2003-08-27 Finding the Story - Broader Applicability of Semantics and Discourse for Hypermedia Generation http://www.ht03.org/papers/pdfs/8.pdf Abstract from http://www.ht03.org/papers/ "Narrative, Semantics, Discourse, Hypermedia, Clustering, Concept Lattices, RDF, SMIL: Generating hypermedia presentations requires processing constituent material into coherent, unified presentations. One large challenge is creating a generic process for producing hypermedia presentations from the semantics of potentially unfamiliar domains. The resulting presentations must both respect the underlying semantics and appear as coherent, plausible and, if possible, pleasant to the user. Among the related unsolved problems is the inclusion of discourse knowledge in the generation process. One potential approach is generating a discourse structure derived from generic processing of the underlying domain semantics, transforming this to a structured progression and then using this to steer the choice of hypermedia communicative devices used to convey the actual information in the resulting presentation. This paper presents the results of the first phase of the Topia project, which explored this approach. These results include an architecture for this more domain-independent processing of semantics and discourse into hypermedia presentations. We demonstrate this architecture with an implementation using Web standards and freely available technologies." Goals for the "GigaCE Topia" project: when you ask a question (a query) it pulls together items - concepts - media items the results come back not as a list but as a SINGLE PRESENTATION [i'm thinking iPages here]. "semantic relations between media components form basis for discover structure" ... topia does two things: 1. finds all matching objects 2. another search to impose a narrative (find a story) that connects these objects works on RDF marked up stuff. [woo!] there's an online demo. [woo!! this would be a fantastic demonstration of the benefit of marking up a controlled domain, eg a website] now the result of the search is an *outline* [ahem, a hierarchy, sorry ted nelson...] which is illustrated with found pictures. the outline structure comes from the commonalities between the objects. topiary hypertext is a concept: - calligraphic hypertext - start with no links, add one link at a time, content and links are separate - sculptural hypertext - start with fully connected graph, selectively remove links - topiary hypertext - plant relations and new relations automatically sproat, new additions gracefully grow into existing whole. user/gardener can choose/make own paths through garden [see also: wikis. see also: poiesis, philosophy of information, constructivist cyberspace] topia is like this. topia's story: (user query) semantics (concept lattices) -> clusters (cluster analysis) -> discourse (style sheet) -> presentation [i like very much the discourse step. that it's not just node/arc, but the entire thing is a conversation.] the concept lattives bit is nice: it figures out what concepts (attached to the found objects) are important, and so you can impose a hierarchy by what concepts are dominant and what sub-concepts divide those up. they use a basic discourse structure: . hierarchy . sequence . recurrence . tangent [a book, is a sequence. chapters, subsections, read in order. a recurrence is a main theme, concept transclusion. then there are tangents which is like serendipity. this is his BASIC HYPERSTRUCTURE -- a datastructure that intrinsically has narrative within it, that's enormously important. wow. there's a lot to learn from the hypertext community and how they build stories into aggregations of data... and this is just one style! you have your style, then in the algorithm to do "discourse" you sort your concepts and results into that, and it makes it more understandable. much better than just "here is the picture box, here is the list of news articles bit" -- that's formatting not narrative. maybe my related-posts search should work like this. instead of an ordered list of nearest-posts, look for other metrics and find link-heavy posts, links, etc, and have a narrative template (or mould) that the results flow into.] future work includes more complex discourse that works like this. the teasers look cool, about how the computers and humans can work together. [this is proper hypertext meets the semantic web stuff. fantastic, we can learn loads.] more: http://homepages.cwi.nl/~media/conferences/HT2003 http://topia.demo.telin.nl -- this is the demo [the demo is remarkable. search for battle, and it divides up artists, themes, material, etc. this would be *awesome* for the Public Lettering Record.]