get it done, I’ll often switch to a second Space to complete it before switching back to my primary Space. Other than that, I never use the tool — it’s simply too cumbersome.
Now, on a completely unrelated note, I’ve recently obtained Will Wright’s latest game, Spore (by EA). I have to say, it’s pretty fun. It didn’t last long for me, but it’s not the kind of game you play repeatedly for a long time anyway.
Point being, Spore is split up into 5 ‘stages’ of play, each of which is sort of like a separate mini-game. Howe you choose to play each stage affects the other stages, and your creature retains properties you give it from one stage into the next… but other than that, the stages are rather disparate.
The first stage is the one relevant to our topic — the ‘cell’ stage. In this stage, the player’s creature is a simple ‘cell’ organism in a (relatively) huge pool of ’primordial soup’. In this stage, the playing area is essentially two-dimensional, just like our desktops currently are (if you’ve played Spore before, you may be beginning to see where I’m going with this). However, at various stages of microevolution, one’s cell would visually float upwards a ‘level’, so that cells that had recently been much larger than you were now about the same size as you, and cells that had been smaller would now be floating ‘below’ you, inaccessible. Occasionally throughout this mini-game, the player would see absolutely vast cells floating far ‘above’ them — cells so large that they were not yet intractable-with, and as such posed no threat to the player’s cell… yet. After another level or two of advancement, the player’s cell would again encounter this species, but this time it’d be the same size or marginally larger than the player, and the player would be able to interact with it. You can see something quite strikingly similar to this concept in the flash web-game flOw.
The pool in and of itself, as a playing area, was essentially infinite — you could ‘swim’ in any direction for as long as you liked, and still encounter activity and life. I think this metaphor could map quite well to a window management system. An infinitely horizontally- and vertically-scrollable workspace that extends in all directions, with an additional restrained third-dimensional component. You could float ‘up’ and ‘down’ through the workspace, with windows ‘above’ you becoming larger and more transparent as you dug ‘down’… and windows ‘below’ you becoming smaller and more transparent as you floated ‘up’.
This could be a very efficient method for organizing active projects — if you had several similar sub projects, they could have groupings of window arranged on or near the same ‘level’, and a window that’s important to all the sub projects could be ‘above’ them all. When you need to reference that window, you float ‘up’ enough to see it, then back ‘down’ to your active projects. All the while, you’ve got a completely unrelated project with completely different applications’ windows in a different location in the active workspace, far off to the left.
This system would obviously be very open and customizable — you use it however you want to. You could simulate Apple’s Spaces by having a three-by-three array of groups of windows — float ‘up’ enough, and you’ve got a view identical to that of Spaces, and then float back ‘down’ to the group you want to work with. The system could provide a tool similar to Exposé that simultaneously moves all windows out from all other windows (temporarily), and floats you ‘up’ enough to see all the windows you can currently see, but in their new positions — and then float back ‘down’ to your original view (re- positioning the windows as they originally were) after selecting one to be ‘moved to the front’ (i.e. floated up to your current level).
The ability to ‘bookmark’ a certain view would be essential, so that you could quickly jump between tasks without too much time spent tediously floating around (with the manual spatial maneuvering obviously still available for when you need to re-arrange things or relate things). Global hotkeys for bookmarks would expedite things even more. Jumping to a bookmark could include a breif and requisite animation wherein the user’s view was floated ‘up’ just enough to see both the original location and the destination, and then back ‘down’ to the destination — giving the user the illusion of an actual physical jump between the two virtual locations.
The interface widgets and images involved in such a windowing system would have to be fully scalable, so that they worked just as well (visually as well as functionally) at half or double size as they do at their base size. This way the layer directly above and all the layers below the user’s current layer could continue to be fully interactive and usable, giving them more freedom to navigate as they choose. Apple has already introduced similar technology in their patent filings for resolution independent interfaces, but there’s already been plenty of buzzing about that topic already.
There’s been many attempts made at various sorts of 3d interfaces, and none of them have really worked out too well. I think this is simply because we’re trying to map two-dimensional concepts to a three-dimensional interface on two-dimensional hardware — that can’t possibly turn out well. What I saw in Spore that inspired me, was how you were given a restricted three- dimensional working area — you were actually working in a two-dimensional space, but that space mapped to a single layer in a super-workspace in three dimensions.
Although I don’t really have the programming-fu to do anything about it (the extent of my desktop-UI programming ability is some mucking about with Ncurses in Ruby), I think this idea has potential.