| |
  |
| Billowing Fisheye Text Interfaces |
|
 |
|
posted by Editor on Monday December 22, @06:03PM
|
|
|
|
|
zeitgeist writes "Billowing or Fisheye text interfaces have been around since at least 1999. The idea is similar to the OSX Dock's technique of enlarging icons near the pointer. Here's an implementation in CSS/Javascript, one in Flash, and a third in Java." The original technique proposed allowing the size of type to be varied dynamically for any purpose, i.e. locally enlarging and shrinking type to highlight areas of a text. It also proposed combining billowing with undulating (using either 2D or 3D simulations of rippling pennants in the breeze) to get away from rectangular windows.
|
|
|
|
< Showing PersonalBrain with Flash | Telepresence Via Zoetrope > | |
|
|
|
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.
|
|
|
|
|
|
This discussion has been archived.
No new comments can be posted.
|
|
|
|
|
by Anonymous Coward on Friday December 26, @09:59AM EST (#1)
|
|
|
|
|
a. you can only see 3 lines of text, the rest are unreadable.
b. the size of the target is proportional to the number of items instead of being fixed. the more items, the more precision you need with your mouse to hit the line.
c. for visual context, you have the scrollbar, that will tell you where you are in the list. it requires less space than a lot of smudgy lines that no one can read.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
So, this implementation is useless because it has been taken too far. The level of compression used obscures too much of the surrounding context, the preservation of which is the motivation for the use of "fisheye" variable enlargements.
By maintaining a greater readability in context (not diminishing so greatly or as quickly) the information could still be well represented and navigated. Additionally a feedback mechanism could be provided via the scroll-bar as to the percentage of the information traversed / current location.
I definitely agree with you that the level of precision required to navigate and perform selections outweighs any benefits in this case. Perhaps the additional lesson is that this isn't necessarily good for menus, or for interacting with the data itself; instead this could find use for simply interacting with the data *views*, such as examining the detail on an otherwise trend-represented line graph.
I wouldn't call it useless, just poorly used in this case.
|
|
|
|
|
|
|
|
|
by Anonymous Coward on Monday December 29, @12:39AM EST (#3)
|
|
|
|
|
I'm not sure what use this whole technique has for text. With text a text list, you can either read each item, or not. Unless zooming far out exposes some kind of overall structure, that's not visible close up, what is the point?
I would image that something like this might be more useful for browsing image thumbnails, where zooming out can give a useful overview.
It also bothers me that everything else on the page moves around when you browse. Anyone know some way to keep things a bit more static?
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
It's a very difficult technique to get right. The math in good fisheye menus has to be very precise to get the spacing right.
That said, the post misses a *very* important point. Recent work, CHI 2002 & 2003, shows that the expanding menus work best if the menu expands *prior* to the mouse reaching the icon. In this case, using trajectory sensitive expansion, an actual performance benefit may be had, in addition to the ooh-ahh.
|
|
|
|
|
|
