2003-08-29 What is hypertext? Peter J Nurnberg http://cs.aue.auc.dk/~pnuern/ intro: came up with the term "structural computing" [looks interesting: http://www.cti.gr/sc3/ -- "Structural computing as a school of thought (e.g. Structuralism)"] not the idea, but the community. "this talk is a hypertext" -- "primarily structures built over existing work" [heh. on the left screen ken is playing "heroes of might and magic 3". every so often he finds a roadsign, and on the roadsign is says which bit of the talk nurnberg is going to do next. oh yeah, the here is the "reader"] part of the talk that'll come up as we see the roadsigns. * axes: - "space" in which to locate work * example applications: - what hypermedia could be * related questions: - concerning the community * conclusions # 1. related questions first (ken killed an ogre): . where is hypertext? front end: in the user's mind, eg "spatial hypertext" application interfaces, eg emacs info and MS help (this, the front end, is often where web people think about the structure of hypertext) system: in clients/apps/middleware, eg callimachus in infrastructure eg hyperbase in dev tools eg construct (a dev tool which itself is part of the built environment [eg smalltalk in a non-hypertext way, i guess, and zigzag] back-end: in OS: HOSS in computing environment: ??? [when he says hypertext, he also says "structural awareness". what would it means to have structural awareness at the swap level, or on the network? iiiiiinterestin,] . what questions are important? philosophical/theoretical <-> design <-> implementation (there are a lot of these recently) <-> evaluation (there need to be more papers on this) <-> reflection . what is the purpose of the community? - standardisation - evangalisation (the community isn't comfortable with this [however, i think the crossover between this community and social software is enormous. the social/technical distinction that people have been fighting simply doesn't exist in this mindset. hypertext community as already where, say, the expanded weblog world is trying to be.]) - canonisation # 2. axes classifying work . manual/automatic structure generation manual -> support low-level knowledge tasks. this is the englebart theory. improve very small tasks you do often, eg you invent the mouse. another example: imagine a pencil tied to a brick, you wouldn't write with that. untie the brick, make high level tasks much more possible. eg, associative storage and retrieval. memex & trails, it doesn't actually do much. automatic -> a machine builds the structures for you. support high-level tasks. agents do it. this is data mining. the hypertext conference doesn't often refer to this kind of structure -- but could we learn from data mining people? microcosm straddles both of this, which is why it's interesting [ http://www.mmrg.ecs.soton.ac.uk/publications/Project-Microcosm.html http://wwwcosm.ecs.soton.ac.uk/ ] . contextual/essential structure contextual -> open hypermedia. believes that structure is completely dependent on point of view essential -> embedded markup, data has inherent structure contextual means that we can separate the two, some people choose different structures. [there are two tensions of the web. we want to bake in small scale structure and simultaneously get rid of macro structure.] . percieved/--- structure [another odd comment. on a system called 'microcosm' he mentioned the "southampton model of system design". i didn't know there was one!] perceived -> user's feel for structure. there is no structure on the www, it just exists between the user's brain and the display. it's just files. implemented -> system's "knowledge" of structure counts. eg hyperbases, structural computing. in principle it's possible to make structured datastores etc. [what do these axes mean for weblog software? could we have really different full-scale applications, not just doing small tasks but something really interesting? the people who want a weblog in rdf, etc, think it's important to have a implemented structure, and they believe that perceived structure is not enough: that you need it at the grass roots to build on. i'm inclined to agree.] # 3. example applications . digital libraries browsing interfaces, supporting serendipity [this is a social software issue too]. browsing in the hypermedia sense is much like browsing in the human sense... it's not like that with search-and-query. intellectual property, how to support 'linking' rights. say you want to implement a trailblazer system, to make money out of selling trails you make to content you didn't write. that's really hard indeed. on the web, the trail isn't even an actual object. . software engineering feature engineering: tracking *constellations* of changes to source (this is larger than a change to a file, which doesn't map to a feature, and isn't as big as a version. [again, social software. that CVS at the moment is dumb wrt the ways people actually do it. why do geeks insist on making software independent from the various ways people use it? ...because when it's tried, it's really hard to use for most people: MS Word.]) patterns: templates for aiding programmers . linguistics historical linguistics: how languages evolve -- typical language things are done as hierarchies, trees! but really people use wave models, and these documents are really hard to use, manage, version. they're very structured. wordnet: how words relate to one another [yay!]. and, dictionary makers, lexicographers. there's a conference in russia next week, and about half of the undergraduate papers are all about computational lexicography and *hypertext*! # conclusions what is hypertext? (what does this community do, not in general) -> structure in knowledge work what will hypertext be? -> more fields involved, more evaluation, more people. ==== during questions somebody references a story where TBL turns up with WWW and everyone says 'that's not hypertext, it has dangling links', so he goes away and takes over the world [laughs]. interesting point that standardisation inside a given field stifles research inside that field. for outside the field, fair enough, some people want that. question: copyright is a non-issue on the web because money is made in other ways (adverts, cost savings), so what's the problem here? answer: the bush model where people have professions as trail blazers. you're trying to sell access to a structure, not the content it points to! [this is interesting. in a way teaching is structuring, trails, and you sell that. and once you put a market on something you encourage innovation - this is the point behind patents - the problem is that: we have no way of dissolving markets once the job has been done (we need this!); introducing a market deliberately seems unfair, introducing scarcity. is there a non-damaging way of having markets? although, maybe, it's enlightened self-interest: if you have a market are people on average better-off than without one? even though there's less equality?] question: spacial hypertexts in people's minds. people's minds are often unclear. spacial hypertext is so that people have something to do before it *becomes* clear. should structure be clear? what about non-structure? is clarity achieved by interactivity? structure vs behaviour [this is a difficult and big issue, apparently] could this group have a low-level standardisation of the concept of linking? basics of linking -> inclusion, or reference... or both?