HOME
The Nexus of Search, Models, and Meaning When you're trying to do something on the internet -- find a digital camera or the meaning of life -- getting search results is just a good start. You still have to make sense of the results. This frame is a snapshot from THOUGHTSHUFFLER, a software prototype of ‘a spreadsheet for ideas’. The user interface presents rows and columns containing text. Clicking on a piece of text causes a ‘shuffle’ of the display -- in an ordered sense, not random -- to create focus on that text. As a result, the user can: drill down into a document to get the 'gist' or focus on a sub-topic submit queries and review results from a search engine see the context of every instance of any combination of phrases juxtapose similar texts and enable a fine-grained, easily controlled ‘compare and contrast’ display Behind this front-end is an engine for proposing the meaning inherent in any text in terms of short phrases that it suggests ‘make sense together’. The set of phrases, called ‘an entailment’, is experienced by the user as a valid distillation or essence of the text. As a result, the user can: request a summary of a page or document focus on a subset of content by sub-topic and get a summary give feedback on the validity of any THOUGHTSHUFFLER proposal move through hierarchies of detail in the document do any of the above on search results THOUGHTSHUFFLER's user controls and display are simple and can scale to long documents or large numbers of search results.   Although far more constrained than the full interactive application and displaying obsolete interface designs, here is further online content: a scenario-based walk-though and another with web content; an authoring example; a self-explaining streaming movie; and a limited demo. THOUGHTSHUFFLER is being developed by Paul Pangaro, who started prototyping knowledge-based user interfaces at Negroponte’s MIT lab [I]. He later developed a version of THOUGHTSTICKER [II], derived from Pask [III]; designed interfaces to search engines [IV]; and wrote code to generate static HTML from a domain model [V] [VI]. He has written, presented, and lectured on the prescriptive advantages of a cybernetic approach to the design of complex systems. He is often focused on problems in search, served as CTO for Snap.com at Idealab , and created new ranking algorithms for The Poetry Foundation. Pangaro holds a BSci from MIT in computer science / humanities, and a PhD from Brunel UK in Cybernetics.
© Copyright Paul Pangaro 2001 - 2005. All Rights Reserved.