August 12, 2004
I’m having some trouble getting online at the conference, but will be trying to blog a bit at ACM Hypertext 2004. (Right now I can’t get email, but I’m able to web browse/post.) Anyway, here are some notes from yesterday (which I’ll be updating a couple times after this first posting).
Doug Engelbart’s keynote: “Facilitating the Evolution of our Collective IQ: What Universities and Professional Societies Can Do.” He’s got more slides than he can get through, but Jim Whitehead says the slides can go online. Doug says he’s presenting a challenge to this community. His 1951 lifetime goal: “As much as possible, to boost mankind’s collective capability for coping with complex, urgent problems.” Not office automation, but human augmentation. An example of the kinds of problems that Doug wants to help people solve are the large-scale, urgent ones that the AC/UNU Millennium Project identifies.
What are we trying to augment? CODIAK: COllectively Developing, Integrating, and Applying Knowledge.
Doug says the only possible way the real evolution can come about is through unencumbered software. Patents kill our ability to evolve.
Technology allows accelerating rate of opportunities, but paradigms lag. How can we accelerate paradigm change so we can actually address big, pressing problems?
A DKR (Dynamic Knowledge Repository) that’s shared between organizations (NIC: Networked Improvement Community). This helps all the organizations to improve their work. It sounds like Wikipedia would be a very general knowledge repository, rather than one that’s really focused on improving the work of people in a particular field (and probably in some ways less advanced that Engelbart envisions). What kind of connection could be made, though, I wonder, between something like Wikipedia and the Bootstrap Alliance? I know there’s been some work trying to add Engelbart’s purple numbers to wikis and weblogs. But maybe it’s just not focused enough for the combination with his ideas to make sense?
Engelbart says: Marketing people will tell you that the market is the best way to get the cheapest thing to the customer that they want. Yes, if the customer really knows what they want.
Engelbart sees the need for an Open Hyperdocument System (OHS). For Engelbart, “hypertext” is not just about links. For example, even in 1968 his system allowed you to change your views of your document. “Hyper” for him is many new ways of working with documents. (There’s a great slide I have to get back to.) He’s unhappy with how things became so much about “pointing and clicking” (this from the inventor of the mouse). He views it as our poor inheritance from the office automation folks, trying to make things simple in the context of primitive word processors. It’s like not being able to express things in language, not having verbs, just having to point at things and gesture. The NLS/Augment approach had nouns and verbs (typed at a regular or chording keyboard, of course.) Engelbart’s approach is obviously more powerful. It does mean, however, that at first it’s like moving to a new place and having to learn the language there. Shows some of the NLS/Augment verbs and nouns. Had a shorter list for beginners. Had ones for subdomains — email, spread sheets — in the 1970s, rather than at time of 1968 demo. The command language, combined with highly flexible and explicit addressing, allowed all kinds of powerful macros. He has some cool-sounding example I’m not completely following, where the system expands out the addresses of things he’s working on and people he’s working on them with, whatever he’s currently working on. From the 1960s, the ability to connect to any object in any file (to address it) — not just jump-links. There’s a Draft OHS Project Plan (The HyperScope) from 2000, is on Bootstrap site. He also suggests people check out a 1992 paper (“Toward High-Performance Organizations”) and the 1962 report we have in The New Media Reader. The slides have a cool walkthrough of how someone might read the OHS paper in Augment (the system in which he wrote it). Showing different views of different levels of detail of different parts of the document. How different it is from reading the HTML version, where all people can do his scroll and jump! Shows a plan for transforming legacy files, using XML and link database to add high-resolution addressing, then having view generators that make the whole thing available in today’s web browsers with a minimal set of new user controls.
The kind of discussion he envisions taking place via high-resolution linking is quite similar to what people try to put together via blogs and trackback. But being able to reach into many more types of documents. I’m sure Matt Webb, sitting next to me, also typing, is writing something similar. Engelbart talks about publishing immediately, rather than waiting for peer review (as the high-energy physics folks do now).
He’s running out of time, pulls back to see future slides. Sad that he’s using PowerPoint. He told me at the reception last night that he wanted to use the new Augment client they’ve been working on — a replacement for the old ones where the areas (windows) can’t move and the connection has to be made over a modem. But they’ve recently run into a problem with this client’s networking. He talks briefly about who might be flexible enough to start the work he’s advocating — hinting broadly that the community around ACM Hypertext should start.
Andruid Kerne asks about the idea that we can have repositories of the best knowledge? Shouldn’t it be heterogeneous? Doug says that we should work hard to be able to represent disagreement, etc.
Mark Bernstein asks why this has been so hard to accomplish, going back to 1987 (first HT), or 1968. Doug says, there are a number of possibilities. One is that it’s bullshit. The audience laughs. But he thinks the main issue is paradigms. And people in IT taking the tactical, rather than strategic, approach.
Monica Schraefel asks, how can we evolve in a way that isn’t typing keystrokes? Your examples sound like vi to me. Doug basically says, Good question — why aren’t people working on that? And yet he’s not convinced that what is easy for a beginner (like MS Word instead of vi) is any more sensible than thinking we should build up the tricycle rather than moving to a bicycle.
Next session, combining two themes of media hypertext — scholarly and story-oriented.
David Kolb talks about writing a hypertext and a book at the same time. The compact conclusiveness of the book and the open digressiveness of the hypertext. It’s not an informational hypertext or a literary hypertext — but an argumentative hypertext. The tension between focusing conclusions and opening space for readers to engage, explore, and come to conclusions. Philosophy has this tension always, but it’s exaggerated. In writing the hypertext, there’s the tension toward writing little mini-essay nodes (self-contained, with mini-conclusions, like the library model of the hypertexts George Landow creates with his students). There’s tension toward creating constant roadsigns, to let people know where they are (he shows some they’re working on for the web version).
(The one for this conference, which is smaller, has a navigation bar with fixed links, a changing set of links to nearby places, and information on link types. This was put in, in part, in response to the peer reviewers of the hypertext.)
He resisted pressures in various ways. One was creating short nodes. (Though on the web they’ll be less short, because there are images on almost every page.) Some of these short nodes reappear repeatedly in the hypertext, but with different links within them. This is similar to some of the link cycle work people have done in Storyspace. Also, he has some links that make almost-disorienting jumps — so that the argument would gradually come into focus, rather than proceed rigidly, step-by-step. He also has a set of pyramid outlines (outlines within outlines) that have tangles of nodes at the next level down. And there are some areas that cannot be reached by the outline, only by associative links from things within the outlines.
So, the hypertext is composed of many regions, which one explores. There’s an outline that gets you to some regions, but not others. People will read it many different ways, and probably quite differently from the book. The goal was to have the book and hypertext published together — but some difficulty is arising with this plan.
Issues. How to make sure that the reader sees crucial nodes? Rich link structures can make this difficult, but the navigation can be arranged to help. He has pages with few links, and also with many links. He describes a kind of stack-based reading, similar to how he says people read philosophical texts generally. He uses link types to help people understand the links (what is at the other end, what the rhetorical goal of the link). The link types grew greatly during work on the hypertext for the conference, and he’s going to try to restrict them for larger project. He’s worried that type links will discourage readers from following the large-leap links. And how does this all tie in with readers who drop in via web searches? And then some more conceptual issues. Followed by a return to the “book vs. hypertext” question. He concludes that it’s not a zero-sum game.
I think about how it might integrate with the web more. How can it matter more that the web version is on the web? Maybe a series of thought experiments. What would it change if the book went up as a path through itself, as a blog, over time? Accepting trackbacks and comments? But responding to these doesn’t seem part of how Kolb imagines himself as an author. An argumentative hypertext, but not wanting to engage in small-scale argument, in direct exchange with readers. (Is there any way that we could peer review an argumentative hypertext that is the kind of argument that takes place in the blog(o)sphere? Of course, in a way the blog(o)sphere has its own, rather different, peer review.) Do blogs actually encourage/allow the development of the sort of scholarly argument that someone like Kolb seeks to create? To go further afield, what if Kolb had a wiki? Getting into edit wars would certainly not meet his goals, but what of readers adding more information about (say) the places that are discussed? But maybe the path that would make the most sense is taking up the question that Kolb himself poses. What about people who drop in from Google? Is there a way that a scholar like Kolb would want to create an experience like The Book of Endings, where each piece is meant to be a place to drop in and/or drop out? What about an RSS feed of the book blog? And so my thoughts continue…
Richard E. Higgason gestures toward Mark B’s writing that there are no good hypertext mysteries. Higgason says he’s always felt like reading many hypertext fictions was like trying to put together a mystery, and mentions others (including Jill writing about “clues” in her Afternoon paper). He approaches Lust in a self-consciously detective/PI manner. Especially appropriate given the violent (talk of blood, screaming, knives) language of the piece. Higgason develops this, but then notes that he’s only looking at the content of nodes, not the structure of links. Looking at the link structure shows how essential the Prologue is. Many starting points there, almost all of which lead to a nine-node cycle that reads like getting caught in memory. But a few lead to “Summer” — which can be read as a happy ending. One word (“sun”) starts one on the “this night” path which takes one linearly almost through the whole text. Eventually, Higgason comes to realize there’s a fifth man, never mentioned by name, who is who the narrator is with at the time of narration.
Mark Bernstein points out that people have been puzzled by Lust for years. What would happen if you made it simpler, clearer, less frustrating? He talks about similarity between links and metadata, and how changing the structure of either changes the meaning. And that has implications for our thinking about things like adaptive hypermedia.