Episode Twenty Three: The Difficulty

by danhon

1.0 The Difficulty

I’m empty, today. Wife and baby are off for a week (possibly longer) trip to see the maternal grandparents, one of whom we’re worried about. I am now catapulted into bachelor land, with a profusion of games consoles, a PS4 I haven’t even plugged in yet, more movies and unfinished games than I know how to watch or play and the weird inkling feeling that I shouldn’t be ordering delivery pizza. Yet.

There are so many things that I want to do: like live-blogging Hyperland or watching through Stealth, Elysium, the Robocop remake, the Total Recall remake, and doing a commentary, or, more honestly, falling asleep for a bit. Everything feels quite disjointed, and it’s not just because the house looks like a 12 month old whirlwind has it it.

This is just a one-off, though. This is what depression can feel like. It might feel like it again, but the new medication that I’m on has taken the edge off, somewhat. It’s hard to explain what it feels like other than a complete emptiness; in one conversation I’ve had with a doctor it’s like looking inside your own head and finding something filling up the space, but that it’s a featureless sphere, impenetrable.

Medication, then. And writing.

2.0 Hyperland

With that, then, tonight’s activity: watching Hyperland[1], a 50 minute documentary written by Douglas Adams for the BBC. You can watch along, if you want. It’s important to remember that Hyperland was written in 1990, before any talk of the Information Superhighway; indeed what everyone was excited about at the time was the promise of Multimedia – the mixing together of sound, video and text into some new hitherto impossible whole. If you were a child in the UK in between 1984 and 1986 you likely participated in a nascent, country-wide attempt at producing a multimedia work; the BBC Domesday project[2].

I think it’s important to take a look at Hyperland (and the other historical fiction that we have, that paint a vision of a possible future) if only because it’s interesting to see what our grasp used to be. It feels, and here I recognise that I’m falling into the bias of the Wonderful Yesterday, that we had more fantastic dreams and that instead what we have now is some sort of worldwide crisis of lack of imagination. Was it because our reach exceeded our grasp, and now we’re able to simply produce what we had dreamed of grasping when we were younger?

Hyperland opens on a slumbering Adams, recollecting a dream that he had to get rid of his television: “I dreamt just to get rid of the thing. I thought, there must be a better way of spending my time, like walking to the nearest scrap heap and throwing the wretched object away.”

We see the power of television: of Margaret Thatcher showing how television can be so selective in the alluring images it presents to us while we pan up to yet another television in the scrap heap showing soft-core porn. The multitude of content that television offers, from an evangelical preacher to a documentary about breadmaking. And, of course, advertising. Adams is perplexed at this: “It seems such a waste of technology.”

“Maybe, in the future, we’ll find a better use to put it all to.” – and we cue dream sequence music with the strumming harp.

This is a familiar refrain. The potential of technology – and what Adams was seeing was the power of the image and its might in broadcasting singular images; nation speaking unto nation. Surely, there are those who ask, we can do better than this?

As we skip into the future, Adams sees news broadcasts of Japan pegging the Yen to the Dollar in exchange for California (we need only substitute China), and that the effects of Global Warming are finally becoming apparent (well, we all know about that). But, as Tom Baker interrupts, surely he is tired of linear, non-interactive television.

Baker asks whether Adams is tired of a television that sort of happens to you, that doesn’t involve you. And for us now, the familiar strike of lean-back versus lean-forward entertainment, of passive versus participatory content. Of user choice.

“My name is Tom,” he proclaims, “and I’m your agent.”

Ah, agents. We had such dreams for these personifications of algorithms in the belief that people would find it easier to act with a simulacra of a human being than with, well, rawly exposed user interface controls. The truth as it turns out, nearly twenty five years later, is somewhere in between. Now we have virtual assistants at airports, instructing us on how to dispose of our excess liquid goods, but every day, millions – billions – of people interact with something more like a command line prompt when they type their questions into Google’s maw.

As an Agent, Baker explains, he’s here to do Adams’ every slightest bidding. To fetch, carry, work tirelessly all on Adams’ behalf, always ready, always willing. An artificial and completely customisable personality, a software application running on his computer, providing instant access to any digital information existing in the world anywhere – or, what we would refer to as “on the internet”.

“Is there anything I can do for you now, Mr. Adams, sir?”
“Well yes, you can stop being so obsequious for a start.”

Now we’re presented with Baker’s control panel, which materialises around him. The thought that people would care as much about the *how* information is presented to them as opposed to its accuracy or timeliness (or, if in this particular imagined future, all information retrieved is correct and timely, all that matters left is the manner in which it’s presented) is interesting. We do not have, at the moment, this level of customisation in our interaction with services like Google Now or Siri. What we do have, instead of a curt/servile slider to adjust “manner” is, in some cases, some degree of control over tone of voice, but normally even this is added as an afterthought. Facebook still has, as language options, English (Pirate) and English (Upside Down), and on Google there used to be vestiges of such a sense of humour in the ability to browse the site in Pig Latin. But, as a rule, tone of voice is one-size fits all, and this has as much to do with the hard AI problem of natural language processing as it does will (how many writers does Apple employ for Siri, anyway?) Indeed Siri’s personality has come under fire for having enough of a point of view to make her unique.

All of the control panel options that Adams plays with are within the realm of current feasibility – they just require a big enough team of 3D modellers and texture artists for the dress and species/3D rendering options but as always, it comes down to the uncanny valley of actual user interaction through speech. I’m not sure what the state of the art is in accurately reproducing accents, but let’s just say I’m assuming a large number of voice talent artists would be required to build up the phoneme database or whatever.

(It’s here that we encounter our first application error and crash in a manner reminiscent of Mac OS 7’s Bomb, and the usual promise that the next version of the software will be more reliable. It may have been 1990, but it looks like that particular trope of the software industry had by then been incredibly well established.)

Baker then presents to us a crash-course in the history of hypertext, covering Vannevar Bush, Ted Nelson, the MIT Media Lab and the Multimedia Lab in San Francisco, then Robert Abel.

It’s the agent aspect of Hyperland that continues to stand out at this point. You have a sufficiently Turing-fooling avatar that Adams is able to converse with, a piece of software that can do natural language recognition as well as speech, summarise information about the subject of hypertext, pick out the most influential people in its evolution as a mode of communication and also talk reflexively about itself. It can also, Librarian-from-Snow Crash-style, take conversation forks and handle interruptions and then return seamlessly.

It’s interesting that one of the hallmarks of the future of interaction with information would require so much personalisation and assistance. As if humans would not be able to navigate such data on their own (arguably, we cannot), but that we would require instead a fake persona, one that we could converse with as if we would another human, to help us understand and navigate. One wonders if this is a particularly paternalistic view of how we need to experience the mass of information out there, rather than a more libertarian have-at-it attitude.

Here we digress as Adams asks why there are moving icons on the screen – Baker explains they’re “micons”, invented by Hans Peter Brandmo at MIT in 1989, essentially anchored animated gifs and we’re whisked off into demo land where a micon of Ted Nelson appears on top of Hans’ video interview and Adams taps it to jump to a video interview.

Adams asks how to go back – something we all asked when confronted with hypertext for the first time – and Baker quips that one could ask him, or instead tap the ‘go back’ button, which “looks like a minimalist fish hook.”

Adams complains that he’s “completely lost” and after a conversation with Baker, Baker provides a “guide” and they start at the beginning, with Nelson explaining the work of Bush. The “completely lost” and “guide” nature is familiar too: those who aren’t familiar with the internet can describe themselves as overwhelmed, and many of our first interactions with the net were through directory sites like Yahoo! and other editorial like the NCSA Site of the Day. When you could travel anywhere, view anything, and it was all at your behest, where would you start?

Nelson introduces us to Bush’s Memex through more video. At this point, and this is purely an artefact of this being a documentary made for television and the format of the documentary being video, as Nelson is speaking, various micons pop up (microfilm, for example, pops up as he explains that the Memex acts as a front-end for a store of information), it’s worth point out that there’s a degree of authorship on show in Adams’ demo. Because there’s no text, as such (and one would instead imagine that if Hyperland were an actual application instead of a film, then the subtitled text would itself be hyperlinked) it appears that there’s only limited user choice: editorial choices are made at opportune moments to introduce the reader to other concepts (an alternate explanation, though, would be that Baker himself as agent is injecting relevant micons into the video stream as Adams is watching, based on what Baker assumes or knows to be Adams’ level of knowledge about the topic. Seeing as we’re already practically envisaging strong AI, we may as well.)

Nelson remarks that Englebart (inventor of word processing and the mouse) was probably influenced by the idea of the Memex.

Adams then interrupts and asks for Baker to come back, and points out that he had been watching video of Brandmo, but now there appears to be a Ted Nelson strand running through everything (isn’t there always). He asks Baker: “Does it all interconnect?”

“Absolutely! The system is constantly at work in the background sifting its data for connections. Every time you start something, I start to line up other things that you might want to see.”

And again, we’re at a curiously different envisioning of hypertext and hypermedia, one mediated by impressively intelligent agents doing our bidding that are able to understand not only our intent, but, by implication, the semantic content of every piece of digital information on the world’s networks. It’s worth it to point out that the links between concepts in Adams’ future aren’t static as they are on our world wide web, they aren’t necessarily manually encoded as humans would do on Wikipedia. They emerge, agent-backed, from an innate understanding of the knowledge embedded in the system. It’s not clear how the agents would do this, but it implies a strong ability of natural language processing and meaning to understand how different concepts interrelate.

Adams starts Baker off on a query about “the Atlantic”. Baker throws up Ecology, Oceanography, Shipping, a Live Feed and Literature, all represented as micons. In contrast, Wikipedia will throw up the article for the Atlantic Ocean which covers subjects such as geography, cultural significance, ocean floor, water characteristics, climate, history, economy, terrain, environmental issues, bordering countries and major ports and harbours.

Adams taps on the Live Feed and up pop micons for the Labrador Basin and the Azores-Gibraltar Ridge. Adams expresses incredulity that he could actually access live video data from submersibles via satellite. On Literature, up pop micons for Coleridge, Melville, Conrad, Hemingway, CS Forrester and More…

And then, upon tapping Coleridge, micons for Critical History, Drugs, The Rime of the Ancient Mariner and Other Works. Adams taps on Other Works, and from then, Kubla Khan. But then Ted Nelson turns up, and Adams asks why.

Baker remarks that he “likes to arrange meaningful coincidences, and this is quite a good one.” The idea that an agent would engineer serendipity is in fact one that we *do* have: all the way back to Flickr and its interesting photos page and the intent behind Dopplr to help engineer serendipitous connections between travelers. This serendipity is now engineered in other ways where it’s not necessarily presented as serendipity or meaningful coincidence, but instead as something where the utility is far more stark: Jobs You Might Be Interested In, for example, or Other People Also Purchased. It’s not too hard to imagine a future where Google or another company with lots of access to browsing data starts providing recommendations for content, Stumbleupon-style, in a truly mass manner.

Nelson then takes us through “that unifying world” of anyone being able to add documents to a central repository, accessible through telephone, laser beam or satellite, it would have to be available to everyone, everywhere. And because of literary tradition, he calls it Xanadu. And as we cut back to Baker reading the text of the poem as it scrolls by on screen, Adams taps words for definitions and phrases for explanation.

Baker then shows Adams a piano roll of Bach’s fugue in C Minor that Adams fails to recognise until its played, and Adams supposes that such a system, one that is good at recognising shapes and dealing with music, could also be good at teaching music and unearthing the patterns that are hidden beneath the surface.

Now we see Robert Winter’s interpretation of Beethoven’s Ninth Symphony in a demo of an early multimedia application published by The Voyager Company in 1989, which, seeing as it’s a real application that shipped, isn’t one that I feel we can learn particularly much from as it’s an example of the promise of multimedia applied and can draw a direct line from Robert Winter’s CD-ROM all the way through to something like The Orchestra on the iPad[3]. In a way, it’s disappointing to see that conventions set down in 1989 still hold strong and that there haven’t been other, more successful ways of presenting music.

When we come back, Adams is musing on the fact that once information is digitised, it’s easy for computers to see shapes and patterns in the data that’s been collected. Baker races ahead when Adams mentions Vonnegut and, mentioning that he’d scanned all of Vonnegut’s work, found a piece of writing about shapes in stories that he presupposed Adams was going to mention in an “I’m Feeling Lucky”, or autocomplete-y kind of way. Again, because Baker is a personified agent, his interruption feels more of, well, an interruption, rather than the UI trope now of autocomplete-as-you-type being the similar hovering over your shoulder as you think, but the conversational nature intrudes more.

What Baker does here is interesting and points more to his fantastical intelligence. Assuming (correctly) that Adams had just spotted a link between the ability of a computer to identify patterns and Vonnegut’s theory that stories may be graphed on a good/ill fortune y-axis and that time progressed through plot be, well, plotted on the x-axis, Baker barrels on and outlines how a plot from Nelson Rockefeller’s life, EastEnders or a bag lady’s might fit in to Vonnegut’s graph, before moving on to mapping various creation myths on to Vonnegut’s graph, too and showing how they exhibit similar shapes.

There’s a bit of irritation I have with Baker’s personality here, almost as if he’s Hypersplaining – so eager to teach and to explain things to Adams, it’s as if Adam is unable to get a word in edgeways (which, this being a documentary written by Adams, is clearly by design).

We then progress through a Google Image Search-esque search based upon Baker agreeing with Adams’s supposition that seeing as all digital information is just shapes, computers can look for them pretty easily, and off we go on a trapezoid hunt before landing back at Picasso’s Guernica (and by extension, Bob Abel’s Picasso’s Guernica, of which we experience a demo).

What’s striking, looking now at what must have been so exciting about multimedia, was that it was simply a new way of packaging and providing an accessible reference to a particular subject or piece of information. While the Beethoven’s Ninth example allows us to jump around and explore different parts of the score visually, the Guernica example seems to mainly consist of a number of video interviews of interesting sources related to the material. Sure, there’s user choice here in that the editorial journey is chosen by the user rather than a tv programme editor, but perhaps that’s 25 years of cynicism leaking in at me toward the experience of, well, Yet Another Multimedia CD-ROM. Certainly, the fact that this information is now more accessible rather than not at all, is a major step forward, but there’s something distinct about the way self-contained information is presented on these CD-ROMs versus network, hyperlinked information. CD-ROMs of those types ended up feeling like particularly dense TV shows with chapter marks, and not necessarily as big a deal as fundamental interconnectivity and user-initiated publishing.

Abel finishes up by cueing up a 90 second film clip of “what you, as a user” might be able to experience if you took your own version of Guernica. And that would certainly be an interesting thought indeed, if the rights to the material on the CD enabled remixing and republishing. But as we know with the benefit of hindsight, that rarely happened – and again, we’re dealing with an edited package of audiovisual content.

Adams muses that if this type of experience is to be understood properly, a new grammar or language needs to be invented to allow people to deal with it. Baker uses examples from the language of the moving image – jump cuts, zooms, pans, etc. – and it still feels, unlikely as it is, that we are as far away from a grammar of interaction despite our best efforts to catalogue it and design for it, in chief because software is so malleable.

The Multimedia Lab then takes us through a backgrounder in Multimedia – how it’s a new medium and how they’re dealing with creating and experimenting with the grammar for it. About how multimedia allows you to “bounce around” and compare different points of view, which strikes me not necessarily as a particular aspect of multimedia itself, but again, more of hypertext. In a somewhat prophetic comment, one of the researchers mentions how an audience involved in a particular drama might want to find out more about it, and would embark on their own idiosyncratic path to find out more background information, which happily describes anyone who’s ever fallen down the rabbit hole of a JJ Abrams story.

We’re taken through a demo of Life Story, a multimedia version of a BBC dramatisation of the discovery of the structure of DNA by Watson and Crick, and again we see this interesting assumption that people would pause one task and then operate on another – watch a bit of video, then do a bit of interaction. Maybe it was because the true impact of Moore’s Law hadn’t hit home yet, but in twenty five year’s time, computing power would be so cheap and accessible that you could sit in front of your TV and have it streaming video in the background *whilst* you purposely paid attention to the second computing device in your hand. In this case, the research project of Life Story is essentially chapter marks and navigation for a bunch of linear video, which then pauses and adds context. The researcher cutely talks about each screen (ie tapping on the links section from the transcript) as a “new pamphlet of information” that would teach the viewer about DNA.

Our next demo is children playing with Amanda Goodenough’s Inigo Gets Out, “a form of interactive storytelling,” a Hypercard stack being played with where areas on the screen are hot-clickable. Inigo Gets Out is instantly recongisable as being somewhat like a graphical point-and-click adventure game, albeit, in story format, and its ultimate progeny is that of the kids iPad book with directly manipulable areas. Goodenough raises the valid point that you may well have branching storylines that converge on the same ending (which you pretty much have to do unless you’re the kind of person who has figured out how to deal with the combinatorial explosion involved in producing branching narrative) but if the branches themselves that lead to the same ending are satisfying stories in their own right, then that’s perfectly fine. This is pretty much the only (and entire) problem I have with such choice-based branching interactive stories: it’s hard enough to write one good story, never mind fifty. Or even ten.

We then skip over to Brad deGraf and Michael Wahrman talking about digital puppetry, and anything historical to do with realtime computer graphics is always going to seem amusing in hindsight. They say, though, that they had “too-realistic” human faces in their rendering and that they had to dial it back so as not to confuse people, which just makes me wonder if they hadn’t instead fallen unintentionally into the uncanny valley. Scarily, one of them suggests that the rendered faces could be truly obnoxious and your fridge could moan at you if you wandered into the kitchen and didn’t pay attention to it, and it would try to attract your attention. Though it’s possible this segment was merely baiting the Sirius Cybernetics Corporation.

And finally, we see a NASA prototype virtual reality display from AMES called Cyberiad, and one of the reasons why I thought it would be fun to take a look at Hyperland again. Because with things like Oculus Rift and Facetime echoing Tom Selleck’s You Will narration that I would tuck my baby in at night from a phone booth (close), it feels like we are finally managing to reach the grasp we sold ourselves in the early 90s. Just in a different way.

We had all of these ideas that had been brought to life in so many different, yet ham-fisted ways. We had a nascent worldwide communications network, first in ARPA and at the same time in BBSes, then, in Usenet and Telnet and Gopher, then in the web, then somewhat persistent but slow connectivity, and now persistent, fast, mobile connectivity in our pockets. It turned out that we could imagine all of those things and that now, in their mundanity, we’re able to bring to life what we were excited about twenty five years ago. How then, do we make what was new and exciting (but inaccessible) twenty five years ago, new and exciting again? Because it should be. The fact that we can communicate, practically for free, with so many people around the world, that we can publish, that we can interact, in so many different ways and in different contexts should be amazing. And, because we are human, we do it in as many ways as we are human, and in all the boring and mundane ways.

The fear, as summed up at the end of the programme, is that all this technology and virtual reality would remove people from ‘real reality’. And that was before we even had this stuff in our pockets. Adams moves the time-slider up to 2005, doffs a 2005 BBC virtual reality cap, and off we are again into a Gibsonian chrome-textured matrix.

This was a bit of a different one. Something akin to thoughts smattered in amongst a transcript. I really can’t emphasise how much of an influence Douglas Adams was upon my formative years. Hyperland occupies a special place in my memory not only because of Adams mark on it, but because it led to a future that I wanted to see and that I wanted to become real. I’ll leave for tomorrow thoughts on where, exactly, agents like Baker disappeared to.

[1] https://vimeo.com/72501076 and http://en.wikipedia.org/wiki/Hyperland
[2] http://en.wikipedia.org/wiki/BBC_Domesday_Project
[3] https://itunes.apple.com/us/app/the-orchestra/id560078788?mt=8