144 |
they show thumbnails, or what they show). |
they show thumbnails, or what they show). |
145 |
Emphasis seems to be on the methods for generating the f+c structure, |
Emphasis seems to be on the methods for generating the f+c structure, |
146 |
i.e., finding landmarks, paths between nodes, what nodes to show, etc. |
i.e., finding landmarks, paths between nodes, what nodes to show, etc. |
147 |
Some features use in the example views: |
Some features used in the example views: |
148 |
|
|
149 |
- nodes are shown as cubes (focus node as a sphere) with titles |
- nodes are shown as cubes (focus node as a sphere) with titles |
150 |
- node size/height/color depicts importance/access frequency |
- node size/height/color depicts importance/access frequency |
302 |
|
|
303 |
When a link is followed, the torn shape of the target buoy |
When a link is followed, the torn shape of the target buoy |
304 |
animates to the full shape of the document, as if larger and |
animates to the full shape of the document, as if larger and |
305 |
larger parts were torn off. |
larger parts were torn off. [XXX: screenshots] |
306 |
|
|
307 |
The hardware implementation uses a noise texture for creating |
The hardware implementation uses a noise texture for creating |
308 |
the variation in the torn shape with texture coordinates tied |
the variation in the torn shape with texture coordinates tied |
345 |
Paper |
Paper |
346 |
----- |
----- |
347 |
|
|
348 |
- fragments of documents are shown |
Although the silhuettes of the buoys are different, |
349 |
|
the fragments of the documents still seem quite similar. |
350 |
|
The user could identify the related documents by |
351 |
|
reading the text of a fragment, but that requires attention. |
352 |
|
|
353 |
- fragments of PDF files seem similar; target fragment |
Using a unique background texture for each document changes the |
354 |
not useful per se |
situation dramatically: the user can perceive the identity |
|
|
|
|
- using a unique background texture for each, situation |
|
|
dramatically changes. |
|
|
|
|
|
A unique background texture consisting of random shapes |
|
|
and colors is chosen for each document using the identity as a seed. |
|
|
The background textures help the user perceive the identity |
|
355 |
of the most familiar documents at a glance, |
of the most familiar documents at a glance, |
356 |
even when only fragments are shown. |
even when only fragments are shown. |
357 |
|
Furthermore, when moving from node to node, the pre-attentive |
358 |
|
cues of identity help the user maintain a sense of direction. |
359 |
|
|
360 |
|
The background textures are randomly chosen using the identity as a seed. |
361 |
|
That is, each document has a unique backround texture, but the texture |
362 |
|
is not in any way related to the contents of the document (except |
363 |
|
that a hash of the contents could be used as an identity of an immutable |
364 |
|
document). |
365 |
|
That way, the textures in any view are as different as possible, |
366 |
|
even if the documents are similar. |
367 |
|
Furthermore, the unique background of any document can be instantly drawn, |
368 |
|
as soon as the identity inside the system is known. |
369 |
|
|
370 |
The distribution of the textures is designed to be maximally |
The distribution of the textures is designed to be maximally |
371 |
diverse and recognizable using a rough, qualitative model |
diverse and recognizable with respect to a rough, qualitative model |
372 |
of visual perception. |
of visual perception. |
373 |
For example, backgrounds with random pixels would all |
For example, backgrounds with random pixels (noise) would all |
374 |
look the same, because the pixels are not perceived individually. |
look the same, because the pixels are not perceived individually. |
375 |
However, shapes and overall colors can be independently random to |
However, shapes and overall colors should be independently random to |
376 |
maximize diversity. |
maximize diversity. |
377 |
Making the backgrounds repeating creates well-defined patterns |
Making the backgrounds repeating creates well-defined patterns |
378 |
and improves recognizability when fragments are shown. |
and improves recognizability when fragments are shown. |
379 |
|
|
380 |
- readability |
Our hardware implementation (libpaper) uses a small set |
381 |
- preattentive |
of *basis textures*, which are non-linearly combined on the |
382 |
|
GPU to create a large set of recognizable shapes. |
383 |
|
The coordinates of the component textures are |
384 |
|
randomly chosen affine functions of the paper location, |
385 |
|
but repeating with a randomly chosen *repeating unit* |
386 |
|
(a parallelogram). |
387 |
|
|
388 |
|
At each pixel, the combined values of the basis textures |
389 |
|
are used for interpolating between the colors |
390 |
|
of a small, randomly chosen palette of *compatible* colors. |
391 |
|
That way, the colors and shapes are independently random |
392 |
|
and the palette can be restricted to light colors to |
393 |
|
maintain readability. |
394 |
|
|
|
- hw accelerated impl (libpaper) |
|
|
- colors |
|
|
- basis textures |
|
|
- texcoords, repeating units |
|
|
- combiners |
|
395 |
|
|
396 |
Implementation on the Gzz platform |
Implementation on the Gzz platform |
397 |
================================== |
================================== |