Static and Dynamic Semantics of the Web
Christopher Henry Fry, Mike Plusch, Henry Lieberman · The MIT Press eBooks · 2003
this paper, we discuss the need for dynamic semantics, and show how dynamic semantics will be enabled by and useful to the new generation of intelligent agent software that will increasingly inhabit the Web. Though there will be autonomous agents, the more interesting agents will cooperate interactively with humans, helping them to achieve their goals more efficiently than a user could on their own. We also present a new language, Water, that fully integrates static and dynamic semantics, while keeping close to the XML framework that has made the Web successful. Beyond that, we envision that Web end users and Web application developers will not be routinely writing code directly in Water or any other formal semantic language, but instead avail themselves of Programming by Example [Lieberman, ed. 01] and other interactive agent interfaces that will hide the details of the formal languages. But transition to these future interfaces will be greatly aided by a foundation that can cleanly integrate static and dynamic semantics. ___________________________________________________ A semantic net [Woods 75] is a network of concepts linked by relations. The Web is, of course, a network of pages, each containing text, pictures, other media types, and links to other Web pages. Though the Web has far less structure than typical AI semantic nets, the Web pages that constitute the nodes of the Web's network often do represent concepts and the links between them represent relations between those concepts. For example, my home page is the Web's representation of me, in a sense. The links leading off my home page - to my publication list, my e-mail address, courses that I teach, etc. - represent the relation of me to, e.g. the articles I have published. The only problem is that these re...