sign in sign up
<<
Home , Andries van Dam, History of hypertext, Hypertext Editing System, NLS (computer system), Project Xanadu, Ted Nelson

9781473980051_OFC.indd 3 12/03/2018 10:45

BK-SAGE-BRUGGER_MILLIGAN-180264-FM.indd 3 01/11/18 4:21 PM Contents

List of Figures and Box ix List of Tables xiv Notes on the Editors and Contributors xv Foreword xxv Steve Jones Introduction xxviii

Part I the Web and Historiography 1

1. Historiography and the Web 3 Ian Milligan

2. Understanding the Archived Web as a Historical Source 16 Niels Brügger

3. Existing Web Archives 30 Peter Webster

4. Periodizing Web Archiving: Biographical, Event-Based, National and Autobiographical Traditions 42 Richard Rogers

Part II theoretical and Methodological Reflections 57

5. Web History in Context 59 Valérie Schafer and Benjamin G. Thierry

6. Science and Technology Studies Approaches to Web History 73 Francesca Musiani and Valérie Schafer

7. Theorizing the Uses of the Web 86 Ralph Schroeder

8. Ethical Considerations for Web Archives and Web History Research 99 Stine Lomborg

9. Collecting Primary Sources from Web Archives: A Tale of Scarcity and Abundance 112 Federico Nanni

BK-SAGE-BRUGGER_MILLIGAN-180264-FM.indd 5 01/11/18 4:21 PM vi The SAGE Handbook of Web History

10. Network Analysis for Web History 125 Michael Stevenson and Anat Ben-David

11. Quantitative Web History Methods 138 Anthony Cocciolo

12. Computational Methods for Web History 153 Anat Ben-David and Adam Amram

13. Visualizing Historical Web Data 168 Justin Joque

Part III technical and Structural Dimensions of Web History 187

14. Adding the Dimension of Time to HTTP 189 Michael L. Nelson and Herbert Van de Sompel

15. before the Web – or, What the Web Could Have Been 215 Belinda Barnet

16. A Historiography of the : Periodizing the Web through the Changing Role of the Hyperlink 227 Anne Helmond

17. How Search Shaped and Was Shaped by the Web 242 Alexander Halavais

18. Making the Web Meaningful: A History of Web Semantics 256 Lindsay Poirier

19. Browsers and Browser Wars 270 Marc Weber

20. Emergence of the Mobile Web 297 Gerard Goggin

Part IV Platforms on the Web 313

21. Wikipedia 315 Andy Famiglietti

22. A Critical Political Economy of Web Advertising History 330 Matthew Crain

23. Exploring Web Archives in the Age of Abundance: A Social History Case Study of GeoCities 344 Ian Milligan

BK-SAGE-BRUGGER_MILLIGAN-180264-FM.indd 6 01/11/18 4:21 PM Contents vii

24. Blogs 359 Ignacio Siles

25. The History of Online Social Media 372 Christina Ortner, Philip Sinner and Tanja Jadin

Part V Web History and Users, some Case Studies 385

26. Cultural Historiography of the ‘Homepage’ 387 Madhavi Mallapragada

27. Consumers, News, and a History of Change 400 Allie Kosterich and Matthew Weber

28. Historical Studies of National Web Domains 413 Niels Brügger and Ditte Laursen

29. The Origins of Electronic Literature as Net/Web Art 428 James O’Sullivan and Dene Grigar

30. Exploring the Memory of the First World War Using Web Archives: Web Graphs Seen from Different Angles 441 Valérie Beaudouin, Zeynep Pehlivan and Peter Stirling

31. A History with Web Archives, Not a History of Web Archives: A History of the British Measles–Mumps–Rubella Vaccine Crisis, 1998–2004 464 Gareth Millward

32. Religion and Web History 479 Peter Webster

33. Hearing the Past: The Sonic Web from MIDI to Music Streaming 491 Jeremy Wade Morris

34. Memes 505 Jim McGrath

35. Years of the : Vernacular Creativity before, on and after the Chinese Web 520 Gabriele de Seta

36. Cultural, Political and Technical Factors Influencing Early Web Uptake in North America and East Asia 537 Mark McLelland

37. Online Pornography 551 Susanna Paasonen

BK-SAGE-BRUGGER_MILLIGAN-180264-FM.indd 7 01/11/18 4:21 PM viii The SAGE Handbook of Web History

38. Spam 564 Finn Brunton

39. Trolls and Trolling History: From Subculture to Mainstream Practices 577 Michael Nycyk

Part VI the Roads Ahead 591

40. Web Archives and (Digital) History: A Troubled Past and a Promising Future? 593 Jane Winters

Index 607

BK-SAGE-BRUGGER_MILLIGAN-180264-FM.indd 8 01/11/18 4:21 PM 15

Hypertext before the Web – or, What the Web Could Have Been

Belinda Barnet

Historiography is always guided by specific before hypertext became synonymous with tropes; it is ‘infected by what it touches as the Web; as Ted Nelson1, who coined the the past’ Demeulenaere (2003). Much has term ‘hypertext’ among other things, put it to been written about the me in 1999, ‘People saw the Web and they over the last 20 years, and I’ve contributed to thought, “Oh, that’s hypertext, that’s how that literature; in the process I’ve been it’s meant to look”’. But hypertext is not the infected by the vision behind the early sys- Web; the Web is one particular implementa- tems from the 60s and 70s. This might be tion of hypertext. In this chapter we will use because I’ve got to know some of the inven- ’s definition of hypertext: branch- tors – I conducted 22 interviews with inven- ing and responding text read at a tors for my book on pre-web systems Memory screen (Nelson, 1992). The Web is without a Machines (Barnet, 2013) – and drank the doubt the most successful and prevalent ver- same Kool-Aid. The first system was built in sion of hypertext, but it is also an arguably counterculture-soaked 1960s California, limited one. Hypertext existed well before though, so a bit of dreaming is appropriate, the Web – the systems here were imagined along with the occasional Yoga . (and in some cases built and obsolete) well This was an era when people had grand before and Facebook. Insofar as the visions for their pre-web hypertext systems, Web is a hypertext system, it continues an when they believed that the solution to the already established line of evolution, and it world’s problems might lie in finding a way is important to understand this pre-history to to organize the mess of human knowledge: to fully understand the Web. represent its true interconnections. The first hypertext systems were deep and Now for the important bit: this story stops richly connected, and in some respects more before the Web. More accurately, it stops powerful than the Web. These early systems

BK-SAGE-BRUGGER_MILLIGAN-180264-Chp15.indd 215 10/13/18 9:58 AM 216 The SAGE Handbook of Web History

were not, however, connected to hundreds of the human, to ‘boost our capacity to deal with millions of other users. You could not reach complexity’ as a species (Engelbart, 1999). out through FRESS and read a page hosted in Thailand or Libya. The early systems worked Here’s a human… He’s got all these capabilities on their own set of documents in their own within his skin we can make use of, a lot of mental capabilities we know of, and some of it he’s even unique environments. Although Nelson cer- conscious of. Those are marvellous machines there – tainly envisioned that Xanadu would have motor machinery to actuate the outside world, and the domestic penetration of the Web, and sensor and perceptual machinery to get the idea of NLS had nifty collaborative tools and chalk- what’s going on… (Engelbart, 1998: 213) passing protocols, none of the early ‘built’ sys- tems we look at either briefly or in depth in For Engelbart, the most important element of this chapter – NLS, HES and FRESS – was the ‘human system’ was language; language designed to accommodate literally billions of is a powerful machine. Engelbart would seek users. That’s something only the Web can do. to harness its nonlinear relationships with As Jay Bolter put it in our interview, a computer system and externalize its ‘net- worked’ structure. What the did was two things. Engelbart still remembers reading about One is that it compromised as it were on the ’s (hypothetical ‘vision’ of hypertext. It said, ‘This is the kind of linkage it’s always going to be, it’s always going to proto-hypertext system that Bush described work in this way’, [but] more importantly it said in his 1945 Monthly article ‘As that the really interesting things happen when We May Think’), and the moment when he your links can cross from one computer to became ‘infected’ with the idea of building another… So global hypertext – which is what the a means to extend and navigate this great Web is – turned out to be the way that you could really engage, well, ultimately hundreds of millions pool of human knowledge. He was a military of users. (Bolter, 2011) radar technician out in the Philippines when he first picked up a reprint of Bush’s article, The goal of this chapter is to explore the in the summer of 1945, and wandered into a visions of the early hypertext pioneers, start- Red Cross library that was built up on stilts to ing over 60 years ago, and in the process, to read it. News of Hiroshima was devastatingly broaden our conception of what hypertext fresh; Engelbart was 20 years old and deep could be. The chapter begins by exploring in thought (Barnet, 2013). He was wondering NLS, then Xanadu and on to HES. how he could ‘maximize [his] contribution to humankind’ as an engineer (Engelbart, 1999). In this article Bush looked towards the Doug Engelbart’s oN-Line postwar world as an engineer and predicted System (NLS) an exponential increase in human knowledge, especially scientific and technological knowl- Dr Douglas Carl Engelbart, who died in edge. How are we to keep track of it all? How 2013, was a softly spoken man. His voice are we to prevent great ideas from being lost? was low yet persuasive, as though ‘his words Some ideas are like seeds. ‘Or viruses. If they have been attenuated by layers of medita- are in the air at the right time, they will infect tion’, his friend Nilo Lindgren wrote in 1971 exactly those people who are most suscepti- (Rheingold, 2000: 178). I struggled to hear ble to putting their lives in the idea’s service’ him in our interview, being deaf myself, but (Rheingold, 2000: 176). Although he didn’t that didn’t matter; he had been describing the think about the article again for many years, same vision in great detail to journalists, his- the ideas in it infected Engelbart. torians and engineers for over 60 years. But five years after he had read ‘As We Engelbart wanted to improve the model of May Think’ in the Philippines, Engelbart

BK-SAGE-BRUGGER_MILLIGAN-180264-Chp15.indd 216 10/13/18 9:58 AM Hypertext Before the Web – or, What the Web Could Have Been 217

claims, ‘I formulated [my] goal on…human their information out to these technicians on intellectual effectiveness’ (Engelbart, 1962a: punch cards and printouts. The idea that a 236). These three ‘flashes’ were to become screen might be attached to a computer, and the framework he worked from for the rest that humans might interact directly via this of his career: surface, was far left field. The computer, screen and mouse would 1 FLASH-1: The difficulty of mankind’s problems become Engelbart’s parallel to Memex’s was increasing at a greater rate than our ability microfilm storage desk, tablet display and to cope. stylus. With these technologies, Engelbart 2 FLASH-2: Boosting mankind’s ability to cope with would ‘update’ an image of potentiality from complex, urgent problems would be an attractive a different era and bring it to digital comput- candidate as an arena in which a young person might ‘make the most difference’. ing. But his ideas did not garner the peer sup- 3 FLASH-3: Ahah – graphic vision surges forth of port Engelbart was seeking. me sitting at a large CRT console, working in After I’d given a talk at Stanford, [three angry ways that are rapidly evolving in front of my eyes guys] got me later outside at a table. They said, ‘All (beginning from memories of the radar-screen you’re talking about is .’ I said consoles I used to service). (Engelbart, 1988: 189) no. They said, ‘YES, it is, we’re professionals and we know, so we’re telling you, you don’t know This vision of a radar-screen console attached enough so stay out of it, ‘cause goddamit, you’re to a computer is important. Engelbart trans- bolloxing it all up. You’re in engineering, not infor- mation retrieval.’ (Engelbart, 1999) ferred this technology from the radars he was servicing in the Philippines to comput- , in large part, were still seen as ers as he had learnt about them in engineering number crunchers, and computer engineers school (CRT phosphors came into common had no business talking about the human use around WWII). This image of a human beings who used these machines. Fortunately, sitting at a screen is the image from which one of the few people who had the discipli- all future work would depart. By the time nary background to be able to understand Engelbart started work on NLS ten years the new conceptual framework was moving later, progress had been made on ‘presenting through the ranks at the Advanced Research computer-stored information to the human… Projects Agency (ARPA). This man was by which a cathode-ray-tube (of which the tel- J.C.R. Licklider, a psychologist from MIT. evision picture tube is a familiar example) can ‘The hope is that, in not too many years, be made to present symbols on their screens human brains and computing machines will of quite good brightness, clarity, and with be coupled together very tightly, and the considerable freedom as to the of the resulting partnership will think as no human symbol’ (Engelbart, 1962b). In 1951, how- brain has ever thought’, wrote Licklider ever, Engelbart had to mentally extrapolate in a 1960 paper called ‘Man-Computer from radar screens. As he told me in 1999, Symbiosis’ (Licklider, 1960: 131). Licklider began financing projects that I put together what I knew about computers and developed thought-amplifying technologies. what I knew about radar circuitry etc. to picture working interactively, and it just grew from there’ ARPA support began in 1963, ‘at varying (Engelbart, 1999). levels – during 1965, about eighty thousand dollars’ (Bardini, 2000: 23). But as Bardini This was a radical idea in the 1950s. At points out (in his book and in a personal that time, computers were large electronic communication), it was Bob Taylor, initially devices stored in air-conditioned rooms, at working as a psychologist at NASA, who many degrees of separation from the ‘user’. mustered the strongest support for the project. They were attended to by technicians and fed Initially he put in 85,000 dollars from NASA

BK-SAGE-BRUGGER_MILLIGAN-180264-Chp15.indd 217 10/13/18 9:58 AM 218 The SAGE Handbook of Web History

mid 1964 to mid 1965 (Bardini, 2000: 23). anywhere (‘unless there was a specific reason Then when he moved from NASA to ARPA not to do so’ (Duvall, 2011)). This approach in 1964 he told Engelbart that IPTO ‘was to information was entirely new – an approach prepared to contribute a million dollars ini- that assumed from the outset that connectiv- tially to provide one of the new time-sharing ity is important, that the relationship between systems, and about half a million dollars a and among ideas was just as important as the year to support the augmentation research’ unit of information (or ‘statement’) itself. As (Rheingold, 2000: 86). That is the equivalent Duvall remembers, of around 14 million in today’s dollars. Around this time Engelbart asked a bright The thing that I would say distinguished NLS from young SRI engineer named Bill English if he’d a lot of other development projects was that it was present a paper for him. English said yes, and sort of the first – I’m not sure what the right word is – ‘holistic’ is almost a word that comes to mind – joined Engelbart’s project shortly thereafter. project that tried to use computers to deal with As English told me in 2011, ‘I saw what he documents in a two-dimensional fashion rather was doing and I was interested in it. So I joined than in a one-dimensional fashion. (Duvall, 2011) him and that’s how it all began’ (English, 2011). He became Engelbart’s chief engineer Content insertion and navigation involved in 1964, and began work on some of the basic four basic commands: Insert, Delete, Move ideas we’ll explore in the next section. and Copy. The mouse served as a pointer to Bill Duvall joined SRI in 1966, but didn’t indicate where content was to be inserted start work on the NLS project until 1968. He or deleted in existing text. Most important, asked to be moved onto Engelbart’s team, but however, the Link function allowed cross- first he had to meet with the head of engineer- referencing to another statement – and the ing at SRI. user could define cross-references at any level in the hierarchy. Links were character [He] was a very traditional engineer. I was a strings within a statement ‘indicating a rela- 24-year-old kid, or 23 years old, and he had this big corner office that had bookshelves to the ceil- tionship’ (Engelbart, 1999) to another state- ing and the little ladder that goes around, and a ment, and they could be made in the same file big desk. He sat on the other side of his desk, and or between different files. I sat in a chair and he looked at me, and he said, The NLS design changed as it was cre- ‘You don’t really think what they’re doing up there ated, evolving around the technical activities is science, do you?’ I think that reflected a lot of the official attitude towards what Doug was of the project team itself as they documented doing. (Duvall, 2011) their efforts. In the process of designing the software, the team generated a number The freshly outfitted laboratory, the Augmen­ of technical reports, source code versions, tation Research Center (ARC), began its communications, release notes, problems work in 1965. It started with a series of and associated solutions. These were linked experiments focused on the way people select together and tracked by date and version. objects on a computer screen. In the context Consequently, between 1969 and 1971 of screen-based interactivity, Engelbart and (Bardini, 2000) NLS was changed to include English’s ‘mouse’ consistently beat other an electronic filing arrangement that served devices for fast, accurate selection in a series as a linked archive of the development team’s of controlled tests. Yet it took over 20 years efforts. This eventually cross-referenced to enter the commercial market, a time period over 100,000 items (Engelbart, 1988). It was that Engelbart considers strikingly long called the software Journal, the most explicit (Engelbart, 1988: 196). model of a hypertextual environment with At the heart of NLS was a basic philosophy embedded associative links to surface in the that you should be able to link to anything, digital environment in the 1960s.

BK-SAGE-BRUGGER_MILLIGAN-180264-Chp15.indd 218 10/13/18 9:58 AM Hypertext Before the Web – or, What the Web Could Have Been 219

As mentioned previously, the entire NLS back to the group at SRI via a temporary system was set up for perfect storage and microwave antenna. The setup was very recall. Everything was tracked and identifia- expensive, and although no special system ble. Every object in the document was intrin- capabilities were employed (NLS was run sically addressable, and most important, these just as it was used back at the lab), the organ- addresses never disappeared (Engelbart, izational and presentational machinery used 1999). They were permanent, attached to the almost all the remaining research funds for object itself, which meant that they followed the year. the object wherever it was stored, so links Engelbart sat up on the stage beneath the could be made at any stage to any object in projection screen, a mouse in one hand and the system. As Duvall remembers, his other hand playing a special one-handed keyset. He manipulated the audience’s atten- [In NLS] we had it that every object in the docu- tion by controlling their view of the informa- ment was intrinsically addressable, right from the tion being explored; he drilled down through word ‘go’. It didn’t matter what date a document’s the data structure and presented it in multiple development was, you could give somebody a link right into anything, so you could actually have different views, each piece connected to the things that point right to a character or a word or last by a link. The screen was divided into something. (Duvall, 2011) neat windows containing explanatory text or graphics about NLS, and also about the The NLS team called this a ‘frozen state’ presentation itself in the hypertext Journal. It addressing scheme (Engelbart, 1997), which was dubbed ‘’ by is in contrast to the World Wide Web, where . A video of this demo is the finest level of intrinsic addressability is now available on the Web. the URL (Universal Resource Locator, a This was the first public appearance of character string that identifies an Internet the mouse and the first public appearance resource, invented by Tim Berners-Lee in of hypertext, screen splitting, computer- 19942). Unlike the NLS object-specific supported associative linking, computer con- address, the URL is simply a location on a ferencing and a mixed text/graphics interface. server; it is not attached to the object itself. It proceeded without a hitch and received a That said, NLS was working in its own lit- standing ovation (Rheingold, 2000). tle environment, on its own documents – not As a result of the NLS demo, many of the with billions of users. technologies from NLS also By 1968 NLS had matured into a massive migrated into computing over the next few database and set of paths through this data- years, the mouse in particular. Ted Nelson base, the first digital hypertext system. It eventually incorporated the mouse into his was time to take NLS out of the Petri dish Xanadu design (Barnet, 2013). NLS also and set it to work in front of the engineering allowed consideration of the modern windows- community. Engelbart took an immense risk icon-menu-pointing-device (WIMP) interface. and applied for a special session at the ACM/ Engelbart was working from a vision he IEEE-CS Fall Joint Computer Conference in had as a young engineer. This vision changed San Francisco in . ‘The nice the world. Technical visions have no essence; people at ARPA and NASA, who were fund- there is no transcendent design, no Platonic ing us, effectively had to say “Don’t tell me!”, form we are striving towards. There is, how- because if this had flopped, we would have ever, a recurrent dream – an elusive ‘blessed gotten in trouble’ (Engelbart, 1988: 203). break’ – and this dream is an ancient one. It The conference was set up using a video comes from long-standing cultural desires projector pointing at a huge, 20-foot screen and anxieties about the ephemerality of that was linked to a host computer and piped human memory and knowledge.

BK-SAGE-BRUGGER_MILLIGAN-180264-Chp15.indd 219 10/13/18 9:58 AM 220 The SAGE Handbook of Web History

The dream is to create a perfect archive claimed to have woken from a laudanum- for human knowledge, a machine that might laced reverie with ‘two or three hundred’ ‘extend, through replication, human mental lines of poetry in his head. He had noted experience’ (Nyce and Kahn, 1991: 124) and down but a few lines when he was interrupted capture the interconnected structure of knowl- by a visitor, and when he returned to his work edge itself. Most important, this would be a later he found that the memories had blurred machine that we can control, whose workings irretrievably. His mythical landscape, this are transparent to us and whose trails do not vision of Xanadu, had passed away ‘like the fade. Which brings us to the next system – images on the surface of a stream into which my favorite system and recurring vision: the a stone had been cast’ (Coleridge, cited in Magical Place of Literary Memory: Xanadu®. Nelson, 1987: 142). One of Nelson’s most cherished memories is based on a vision of gently moving water. The Magical Place of Literary When he was about four or five he trailed his hand in the water as his grandfather rowed a Memory: Xanadu boat. He studied the different ‘places’ in the water as they passed through his fingers, the What I thought would be called Xanadu® is called the World Wide Web and works differently, but ‘places that at one instant were next to each has the same penetration. (Nelson, 1999) other, then separated as my finger passed. They rejoined, but no longer in the same way’ It was a vision in a dream. A computer filing (Nelson, 2010: 35). These connections were system that would store and deliver the great infinite in number and in complexity, and body of human literature, in all its historical they changed as the water moved. It was a versions and with all its messy interconnec- religious experience, a vision that has stayed tions, acknowledging authorship, ownership, with him for 70 years. At that moment he quotation and linkage. Like the Web, but started thinking about what he calls ‘profuse much better: no links would ever be broken, connection’, the interconnections that perme- no documents would ever be lost, and copy- ate life and thought: How can one manage right and ownership would be scrupulously all the changing relationships? How can one preserved. The Magical Place of Literary represent profuse connection? Xanadu was Memory: Xanadu. In this place, users would proposed as a vast digital network to house be able to mark and annotate any document, this corpus of ideas and evidential materials, see and intercompare versions of documents facilitated by a special linking system. side by side, follow visible from The story of Xanadu is the greatest image both ends (‘two-way links’) and reuse con- of potentiality in the evolution of hypertext. tent pieces that stay connected to their origi- Nelson invented a new vocabulary to describe nal source document. There would be his vision, much of which has become inte- multiple ways to view all this on a computer grated into contemporary theory screen, but the canonical view would be side- and practice – for instance, the words ‘hyper- by-side parallel strips with visible connec- text’ and ‘hypermedia’. As he told me in 1999, tions. Just imagine. This vision – which is ‘I think I’ve put more words in the dictionary older than the Web, and aspects of it are older than Lewis Carroll. Every significant change than personal computing – belongs to hyper- in an idea means a new term’. Nelson came text pioneer Theodor Holm Nelson, who up with many significant changes, and con- dubbed the in October 1966.3 sequently many new terms, some of which I The name comes from the famous poem discuss below. He also recruited or inspired by Samuel Taylor Coleridge, ‘Kubla Khan’. some of the most visionary programmers and In his tale of the poem’s origin, Coleridge developers in the history of computing, many

BK-SAGE-BRUGGER_MILLIGAN-180264-Chp15.indd 220 10/13/18 9:58 AM Hypertext Before the Web – or, What the Web Could Have Been 221

of whom went on to develop the first hyper- ‘and my world exploded’ (Nelson, 2010: 99). text products (although this doesn’t impress He understood immediately that computers him: ‘the problem with inspiring people is were all-purpose machines that could be put that they then try to credit you with things in the service of information handling. Like you don’t like’ (Nelson, 2010). Engelbart, he did not believe that computers Much has been written about Project were mathematical tools for engineers. He Xanadu over the years (the ones I cite saw a solution to his problem of informa- throughout this chapter are Landow, 1992; tion handling, and he also saw a future where Rheingold, 2000). Nelson himself doesn’t paper might be eliminated. ‘The prison of have the time to keep up with it all – and even paper, enforcing sequence and rectangularity, when he does, as he put it to me in 2010, had been the enemy of authors and editors ‘anything people write about me will be for thousands of years; now at last we could insufficiently praising, and so it’s very hard break free’ (Nelson, 2010: 120). to read it’ (his comments on my own work One of the first ideas was based on his were prolific). The remarkable thing about own ‘terrible problem’ keeping notes on file Xanadu is that, despite countless setbacks, cards. The problem was that his cards really it refuses to die. Its logo is, appropriately needed to be in several different places at enough, the Eternal Flaming X. Paisley and once; new projects were built on earlier ideas, Butler (cited in Smith, 1991: 262) have noted new documents were built on earlier ideas, so that ‘scientists and technologists are guided these items should be reusable by reference. by “images of potentiality” – the untested Perhaps the computer might solve that prob- theories, unanswered questions and unbuilt lem; it could link them together. This idea devices that they view as their agenda for five would later prove important in the design years, ten years, and longer’. Often accused of Xanadu. of hand waving and lucid dreaming, Nelson’s Xanadu has nonetheless become the most Many of my file cards belonged in several places at important vision in the history of computing. once – several different sequences or projects. Each card – call it now an entry or an item – should Nelson wears a lanyard across his neck and be stored only once. Then each project or sequence shoulder with three pens attached to it (when would be a list of those items. (Nelson, 2010: 103) I first met him in 1999 it had sticky notes, at another point a stapler – the ‘system is evolv- Imagine if the same item could appear in ing’ as he put it in a personal communication, multiple places. You could connect each item 2012). He has been wearing it since 1997. The to its original by an addressing method and belt is filled with tools to connect things with, retrieve them on a computer screen. Because tools to deal with a world of paper. Like Bush, this would be reuse by reference rather than Nelson is painfully aware that ideas are eas- by copying, you could trace each item back ily lost in conventional indexing systems, that to its original source. This idea – that you they are disconnected from each other, and should be able to see all the contexts of reuse, that ‘serious writing or research’ demands and that you should be able to trace items connecting ideas together. Frustrated by the back to their original source – would ‘drive lack of a global, real-world system that might my work to the edge of madness’ (Nelson, do this for him, and ‘outraged’ by the con- 2010: 104). It would later become the kernel fines of paper (Nelson, 1998: 1–2), he feels of Nelson’s most innovative idea: transclu- the need to do this manually. sion. , and also the ability to vis- Nelson’s struggle against the paradigm ually compare prior or alternative versions of of paper led him to design an alterna- the same document on-screen (‘intercompar- tive. In 1960, at , he ison’) were integral to the design of Xanadu. took a computer course for the humanities, Over the next 40 years, Nelson would hone

BK-SAGE-BRUGGER_MILLIGAN-180264-Chp15.indd 221 10/13/18 9:58 AM 222 The SAGE Handbook of Web History

these ideas and experiment with them in vari- of elements (transclusion). He thought about ous incarnations. the architecture of the system and decided In 1960 Nelson announced his term pro- to have sequences of information that could ject: a writing system for the IBM 7090, be linked sideways. As with his first design, the only computer at Harvard at the time, this would all occur on a computer screen, stored in a big, air-conditioned room at the visually, in real time. He called this system Smithsonian Observatory. In the 1960s ‘Zippered Lists’. computers were ‘possessed only by huge Zippered Lists permitted linking between organizations to be used for corporate tasks documents: like the teeth in a zipper, items in or intricate scientific calculations’ (Nelson, one sequence could become part of another 1965: 135). The idea that expensive process- (‘EXCEPT’, Nelson wrote in response to this ing time might be wasted on writing, of all chapter, ‘the two sides of the zipper don’t have things, was deemed crazy by the engineer- to be in the same order’). Versions of a docu- ing community. Nelson ignored this. He pro- ment could be intercompared; an item could posed a machine-language program to store be an important heading in one sequence and documents in the computer, change them on- a trivial point in another, and all items could screen with various editorial operations and be written or retrieved in a nonsequential print them out. But this was no mere word fashion. Links could be made between large processor (which in any case didn’t exist at sections, small sections or single paragraphs. the time); Nelson envisioned the user would Writers could trace the evolution of an idea. be able to compare alternative and prior ver- Crucially, the design also got him pub- sions of the same document on-screen. lished. Nelson’s first paper explaining the The second part of Nelson’s idea took term ‘hypertext’ was presented at the ACM shape in the early 1960s, when there was ‘a 20th National Conference in August 1965. It lot of talk around Cambridge [Massachusetts, was not the first time that the word ‘hyper- where Harvard is located] about Computer- text’ had appeared in print, though. That Assisted Instruction, for which there was a was in an invitation to a talk Nelson gave at lot of money’ (Nelson, 1992: 1/26). It was Vassar College, ‘Computers, Creativity and not so much a design at this point, Nelson the Nature of the Written Word’, on 5 January stressed in response to my request for paper, 1965. (Actually, it’s the word ‘’. A ‘it was an idea that may have been on only copy of this invitation appears in Possiplex.) one file card’ (Nelson, 2010). At this time he One of the first people who thought they conceived of what he called ‘the thousand might try to build part of Nelson’s design was theories program’, an explorable computer- Andries van Dam (I stress ‘part’ here because assisted instruction program that would allow van Dam had ideas of his own that he wanted the user to study different theories and sub- to explore at the same time, such as print text jects by taking different trajectories through editing). We explore this in the next section. a network of information. Van Dam would be the first person to This led to another idea, which Nelson attempt to build part of Nelson’s vision. drafted as an academic paper while teach- ing sociology at Vassar College in 1965. He wants to stress at this point that there was no single eureka moment as ‘the ever-changing Seeing and Making Connections: designs had been swirling in my head for five HES and FRESS years’ (Nelson, 2012). The concept was clear enough, however, to put it down on paper. We will now trace the development of two This revised design combined two key ideas: important hypertext systems built at Brown: side-by-side intercomparison and the reuse the (HES),

BK-SAGE-BRUGGER_MILLIGAN-180264-Chp15.indd 222 10/13/18 9:59 AM Hypertext Before the Web – or, What the Web Could Have Been 223

co-designed by Ted Nelson and van Dam and talk about work – i.e. software, algorithms, developed by van Dam’s students, and the things that are concrete’, recalled van Dam File Retrieval and Editing System (FRESS), (1999). What Nelson did have was a vision of designed by van Dam and his students. what hypertext should look like, and an has played a major role in infectious enthusiasm for the idea. Nelson the development of hypertext systems and had in mind an entirely new genre for litera- humanities computing since 1967, due in no ture, and he had a new word to describe this small part to van Dam’s work in this area, vision: hypertext. and the Institute for Research in Information and Scholarship (IRIS) he helped establish Nelson’s vision seduced me. I really loved his way there with William Shipp and Norman of thinking about writing, editing and annotating as a scholarly activity, and putting tools together Meyrowitz in 1983. to support that… He talked me into working on Ted Nelson was co-designer of HES, and the world’s first hypertext system and that sounded his ideas about hypertext inspired the HES cool. (van Dam, 1999) project in the first place; van Dam credited Nelson for this contribution both in our inter- Van Dam gathered a team together at Brown views (1999; 2011) and in his public talks and began work later that year, with the and published work. Nelson still feels, how- objective of trying out this hypertext concept. ever, that he has been ‘given no more credit He stressed in his communications to me that than his [van Dam’s] undergraduate students’ the idea was never to ‘realize’ Xanadu in its (Nelson, 2012). Unfortunately, Nelson fell entirety. The intention was much smaller and out with the HES team during its design and more circumspect: to ‘implement a part of implementation in 1967, and he has stated in his vision. We were not able to implement or several interviews with the author that he was even understand the breadth and scope of that unhappy with the result. He is bitter about [larger] vision’ (van Dam, 2011). Nelson, the experience, largely because he feels his however, was under the impression that his vision was sidelined in favor of print text designs would be honored. editing (Nelson, 1999; 2011). We explore this Nelson was initially very excited. He went in more detail shortly. up there ‘at his own expense’ (van Dam, Along with Engelbart’s landmark NLS 1988: 87) to consult in the development of system, HES and FRESS constitute the first HES, but found the experience frustrating. generation of hypertext systems. These are not As observed previously, van Dam and his dusty old antiques from the dawn of computing team wanted to explore the hypertext con- science, however; in many respects, FRESS cept, but they also had their own plans for by 1968 was more interactive than present-day print text editing (which Nelson strenuously HTML. As working prototype systems, HES opposed). The team set out to design a dual- and FRESS had technical chutzpah. purpose system for authoring, editing and printing documents such as papers, propos- als and course notes, which could also be The Design of HES used to browse and query written materials nonsequentially. From Nelson’s perspective, Van Dam bumped into Nelson at the 1967 this was a compromise on his design: it was Spring Joint Computer Conference in Atlantic simulating paper, and hypertext was a mere City. Passionate and eloquent, Nelson told footnote to the system. van Dam about what he’d been doing since This sentiment is understandable; the he left Swarthmore: hypertext. ‘He had noth- reader knows by now that Nelson feels his ing to show for this idea, no prototypes or vision was sidelined. But the HES team were work in the sense that computer scientists trying to convince the world that the whole

BK-SAGE-BRUGGER_MILLIGAN-180264-Chp15.indd 223 10/13/18 9:59 AM 224 The SAGE Handbook of Web History

concept of handling text on computers was an asterisk) to an entrance point in another, not a waste of time and processing power. The or the same, area. world knew text handling as a paper-based The HES team used Ted Nelson’s concept thing. ‘Not only were we selling hypertext, of a hypertext link (though from Nelson’s but at the same time document processing, perspective they ‘flattened’ this by making interaction. Many people were still comput- the jumps one-way), as Doug Engelbart was ing with cards’, recalls van Dam (1999). incorporating the same idea into NLS indepen- HES was set up on an IBM 360/50 with dently, unbeknownst to van Dam, who wishes a 2250 display, and ran in a 128k partition he was aware of this work. ‘I hadn’t heard of of the that controlled the Engelbart. I hadn’t heard of Bush and Memex. 512k of main memory available (there was That came quite a bit later’, van Dam recalls a complete times-haring system operating (1999). Links were intended to be optional in another partition). The user sat facing a paths within a body of text – from one place 12-inch-by-12-inch screen, browsing through to another. They represented a relationship portions of arbitrarily sized texts. Original between two ideas or points: an intuitive con- text was entered directly via a keyboard, and cept. Branches were inserted at decision points the system itself was controlled by pressing to allow users to choose ‘next places’ to go. function keys and by pointing at the text with In early 1968 HES did the rounds of a num- a light pen or with the keyboard (Carmody ber of large customers for IBM equipment, for et al., 1969: 4). The activities of the user cor- example, Time/Life and The New York Times. responded directly to the operations normally All these customers based their business on performed upon text by writers and editors. the printed word, but HES was too far out for The user was able to manipulate pieces of them. Writing was not something you did at text as though they were physical items: cor- a computer screen. They had seen programs recting, cutting, pasting, copying, moving that set type, and maybe some programs for and filing drafts. managing advertisements, but the concept of The HES team did not wish to store text in sitting in front of a computer and writing or numerical pages or divisions known to users, navigating text was foreign to them. except as they might deliberately divide text, create links or number headings. Rather The best I ever got was from people like Time-Life than filing by page number or formal code and the New York Times who said this is terrific technology, but we’re not going to get journalists name, HES stored text as arbitrary-length typing on computer keyboards for the foreseeable fragments or ‘strings’ and allowed for edits future. (van Dam, 1999) with arbitrary-length scope (for example, insert, delete, move, copy). This approach As we now know, however, in less than a differed from NLS, which imposed a hier- decade journalists (and executives) would be archical tree structure of fixed-length lines typing on computer keyboards. or statements upon all content; Engelbart In late 1968 van Dam finally met Doug used 4,000-character limits on his state- Engelbart and attended a demonstration of ments to create a tighter, more controlled NLS at the Fall Joint Computer Conference. environment. HES was deliberately made As we explored, this presentation was a land- to embody a freewheeling character, as non- mark in the history of computing, and the audi- structured as possible. ence, comprising several thousand engineers The system itself comprised text ‘areas’ and scientists, witnessed innovations such as that were of any length, expanding and con- the use of hypertext, the computer ‘mouse’ tracting automatically to accommodate mate- and screen and telecollaboration on shared rial. These areas were connected in two ways: files via videoconferencing for the first time. by links and by branches. A link went from a For van Dam this system set another, point of departure in one area (signified by and entirely different, technical precedent.

BK-SAGE-BRUGGER_MILLIGAN-180264-Chp15.indd 224 10/13/18 9:59 AM Hypertext Before the Web – or, What the Web Could Have Been 225

The line or context editor was old technol- tangled data that surrounds us. For the world ogy. NLS was the prototype for creating, grows more and more complex every day, navigating and storing information behind and the information we are expected to keep a tube and for having a multiuser, multiter- track of proliferates at every click. How are minal, cost-effective system. He went on to we to keep track of it all? design the File Retrieval and Editing System The problem Nelson, Engelbart and van (FRESS) at Brown with his team. Dam identified is just as urgent today. The Nelson is adamant that the legacy of HES Web has *not* solved this problem for us – is modern word processing, and that it also arguably it has highlighted it, in razor-sharp led to today’s . text and 16.7 million colors. The Web is without doubt a world-changing and world- The design of HES became the design of FRESS… opening technology; as van Dam put it to then Intermedia, then imitated by Notecards and then by the World Wide Web. (Nelson, 2012) me in 1999, ‘the fact that I can reach out and touch stuff in Ethiopia, as it were, is still a Van Dam, for his part, thinks this is over- surprise to me’ (van Dam, 1999). Unlike any reaching. ‘I think Nelson, when he feels that of the hypertext systems we have looked at the “bad example” that HES set had reverber- here, the Web stretches between countries ations in the bad design of the Web or brows- and engages literally billions of people. But ers, is giving our humble little effort an order there are things we could improve on. of magnitude more credit than it deserves’ Imagine a system whose trails did not fade. (van Dam, 2011). Imagine if documents and objects could be We do not have here to go into van stored permanently, with their own unique Dam’s next system (FRESS): interested read- address that never vanished, and retrieved at ers can find it in my book. I shall stop here, will. Imagine if any version of these docu- 40 years before the Web, and conclude. ments could be visually intercompared side by side, like the teeth in a zipper – and the quotes or ideas in those documents could be traced back to their original source with a click. Conclusions Imagine if we could separate the linking struc- ture from the content, and that content could I hope that this chapter has presented you with consequently be reused in a million differ- some earlier models of the hypertext concept, ent formats. Imagine if we could capture the and in the process, demonstrated that every deeply tangled structure of knowledge itself, model has its benefits and its shortcomings. but make it better, make it permanent. Hypertext is not the Web; the Web is but one particular implementation of hypertext. It’s the best we’ve come up with insofar as it actu- Notes ally works, most of the time – and it has stayed the course for 22 years. It is not, how- 1 Note that his full name is Theodor Holm Nelson. 2 An anchor tag is not an intrinsic address. ever, the only way hypertext can be done – as 3 While working at Harcourt. the systems described in this chapter show. We have also inherited a vision from these projects: a device that ‘enables associative connections that attempt to partially reflect References the ‘intricate web of trails carried by the cells of the brain”’ (Wardrip-Fruin, 2003: 35). Bardini, T. (2000) Bootstrapping: Douglas More precisely, a tool for thought – a tool Engelbart, Coevolution, and the Origins of that might both organize and ‘permanize’ (to Personal Computing. Stanford, CA: Stanford use Nelson’s term again) the mass of deeply University Press.

BK-SAGE-BRUGGER_MILLIGAN-180264-Chp15.indd 225 10/13/18 9:59 AM 226 The SAGE Handbook of Web History

Barnet, B. (2013) Memory Machines: The Evo- Nyce, J. and Kahn, P. (eds) (1991) From lution of Hypertext. London & New York, NY: Memex to Hypertext: Vannevar Bush and Anthem Press. the Mind’s Machine. London: Academic Carmody, S., Gross, W., Nelson, T.H., Rice, D. Press. and van Dam, A. (1969) ‘A Hypertext Editing Rheingold, H. (2000) Tools for Thought: The System for the /360’. In M. Faiman and J. History and Future of Mind-Expanding Tech- Nievergelt (eds), Pertinent Concepts in Com- nology (2nd Edition). Cambridge, MA: MIT puter Graphics, 291–330. Urbana: University Press. of Press. Smith, L.C. (1991) ‘Memex as an Image of Demeulenaere, A. (2003) ‘An Uncanny Thinker: Potentiality Revisited’. In J. Nyce and P. Kahn Michel de Certeau’ in Image & Narrative, 5. (eds), From Memex to Hypertext: Vannevar Engelbart, D. (1962a) ‘Letter to Vannevar Bush Bush and the Mind’s Machine, 261–86. and Program on Human Effectiveness’. In J. London: Academic Press. Nyce and P. Kahn (eds), From Memex to Hyper- Tofts, D. (1998) Memory Trade: A Prehistory of text: Vannevar Bush and the Mind’s Machine, Cyberculture. North Ryde, NSW: Interface 1991, 235–44. London: Academic Press. Press. Engelbart, D. (1962b) ‘Augmenting Human Van Dam, A. (1988) ‘Hypertext ‘87 Keynote Intellect: A Conceptual Framework’. Report Address’ in Communications of the ACM to the Director of Information Sciences, Air Volume 31 Issue 7, July, 887–95. Force Office of Scientific Research, Menlo Wardrip-Fruin, Noah (2003) ‘Introduction: As Park, CA, Stanford Research Institute. We May Think’. In Noah Wardrip-Fruin and Engelbart, D. (1997) Interview with David Ben- Nick Montfort (eds), 2003, 35. nehum. Online: http://memex.org/meme3- Wardrip-Fruin, Noah and Montford, Nick (eds) 01. (2003) The New Media Reader. Cambridge, Engelbart, D. (1998) ‘The Augmented Knowl- MA: MIT Press edge Workshop’. In A. Goldberg (ed.), A History of Personal , 185–249. New York: ACM Press. Landow, G. (1992) Hypertext. London: Press. Personal Communications Licklider, J.C.R. (1960) ‘Man-Computer Symbi- oses’. In A. Goldberg (ed.), A History of Per- Bolter, Jay (2011) Interview with the author. sonal Workstations, 1988, 131–40. New Duvall, Bill (2011) Interview with the author York: ACM Press. (recorded via phone, Melbourne). Nelson, T. (1965) ‘A File Structure for the Com- Engelbart, Doug (1999) Interview with the plex, the Changing and the Indeterminate’. author, Baltimore. In Proceedings of the ACM 20th National English, Bill (2011) Interview with the author Conference. New York: ACM Press. (recorded via phone, Melbourne). Nelson, T. (1987) Computer Lib/Dream Nelson, Ted (1999) Interview with the author at Machines. Redmond, WA: Microsoft Press. Kyoto University Japan. Nelson, T. (1992) Literary Machines. Sausalito, Nelson, Ted (2010) Interview with the author at CA: Mindful Press. Swinburne University. Nelson, T. (1998) What’s on My Mind. Online: Nelson, Ted (2012) Comments on the manu- http://www.xanadu.net/xybrap.html script for Memory Machines. Nelson, T. (2010) Possiplex: Movies, Intellect, Cre- Van Dam, Andries (1999) Interview with the ative Control, My Computer Life and the Fight author at Brown University. for Civilization. Available at Lulu.com http:// Van Dam, Andries (2011) Interview with the www.lulu.com/spotlight/tandm [March 2012]. author via phone.

BK-SAGE-BRUGGER_MILLIGAN-180264-Chp15.indd 226 10/13/18 9:59 AM