Arie Altena

The Browser Is Dead; Long Live the Browser

The myth of the graphical browser is that it can be used universally. This browser and no other brings the Web home to you as it's meant to be. But the Web has gotten big and meets a variety of needs – from finding the departure times of trains to looking up academic papers and playing games. And it's beginning to dawn on the everyday user that the existence of a universal browser that communicates all these experiences in the best manner possible is a myth.

Netscape and Explorer got big once computer illiterates arrived in the Internet wonderland. They needed the gui's page metaphor in order to get comfortable with the new medium. The capabilities of the medium were expanded, illiterates became literate, and the gui of the Netscape-Explorer continuum – with its imposed norm of downloading the latest version - was, more and more, seen as limiting. The first cracks in the monopoly of the graphical browser are becoming visible. The continuum is being attacked by graphic designers who would prefer to make Flash/Shockwave sites that stand on their own; by the more and more varied uses of the Web; but also by Linux and the free software movement; browser developers; new machines like the Web telephone; the misuse of Quake as a browser; users who keep using old versions out of stubbornness or laziness; and institutes with obsolete equipment.1

1 Only a small group of users regularly has the latest software at home. Choosing to make a site that can only be viewed with 4.0 browsers means choosing a target audience. It does not include, for example, scholars and students who only have access through school. That situation will not go away: a school or university can't add a cool 32 mb of extra memory to hundreds of pcs once a year, any more than the users can download a new version just like that.

Besides, 4.0 browsers have become dinosaurs, unmanageable, too big to manoeuvre with ease. Compare them with the extremely handy Swiss army knife: knife, bottle opener, corkscrew, everything on one handle. But there is a limit to the number of accessories: there are Swiss army knives so thick, with so many integrated attachments, that they still work but are no longer handy. 4.0 browsers are such knives, with their enormous numbers of plug-ins to translate the 4.0 Web experience. Starting up takes longer and longer, it's not clear which functions are available and which aren't, and people whose computers will easily tolerate the 4.0 version will be startled in 1999 by the amount of ram needed for its successor. It seems that a technological limit has been reached - the browser must be written anew if it's to remain functioning easily. Or is it time to throw the myth of the universal graphical browser overboard?

The use of spatial metaphors to grasp the Web's virtual structure and experience is ingrained in the software and in our gaze. It seems as if we can no longer escape it: it's an inherent concept. The Web is a sea of information you surf, you navigate, it's a library you explore, a highway you drive. The Web is an information architecture that extends in three dimensions. The browser is the technological means that makes experience there possible. The browser is the means of transportation for following links in virtual space, the instrument for charting them, and the sense (and the lens) with which the user observes space. In theory every browser can give its own interpretation of the html elements, and place these in a configuration of its own. The representation is different, but the information architecture read by Netscape and Explorer remains the same.

In the structure of the Web lies the possibility to write one's own means of transportation for data travel. Because, as Webstalker's Matthew Fuller has remarked, On connecting to a url, html appears to the user's computer as a stream of data. This data could be formatted for use in any of a wide variety of configurations. As a current, given mediation by some interpretative device, it could even be used as a flowing pattern to determine the behaviour of a device completely unrelated to its purpose.' 2

2 M. Fuller A means of Mutation, see

The structure of the Web makes a variety of experiences of it possible on the software level. This can be seen as the technological expression of relativism: equipped with a different ideology, world view or religion, the human subject observes a different reality.3

3 Originally this Web architecture was set up to be read in different ways. The formatting tags were logical and structural, not physical. What the tag does with the data between depends on the interpretation of the browser. That can mean make boldface, but a more radical interpretation is possible too, like open this content in a new window. Graphical browsers brought this mutability of html interpretation under control: html was adapted to the browsers with the goal of showing the design exactly as the designer intended. That is a limiting way of handling the medium which is not essential for the structure of the medium. Rather, that structure lends itself outstandingly to a variety of unlocking methods, to the implementation of filters, to radical use and misuse.

The browser filters the experience of space and the interface determines how space is experienced. The interface mediates the illusion of a space: at the birth of the Netscape-Explorer continuum there was the page. Despite the spread of the metaphors of ocean and highway, the metaphor of the magazine page is the true building block of the graphical browser. It's a principle that is useful for fanzines, handy for the use of the Web as distribution channel, and understandable for beginning users. The problems begin with the expansion of the use of the infrastructure. As html was forced to become a formatting language, so Explorer and Netscape were forced to become overgrown monsters. They are Formula 1 racing cars built on a go-kart chassis. (And you can take them shopping, too.) The graphical browser aims to pass along all sorts of experiences with one piece of software, one means of transportation.

The means of transport determines the experience of space and the importance one attaches to the elements within it. Different spatial information is important to a bicyclist than to a driver or a pedestrian. Good maps reflect this. Hikers use maps with a scale of 1:50,000, which is so detailed as to be useless for the driver. A highway is of the same importance to a cyclist as a railway embankment: it is an obstacle that, at most, can serve as a point of orientation to the horizon. A real cyclist's map thus will not highlight it in red. Nor is a highway one of the elements that a cyclist wants to experience. He prefers to stay as far away from it as possible and its existence has a minor effect on his experience. By contrast, a small paved bike path through the woods is of utmost importance - surely for the tourist - for it is where the desired experience is to be found. In other words, the means of transportation is a filter for the experience of space.

You ride a bicycle because it's the fastest way to reach your destination, because it's more pleasant than another means of transport, or because you want a certain experience. On the www most people put up with taking the same means of transport every time - for getting train information, looking at promotional sites, random surfing, or reading the newspaper.4

4 With the availability of bigger programs we are inclined to underestimate the scope of smaller programs. Once we're used to the automobile, the distance between Amsterdam and Rotterdam looks impossible by bicycle.

Around the means of transport arises a specific organisation of the experience of space. Or we organise space by the choice of transport. This has most of all to do with speed and range. The invention of train, bicycle, automobile and aeroplane led to a different view of geography. Distances got shorter and cities could be looked at from above, but the observation of the space crossed changed too: a meadow full of dandelions turned into spotty golden stripes on green. With high enough speed, the means of transport becomes a time machine, and there are people who claim they can see from an automobile how space curves with respect to the cabin. It is most of all the development of modern art – namely futurism – in which the influence of means of transport is immediately visible. Many innovations in the representation of space in art can be directly explained via the development of transportation technology (see Stephen Kern, The Culture of Time and Space, 1880-1918, Cambridge, Massachusetts 1983).

Just as some means of transport are healthier than others, it can also be supposed that there are good and bad browsers for mental health. A good browser delivers only healthy information and doesn't bother you with superfluous, distracting manipulations. A bad browser fills your head with banners, distractions, cookies, superfluous pictures, frames with redirects, and so on. Those who find attention and concentration salutary for the mind will reject the use of the blink tag as unhealthy, which is to say someone looking for World Cup results won't get banner ads, because they are disturbing, are a superfluous assault on the senses. Thus a proposal can be made for an ergonomic browser that filters distracting and unhealthy elements, independent of Web site makers' intentions.

Fantasise about a hundred different browsers, each perfectly suited for the combination of Web and machine. A browser for academic archives, one for investments, one for on-line shopping, one for looking up sports results, one for playing games, one for combining contents, one for making associations, one for daydreaming, one for working out aggression. Or we might imagine kits for building one's own instant browsers that would perfectly agree with the experiences one desires as Web user, with built-in filters and one's choice of interface.

Alternative browsers and alternative browser misuse have been proposed by artists, programmers, architects and fanatical gamers. Some are just proposals; many are working applications. They are developments of the idea that the experience of web space can be driven not just on the design level, but also, or instead, on that of software.



Ambulator is a browser designed by B. Mueller for people who want to feed their imaginations with images, who wish to dream away on arbitrarily loading pictures and make up stories in their heads to go with them. The Lycos search engine had already made it possible to picture-zap on the Web (better than tv). Ambulator goes one step further: it's a picture browser based on a search engine. The user starts with an almost empty screen where a search command can be given. Someone who types in, for example, David Bowie is thus ordering Ambulator to search the Web for pages where the words 'David Bowie' appear. Ambulator regards html as a stream of data that can be plucked from at will, not as a message with a beginning and an end whose unity must be respected. Of all browsers Ambulator'' has the strongest built-in filter: it ignores all html except one tag. It searches pages where the search term appears for img tags and the corresponding urls. It then serves only the pictures and makes a collage of them on the black screen.

Ambulator's practical use is extremely limited. The search command 'David Bowie' does not guarantee pictures of David Bowie will be found. What will be found are images on pages that have the search terms. The images might be banners, bullets, Netscape logos or navigation buttons. (In this respect, Ambulator mercilessly shows how miserable so much Web design is.) But because Ambulator ''works in this random way, improbable combinations can also appear on the screen. And that very randomness of images sets fantasy in motion: a story, strange connections, a short daydream or the invention of a joke - at least for those who let themselves be enchanted by the images appearing and disappearing in the black hole of the screen.

Ambulator stops searching only when it ceases to find. It keeps scanning the Web and html for the desired words and the img urls. New pictures are loaded on the screen and those that have been there too long disappear. There is no experience or transmission of any information architecture. Ambulator is a flat surface with pictures that light up. The spatial metaphor does not completely disappear, but becomes irrelevant.



Now that emulators are popular as never before and there are perverse people running Virtualpc on their Macs, the time seems ripe to bring back the experience of the old Web. There is a demand for Atari emulators for playing old games – Elite has clearly not yet been surpassed – and there are dozens of emulators for old synthesisers on the market, now that original Moogs and Vibra 7000s are impossible to find. There will likewise be early adopters among Web users - nostalgics, true - who long for the old Web, when it was still pure. No commerce, no government. What they need to realise those longings is Lynx .

Naturally, they can still telnet to a provider and type Lynx at the prompt. If they're lucky, the only real text browser will start up, and they can surf from a dumb terminal, 1993-style. 5

5 The poor academics who were forced to use Lynx are dying out. Now and then you find an email in a newsgroup from someone who only has uucp access to the Web. The last of the Mohicans. Or will new groups of Lynx users form in developing countries with donations of discarded 386s and 2400-baud modems?

For two years a client-side Lynx has also existed, with which you can emulate the old Web on your computer. This updated Lynx - there are versions for pc, Mac and Powerpc - also supports frames and tables (forms were always supported), but not the img tag.

In the same way that emulators of old synthesisers clearly produce a sound that cannot be attained with standard midi software, Lynx gives something of the experience of the pre-Mosaic Web. Naturally, much is lost - but that is also true of the MiniMoog emulator as compared with elaborate midi software and the Atari emulator as compared with the Powerpc. It's precisely that loss which gives us back something we thought was gone.

But Lynx is more than an emulator (although it would not have to be to claim right of existence!): it is useful and, sometimes, extremely convenient. Lynx does text. Lynx doesn't do pictures, Java, javascript, or any shockwave, Flash or Real Audio whatsoever. (The inflexible Lynx-adept would say you have to look at Real Audio sites with real audio, compress shockwave documents, and that Java always crashes because it was written with a Microsoft compiler that generates scripts only pcs can read.) You miss anything that's an assault on the loading time.

Whoever uses Lynx will notice that, when the Web is busy and graphical browsers would simply say does not have a dns entry, Lynx brings in the desired information without faltering. And what a relief to be able to navigate with the keyboard (no rsi). No more time-wasting clicking or menu searching; just pressing a key once (z=stop, s=save). Client-side Lynx is fast.

In all its primitiveness, Lynx is also a perfect filter. Those who hate banners and don't want animation or dhtml (enough tv around already); those who don't accept cookies and are tired of refusing them; those who don't want their every wish anticipated and would rather figure things out for themselves; those who choose to live in a textual universe and put up with the fact that many sites are unusable and innavigable and others unreachable (this site was made for 4.0 browsers) - those people should use Lynx.

Lynx is also the browser for those who hate bad and non-Web-native design. The philosophy of the Web is that of hypertext: links are there to connect information and the user determines what s/he wants to refer to when. Most sites that were built from the philosophy of the Web and that take into account the shortcomings of network communication are outstandingly readable with Lynx. NASA and most universities, for example, can be reached. Sites that were made from a graphic design borrowed from paper, that make themselves in the image of tv and seek to overwhelm the user with flashing images, cannot be reached. Lynx is there for nostalgics and text nerds, a small group that is scarcely economically interesting. A negligible group. And a group that wants to be neglected. Just let them go their way; they'll take care of themselves.

The Lynx user imposes a certain limitation on him- or herself. In the same way that someone who only gets around on foot consciously limits his/her range of action, but as a trade-off gets more detail out of life, so the Lynx user limits what is depicted on his screen. But in return he gets less junk, less commercialism, less pushiness, less unwanted intimacy, more speed and a familiar old-style hypertext labyrinth. For the Lynx user the Web is an archive of linked text documents. The Lynx user does not want progress. The Lynx user wants to go back to the source and focus on what has been lost. The Lynx user gets off on purity: a monochromatic screen of alphanumeric characters.



Webstalkeris a radical browser – some call it a metabrowser – which passes the interface and the metaphors of the gui continuum by. It breaks through the prestructured context of the combination of html and graphical browser which limits experience and the possibilities for (artistic) production. Webstalker'', designed by Matthew Fuller, Simon Pope and Colin Green of i/o/d, looks like a purely artistic project that is wholly unconcerned with the practical value of the Web as it currently exists.

The starting point of Webstalker is to bring about a breakthrough in the limited character of Web experience. Webstalker takes html documents for what they are: a data stream. The way Explorer and Netscape translate this stream is, as is known, merely one of many - and one that, according to the makers, was determined by the promise of e-commerce and graphic design with its page metaphors. Webstalker is software which foils these limitations of the Web's political, cultural and artistic meaning by showing one of the possible alternatives. And it is more than a purely artistic project.

''Webstalker's best trick is that you begin with a blank screen. The user must produce windows him- or herself and assign them a function. Crawler, map, html-stream, stash, and dismantle are the five ways to approach the data stream. The user manufactures, as it were, his or her own browser.

The user opens an urlin the crawl window and the stalker starts following. It follows all that links that lead off that page until no more new ones can be found. It maps space in the map window and reproduces the data stream of the source. You can dismantle a part of the site to see the structure better and it is possible to save the text from a page so you can read off line. Like Lynx, Webstalker has a certain practical value as a browser for finding text information. It puts the user in a position to circumvent all distracting information, bumpers, and commercialism on a site without knowing a page's exact url.

The emphasis, though, is on the possibility of portraying the relationships, the links, yourself, on making visible the structure of information networks. The site map in the map window gives an impression of a site's construction and structure at a glance, and, potentially, how the site is embedded in a larger network.

Webstalker was developed out of the conviction that the presentation of information must take account of its structure - something which threatens to be forgotten on the Internet. Webstalker does not give the user a website's geography; it portrays the topology as a process. You see the structure unfolding as the stalker searches the Web. Analysts of linking behaviour on the Web can imagine a modified Webstalker as the perfect research tool, an expanded version which generates maps that can be interpreted via graph theory and provide insight into the politics and the economy of linking.

Webstalker has devoted itself to a different notion of what html data is and what it can be used for. The functionality of Lynx, for example, is in accord with the goals of the Web, but the functionality of Webstalker is on another plane: of the meta-level, analysis, or the aesthetic value of graphics (which reduces the structure of Web sites to raw material for pretty pictures). Webstalker escapes the oppressive page metaphor and manages to slip underneath representation. Thus the software itself becomes a sensorium.

Someone who uses Webstalker experiences the Web most of all as an empty universe in which star systems appear if they're asked for. Experience is not determined by hackneyed metaphors. (In thinking, it is nearly impossible to escape metaphors, which is not the same thing as applying a metaphor in design!) Within those star systems a star's structure can be requested and downloaded for further study. Webstalker is the browser for people addicted to meta-levels and analyses, for hard-core users who demand the same austerity and purity from art and browsers that logic and mathematics offer.

Webstalker and Lynx are browsers for protestant text lovers, code-eaters who, fed by code, make images in their own heads. Ambulator is the browser for people who make up stories and wish to be fed by images without being disturbed by text. For the Lynx user the Web is what it always was: a library, an architecture of text and links - without gui - and a hypertext in optima forma. For the Webstalker user the Web is an empty universe where star systems appear. And for those who use Ambulator'', it is an accumulation of unrelated images, a reason to start imagining.

The rationale behind alternative browsers and the use of obsolete ones is to treat the Net as a space for reinvention. The Web is not a space that can only be made visible, legible and navigable with one kind of software. The advantage of the Net over reality is that the Net is only a stream of data which can be made visible, legible and navigable in various ways. That's why it's pointless to condemn users of Netscape 1.0 or MacLynx as losers who've neglected to download the 4.0 version. Just as it would be pointless to regard the development of alternative browsers as a threat to the general access and clarity of the Web. (This would be as limiting as demanding that everyone be a libertine.) The Web, after all, has long been suited for more than just one goal, and it is not controlled by commerce. And as in everyday life instruments are used in ways other than those they were made for (a match to stop a timed switch, a bicycle as a ladder to climb over a fence), software and infrastructure can also be used in inappropriate ways. Zapping with search engines. Surfing with Webstalker. That is not some kind of subversive act of sabotage; it is one of the possibilities at hand: a self-imposed limitation, the wish for a specific experience, a means of reaching a goal or a greater user-friendliness with respect to the desired goal.6

6 Propagating the use of Lynx is not just a case of an old intellectual's nostalgia for the days when the choice was between Lynx and Mosaic (for the happy few with a ppp connection) and the Web was the domain of academics and hackers - no government, no commerce. It is part of a proposal for dealing with the medium in multiple ways, not only on a content level but also on a software level.

Consider the Web as one enormous database. The browser as a lens that makes a data stream visible and interpretable. Let us stop optimalising sites for as many users as possible. Optimalise the site for a specific browser that best conveys the experience of the site, and leave it to the user whether to experience the site in the optimal manner or in another of his or her choice. This is not subversion aimed at resisting the commercialisation of the Net, or at trying to undermine its economic and financial importance or contain its socio-political importance. It is an expansion of the possible experiences. For why shouldn't we develop optimal and safe shopping browsers? Or, like the browsers that bring in perfect Java and Flash, a small browser for Real Audio. Let Netscape and Explorer do what they were made for – gui, html for the page metaphor, multimedia as a mosaic of free-standing elements – so that pessimistic intellectuals can stop cursing at the graphical browser's putative monopoly on Web experience.

translation laura martz