A second set of studies, by Josh Draper. In these, rather than attempting to visually sync up the time that the webcam images were taken, they are displayed behind the text of the discussion. As the discussion is replayed [using the same context-sensitive interaction metaphor as the previous studies], the appropriate images are shown.

split screen and tsunami interface/behaivor on the text. thinking about the difference between the text and cams. text comes in with no regular rhythm. cams are regular - some fps. on playback the cam images would change per some interval regularly. moving up and down through the tsunami text would do a fast CART like scrub of the images.

come to me giselle. a file is sent and recieved. recipeint goes view-in-view. a little interface appears maybe.

Studies in auto-generating webcam captions in relation to archived chats. The text elements' vertical location is determined by their timestamp, resulting in an time-sensitive record of the chat. Images grabbed from webcams are interspersed through out the conversation at the user's discretion. The text element which corresponds to that image's time becomes an auto-caption for that image, allowing for the creation of a text/image narrative over time.

First study - all elements displayed.

Second study - elements greeked out depending on distance from area of interest - all elements still displayed.

Third study - non-caption text elements greeked out and only displayed when not interfering with area of interest - auto-captions visible.