2003-08-27 Which Semantic Web http://www.ht03.org/papers/pdfs/7.pdf Abstract from http://www.ht03.org/papers/ "Through scenarios in the popular press and technical papers in the research literature, the promise of the Semantic Web has raised a number of different expectations. These expectations can be traced to three different perspectives on the Semantic Web. The Semantic Web is portrayed as: (1) a universal library, to be readily accessed and used by humans in a variety of information use contexts; (2) the backdrop for the work of computational agents completing sophisticated activities on behalf of their human counterparts; and (3) a method for federating particular knowledge bases and databases to perform anticipated tasks for humans and their agents. Each of these perspectives has both theoretical and pragmatic entailments, and a wealth of past experiences to guide and temper our expectations. In this paper, we examine all three perspectives from rhetorical, theoretical, and pragmatic viewpoints with an eye toward possible outcomes as Semantic Web efforts move forward." 2nd gen hypertext systems: gIBIS, germ, NoteCards, Aquenet, HOW all geared to support semi-formal representations Dexter model & Halasz's Seven Issues tried to capture this "Semantic Web" rooted in this other tie-ins to HT community: - 'Bridge Law' research (Bieber) controversy: is the semantic web just covering the ground AI did? great promises, nothing happens? qs: . who needs it? . does it work? . is it safe? <-- this is the question to concentrate on today references http://www.well.com/~doctorow/metacrap.htm [:)] summary: - people lie - people are lazy - people are stupid (or make mistakes...) - people don't know things - schemas influence results - there's more than one way to describe something ...these are concerns people already have 2x2 matrix for a framework for understanding the semantic web: - who? human/machine - what? particular/universal 1. human + particular: much of current Web use 2. human + universal: library of alexandria/taming the Web this assumes that the web is unordered, and so the semantic web turns the web into a big library 3. machine + particular: federated knowledge base example the rdf thing, link all the relationships together, so the web is a big database 4. machine + universal: knowledge navigator perspective meaningful structure so software agents can compare and filter information in new and exciting ways... this is just like "Knowledge Navigator" [a video clip from the www] http://de.unna.org/unna/movies/Apple/KnowledgeNavigator/ KnowledgeNavigator65kbit.mov So 3 theoretical problems: . additional overhead: someone needs to learn the syntax and semantics of the formal language to put the microwave on the Web . tacit and evolving knowledge: a microwave is a type-of oven. Or is it? you control in a completely different way. it doesn't heat with heat, you can't turn it up to 400. [popper: the dog's nose isn't a nose.] . the situated nature of knowledge heating a burrito has different representational needs that the buyer who is comparing specs for a kitchenette [hear hear! context is part of the thing, not separable from it] and 4 pragmatic issues: . metadata and community metadata signified membership. different cataloguing rules, etc . the high cost of metadata you have to choose between systems . metadata authority and trust eg porn sites using entire dictionaries in meta tags . metadata consistency html semantic markup used as visual markup however: a google-like approach works most of the time, is robust, and has social evaluation. these sort of things can't be universal, navigate knowledge within domain, with a purpose. plus, other problems: is knowledge stable enough? is knowledge intrinsic or extrinsic? cost/benefit (do people really want to publish metadata that lets them get compared with competition)? and what is the role of negotiation in representing this knowledge? [good paper! but a little old? hm. there's a lot more detail - and evidence - in the paper.] questions: from the W3C Ontology WG, somebody says that a lot of people involved are from the AI community and they've come across these problems a long while back, and the semantic web takes these into account... ontologies are indeed domain-specific, content-specific, there is no one way of viewing the world. [interesting point] the questioner also asks about the role of *inference*. the chair mentions "weblets" which are small manifestations of the semantic web (context and domain specific) which are joined together with something like google. this is one of the two semantic web schools -- the other being the world-encompassing thing. an organiser shouts: use the conference wiki! chair: "yes, use the wiki! then we'll turn it all to rdf and connect it together and then it'll all sit there and... fester."