Nooface
search Nooface:
  
In Search of the Post-PC Interface
 
Nooface
- Home
- About
- FAQ
- Discussions
- Journals
- Messages
- Topics
- Contact
- Preferences
- Older Stuff
- Past Polls
- Submit Story

Resources
- 3D User Interfaces
- Archives

 
An Introduction To The Track Window Manager User Interface
posted by Editor on Friday January 07, @03:47PM
User Interface Innovation Will La Chapelle writes "It's my hope that more open OS Developers might consider incorporating Track (or Tiling) Window Manager functionality at least as an alternative or option to the conventional and frequently copied cascading windows, desktop, and task bar/manager paradigm. A lot if not most the of current crop operation system user interface design (mainstream and alternative) is based on making commercially marketable, attractive and 'familiar' interfaces that are increasingly removed from any real studied analysis of how computers are actually used or real interface utility -- often even ignoring substantial current research and testing that has been done on user interface design. Interfaces incorporating cascading windows, desktop, task 'bar or manager' are very manipulation intensive i.e. the user frequently has to perform an inordinate amount of redundant and repetitive interface manipulation work just to accomplish the simplest task. DCW GUIs (Desktop Cascading Windows) are also wasteful of screen real-estate and inefficient at clearly comporting useful information without substantial manipulation and re-manipulation". Read on for a discussion of the TWM alternative.

What Is A TWM

A TWM -- Track (or Tiling) Window Manager is an alternative to the 'desktop' and 'window' paradigm to the traditional cascading window desktop design metaphor. Fundamental concepts behind TWM design include:

· screen real estate is precious
· computers may multi-task well but people may not
· people may multi-task marginally well but cascading windows make that impossible
· complete windows management/customization using only the keyboard
· avoiding redundancies through inheritance
· reduction of manual window positioning and management

The TWM interface paradigm in essence 'tiles' all windows so that 100% of all screen real estate is in 100% use 100% of the time comporting 'visible' information. TWMs typically incorporate default keyboard and mouse shortcuts that allow for automatically resizing windows to preset 'tracks' or sizes according to user preferences and include zooming features to make one window the largest or full screen application. Interface 'widgets' may include a 'tab bar', and/or integrated and unified tool and menu bars that have 'inheritance' features for the dominant application(s).

How It Works

In a TWM the 'desktop' is the work space and the application interface concurrently -- all open applications may have a window visible as a 'tile' or 'track'. Windows too small to comport normal application content may be reduced to icons, tabs, or some specialized 'ticker' representation of the application that offers relevant information.

Sophisticated TWM interfaces like 'Gadgets' for Oberon have unique modular elements that can be dragged and dropped between windows or tracks -- very similar to Syllable modular interface design but the 'never overlapping' paradigm of a TWM makes sharing these of much more practical utility and value.

While many early TWMs extol the elimination of the mouse, modern TWMs allow for fast and clever mouse driven window interactions; double click to maximize a window, left-click to re-tile, scroll wheel or click-and-drag to size.

Menus, tool-bars, icon trays or even 'in-window' based desktop metaphors can be inherited reducing redundant wasteful interface clutter, excessive mouse movement, manual repositioning and window shuffling.

Oberon Gadgets is one of the oldest and most forward looking TWM interface designs, integrating GUI and programming interface with a high granularity graphic object oriented modules and unique user interaction; even the Oberon boot console 'windows' and is a TWM.

Benefits

· can be designed to less resource intensive then cascading windows systems
· make drag & drop or cut & paste operations across applications much easer
· automates window management
· less redundant mouse movement
· less eye strain (Swiss Study*)
· could leverage considerable existing Syllable technology
· allow for shared 'inherited' interface widgets further simplifying the interface

Summarily TWM is about making the window manager do the window management not the user. Clarity of information presentation, ease of interaction, interface efficiency, efficacy, transparency and form that truly follows function are paradigms that drive TWM design.

Syllable offer many unique technologies and GUI elements and concepts that could if carefully crafted move TWM design forward in both attractiveness and utility.

Links to hallmark TWMs:

· Oberon Gadgets
· Ion
· 8½ (Plan 9)
· TrsWM
· Ratpoison
· PWM

Screenshots:

· Oberon Gadgets
· Ion
· 8½
· TrsWM

Although the TWMs represented here for illustration may not be as 'sexy' or 'cosmetically evolved' as more conventional desktop interfaces; there is certainly plenty of room for elaborate facade and art to be added.

A Sort Of Summary

It's my belief that alternative operating systems are unlikely to ever compete with the mainstream, and while that's not necessarily a bad thing and should not be a source of disappointment emulating and copying well worn design brings all the limitations and flaws of those designs with them -- when capable 'alternatives' are available that can raise the bar of form, function, value and usability -- some of the very things that one would suppose are aim of alternative operating systems.

Emulating familiar interfaces has value but offers little motivation in attracting people to separate themselves from what they are currently using. I propose that alternative operating systems would be better served by truly offing an 'alternative', not just in core design and low level function but real utility at the operative user level and man/machine interface. Approached this way even if an alternative OS never achieves a large audience it will at the very least offer unique utility and value that can not be satisfies elsewhere or by other projects with similar goals. Frustrations and limitations in prevailing design is on place Syllable could distinguish itself by incorporating new interface options that pursue real form that follows function, and a 'better way of doing things'.

<  Aduna Updates AutoFocus Visual Desktop Search | Nooface Site Undergoing Overhaul  >

 

 
Nooface Login
Nickname:

Password:

Don't have an account yet? Go Create One. A user account will let you customize the site's content according to your preferences. It will also allow you to moderate the comments of other users.

Related Links
  • Oberon Gadgets
  • Ion
  • 8½ (Plan 9)
  • TrsWM
  • Ratpoison
  • PWM
  • Oberon Gadgets
  • Ion
  • 8½
  • TrsWM
  • Will La Chapelle
  • More on User Interface Innovation
  • Also by Editor
    •

    'An Introduction To The Track Window Manager User Interface' | Login/Create an Account | 6 comments | Search Discussion
    Threshold:
    Click REPLY To Post Comment
    The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
    Not going to happen (Score:0)
    by Anonymous Coward on Friday January 07, @09:33PM EST (#1)

    The problem with tiling is what you think is a benefit: the window manager manages the windows. I want to be able to stick a status window in an unused corner and count on it to stay out of the way. By trying to keep the screen full, you'd have to start shifting things around and resizing content when other "tiles" open and close, and users seldom want that. What faults there are with overlapping windows are not solved by tiling. Instead, look at what Apple has done with things like Exposé to dynamically scale and arrange windows when necessary to get some of the benefits of tiles.

    If there is any direction to look for an interface that eliminates overlap, it is in something that virtualizes the desktop to be of infinite size so that the user is free to size and arrange windows as they wish without the constraint of their monitor's resolution. Some window managers do allow multiple "desktops", but they're usually all the size of the monitor rather than truly considering the monitor as merely a view on a single, large 2D desk space. And I'm not even going to get into how 3D could eventually get a usable WIMP paradigm of its own, which would make tiling an even more questionable way to use the "precious" screen space.


    [ Reply to This | Parent ]
    Problems with a TWM (Score:0)
    by Anonymous Coward on Saturday January 08, @03:24PM EST (#3)

    There's the trivial issue that "TileWM" is as easy to say but wouldn't have been confused with the non-TWM X11 WM TWM (Tabbed Window Manager, approximately).

    But more importantly, using tiled WMs is a lot more successful if one of several additional characteristics are expected:

    1. Window content is automatically scalable through a mechanism like SVG or NeWS PostScript.

      Why: with this, even idiot applications written to always run fullscreen can at least be scaled down to an often recognizable postage stamp version of themselves, still active.

    2. Window content is predominately text, with a fairly obvious point of focus such as the cursor within that window. My memory of Oberon, in particular, is a perfect example of this.

      Why: Text applications almost always have a read or edit focus, and in some case even just showing a fraction of the key line of text can be sufficient.

      Note that in most window managers under X with focus-follows-mouse enabled and window-autoraise disabled (the original mode) user typically layer larger windows specifically to show just the subsections of interest in each. With text, this process can usually be automated and handled reasonably well in a tiling WM, but beyond text the issue becomes suddenly quite murky.

    3. Applications are all written to automatically put only relevant content in the screen space allotted to them.

      Why: This would allow the larger problem mentioned in (2) to be addressed. However, different user habits, particularly in non-text applications, complicate determining the most interesting content.

    In a tiled WM, there is a need to give the user control over which data within an app is displayed in a smallish window, or which part of an image to show, or in which apps and in which subsection of those apps scaling should be used, and so on. For the most part, I believe that the complexity of those controls can easily approach the same level of complexity seen in conventional WMs such a fvwm, sawmill, etc. There is a benefit to some users, in that once all of the needed customizations have been done, those users should indeed be largely free of having to mess with it, but I think that is only an achievable goal in tiled environment where the applications run within them are actively cooperating with tiling.

    This is far from the case for most end users, running Mozilla, pysol, xterm, yast2, emacs, or whatever. Is there a tiling-smart browser currently, that can keep itself optimally shrink-wrapped in respect to the rest of the shared space? Should emacs resize its window on the fly to fit the current function being edited - and switch to a smaller fontsize when its window isn't selected? If you cram four of each on the same screen, are they still useful - or does it just turn into a subwindow scrollfest? (and answering that Mozilla support tabbed windows is cheating here, since that's -certainly- violating tiling internally, no?)

    Let me point out before continuing that I think Oberon was an elegant, functional interface, and had much going for it. All of the applications I saw in it were written explicitly for Oberon, and the result was quite nice. However, I'm about to hassle a lot of the bullet points from the perspective of working with applications not necessarily written with tiling in mind, as mentioned earlier.

    • screen real estate is precious - � computers may multi-task well but people may not � people may multi-task marginally well but cascading windows make that impossible � complete windows management/customization using only the keyboard � avoiding redundancies through inheritance � reduction of manual window positioning and management
    [ Reply to This | Parent ]

    I'm not a robot like you. I don't like having disks crammed into me... unless they're Oreos, and then only in the mouth. -- Fry

    [ home | contribute story | older articles | past polls | faq | authors | preferences ]