The Dasher Project, a zooming text input interface.
I've seen Zoomable User Interfaces for filesystems and ZUIs for collaborative working; I'm not convinced. They're just novel interfaces to problems that could be solved in other ways. That's not to say ZUIs don't have a place, but we should look at what problem they're actually trying to solve. That is, adding a dimension of detail to otherwise flat data. And that's got to start much lower down than a groupware system, at the level of text documents. Our computer metaphors are based around single strings of texts structured by the newline and EOF [end of file] special characters. And because of this: We have discrete documents; Datatypes in programming languages are limited by this same metaphor, our datastructures are still pretty unstructured; everything is strings, strings, strings.
The alternative, and why ZUI researchers are looking too high: start by structuring the plain text document. Rebuild the last 50 years of computer history. Let us have trees, graphs and tables as fundamental datatypes to accompany the list and associative array. Constructs that break free of the text line. Only then work on replacing the spreadsheet, the www browser, the filesystem.
(Of course, it won't happen like this. We need to work in the meantime. Until then, it'll happen gradually because the requirement's there. Text is being structured by XML. Slowly standards and standard tools will emerge. Then application history will be infected from the present, Excel rewritten as ZigZag, and the rest. And the rest.)
The Dasher Project, a zooming text input interface.
I've seen Zoomable User Interfaces for filesystems and ZUIs for collaborative working; I'm not convinced. They're just novel interfaces to problems that could be solved in other ways. That's not to say ZUIs don't have a place, but we should look at what problem they're actually trying to solve. That is, adding a dimension of detail to otherwise flat data. And that's got to start much lower down than a groupware system, at the level of text documents. Our computer metaphors are based around single strings of texts structured by the newline and EOF [end of file] special characters. And because of this: We have discrete documents; Datatypes in programming languages are limited by this same metaphor, our datastructures are still pretty unstructured; everything is strings, strings, strings.
The alternative, and why ZUI researchers are looking too high: start by structuring the plain text document. Rebuild the last 50 years of computer history. Let us have trees, graphs and tables as fundamental datatypes to accompany the list and associative array. Constructs that break free of the text line. Only then work on replacing the spreadsheet, the www browser, the filesystem.
(Of course, it won't happen like this. We need to work in the meantime. Until then, it'll happen gradually because the requirement's there. Text is being structured by XML. Slowly standards and standard tools will emerge. Then application history will be infected from the present, Excel rewritten as ZigZag, and the rest. And the rest.)