Words in a text are wannabee â€˜hotâ€™ links. So treat every word on a web page as a launch pad for searching, translating, pronouncing, defining, thesaurizing, etc. An old digital dream, already explored in dictionary/encyclopedia mode by Gurunet (â€œSee a word on your screen? Just Alt-Click it!â€), and from a more multilingual/translation perspective by Babylon.
Now comes Hyperwords, the outcome of a Liquid Information project. On October 14, it is due to launch a comprehensive new word-click service. Users will be able to right click on a word or phrase on a web page and access a dictionary definition, a Wikipedia reference (if relevant), a web search engine, a translation engine, a blog search, and more.
If you select a phrase from a web page, the linked translation engine will run the usual clunky word-for-worder (e.g. have an assistant mark-up interesting passages is faux-Frenched as avoir une marge bÃ©nÃ©ficiaire auxiliaire intÃ©resse des passages (lit.= ‘have an extra profit margin interest passages’), but the system can at least correctly guess spelling errors in its dictionary: on Hyperwordsâ€™ own site, the word cooperation is mis-typed as *coperation but captured correctly by the dictionary link.
Hyperwordsâ€™ inventor Frode Hegland makes big claims about his product in a recent French language interview. He suggests it gives a third dimension to the hitherto 2D world of websites, offering better content visualization and interlinkage for â€œknowledge workersâ€. He reckons the menu that displays when you click on a word â€œis hierarchical rather than a long listâ€ (doesn’t look like it yet), and hints that Hyperwords will only offer what is needed at a given moment in your knowledge work (how does it track your needs?). Hegland claims its is â€œthe same principle as human language: you can say a lot of things, but you donâ€™t need to think beforehand about all the sentences you are going to speak. Expressive possibilities develop in the very process of thinking.â€ (my translation).
This Liquid Information vision explicitly invokes the pre-web luminary Douglas Engelbart, who pitched the augmentation of human intelligence through rich human-computer interfaces as an alternative to the old AI agenda of installing automatic text understanding inside the computer. Yet Hyperwords looks more like a Ted Nelson type hypertext application, especially as the click menu offers a link to Nelsonâ€™s about to be implemented(?) and rather laborious Transquoter functionality (â€œbrings in quotations from textfiles and web pages all over the Net and concatenates them into a web page, keeping each quotation connected to its sourceâ€).
A service such as Hyperwords will ultimately depend on the value of the resources it taps into. So far, it looks like a slightly faster way to access existing Google and other resources, where (to parody Richard Dawkins’ definition of the selfish gene), words are just a search engine’s way to generate new searches.