User:Eleanorg/thesis/draft1.1/collision

Notes: Playing Gutenberg - Collision 1, Brussels

This February I was fortunate enough to attend the first 'Collision' worksession in Brussels, organized by Pierre Huyghebaert and others of the Libre Graphics Research Unit and Constant. The meeting was an attempt to critique, and explore alternatives to, the logic of conflict resolution underlying conventional graphic design practice, and the collaborative tools it gives rise to. An introductory text summarizes, "Maps, schematics and books are complex graphic surfaces where meaningful elements fight for space in different dimensions. ...design is essentially the work of organising collisions. In a conventional design practice... tension is deflated through erasure, simplification and filtering. What would be ways to articulate collisions, instead of avoiding them?" (Huyghebaert 2013a, collision blog)

While this is a highly formalist brief, there is a clear ethical imperative behind this line of questioning. Mapnik, the graphical engine used by Open Street Maps, is given as an example: "If a streetname doesn’t fit, it disappears from the map; if there is a conflict, contributors are advised that it all comes down to a simple choice: “yours or mine”." The conflict between competing graphical elements, then, stands in for a conflict between people. Femke Snelting summarized this adversarial approach: "If it collides, erase. If it doesn't fit, take it out." (LGRU 2013a, introductory discussion) Parallels with a black-and-white discourse of "yes or no" consent are immediately apparent ("if they say yes you go ahead and do it, and if they say no you don't" (Cochrane 2013)).

The tenacious assumption behind the Collision project is that digital tools (and ways of thinking) undermine the necessity of this approach. (Footnote: see booklet '16 case studies re-imagining the practice of layout'.) "We can begin to rethink layout from scratch. So what is scratch? What is at the bottom of it? I think the bottom of it is Gutenberg", said Huyghebaert. Framing the question as one of historical progress, Huyghebaert expressed his frustration that graphic designers are still "playing Gutenberg... playing printer with little squares & rectangles. The separation that started with Gutenberg between the writing, which is really square, seems not useful anymore." (LGRU 2013a)

The analogy between lead blocks, which cannot co-exist in the same slot, and social proposals, which must be reduced to winner/loser, is a striking one. The Collision project attempts to articulate how we might get beyond this mode of thinking. And, crucially, it doesn't require that we fetishize digital media or conceptualize it as opposed to paper. The reductive box-model described by Huyghebaert started with Gutenberg; the implication is that a digital - as in hand-made - approach could be just as liberating as a digital - as in computational - one. As Ludovico notes, the opposition between (static) paper and (manipulable) digital media is a false one (see POD discussion, Ludovico PDP p.77). While paper is not analogous to screen (PDP p. 117), 'hybrid' publishing models (such as print-on-demand) mean that printed objects are increasingly (in)formed by computational processes. (Ludovico, Liberation article) This means that we can use 'digital' ways of thinking to approach design conflicts which are not, after all, inherent to the printed medium but in fact arise from social and historical assumptions ("playing Gutenberg").

The first workshop I attended during Collision 1 illustrated this point nicely. Showing us his favourite toy, a plotter printer, Gijs de Heij facilitated a battleships-style game in which two teams plot circles on the same sheet of paper, attempting to avoid 'collision' by not overlapping the other team's shapes. De Heij's custom browser interface allowed us to draw, or write text commands, in real-time, and see our designs rendered instantly by the plotter pens moving over the paper. The game was difficult - owing to the various layers of mediation between our bodies and the page. But importantly, it was also pointless. A plotter - which can go back and re-work the same surface in a fluid way impossible with a laserjet printer - lets us re-imagine the printed page as an open space where interactions can occur, rather than a grid in which lead blocks compete for their share.

The second workshop took a more digital approach to the same problem. Christoph Haag's 'Forkable' creates a (digital) space for multiple design options to co-exist. We were invited to clone a repository containing an image file (an SVG reading 'Badness and Conflict'). Each letter of this text was stored in a different layer of the SVG file, meaning it could be independently manipulated. We were invited to make and save changes to any or all layers, and add our version as a new file to the repository. Thus, multiple versions of the same file existed, with various possible permutations. Rejecting the imperative to chose one over another, this 'conflict' is resolved by a Bash script which picks layers at random from each SVG contributed, re-combining them each time the script is run. One time my version of the letter 'M' may be included; the next, yours may be.

On the final day of Collision 1, I spent the day with Femke Snelting and Pierre Marchand researching a real-world example of collision: the choice of which metadata to render on Open Street Maps (OSM). While anyone can contribute metadata to the OSM database - for example, describing a route as a 'footpath' or a building as a 'shop' - only certain tags are rendered on the official OSM website. Footpaths and shops will be labelled; but if you've tagged a place as a gay cruising area, don't expect that to show up. While Haag's 'Forkable' simply picks at random which options to include, the decision about which OSM tags to render is a complex, quasi-democratic process. There are various interest groups lobbying for their tags of choice to be rendered in the official map - presumably so they don't have to render and host their own versions. Amongst these are the proposal to tag adult shops, and the proposal to tag gay-friendly places. The latter has been proposed by Open Queer Map, a group who are already actively adding metadata to the OSM database indicating places which are gay-friendly. While this metadata is stored in the OSM database, it is not rendered on the official map, and Open Queer Map currently have to render and host their own map in order for this information to show up. There were several people present at Collision 1 who argued that, instead of 'cherry-picking' the most important tags - resulting in the current scenario of lobbying an elite of programmer-curators - OSM should be providing an interface which lets users, not OSM programmers, decide which metadata to render. We should not have to go to several different maps to get all the information we need, but be able to browse the appropriate information through a single interface.

In both the Forkable workshop, and the debates about Open Street Map, an approach emerged which offers possible answers to the question of whether indecision could be a viable design methodology. Indecision needn't mean paralysis, deadlock, or mediocrity. Instead, it could mean provisionality. In each iteration of the Forkable SVG, different people's versions will 'win'. Just as, if OSM gave users the power to chose which metadata to render, each map would still render only a limited range of this data. The difference is that that the decision is not final, and the algorithm used to arrive at the decision is made visible.

There is a danger, of course, of such an approach returning to the 'filter bubble', and serving users a customized pick of the crop. But it needn't go this way. In the Forkable repository, all iterations are visible, and it is clear that running the Bash script will simply produce one of many possible permutations. Similarly, the simple act of looking through the list of OSM tags and seeing which ones are in use allows us to see any given rendering as a provisional and partial snapshot of the database as a whole. In this approach, the 'competing' versions are all retained and on view; not discarded or hidden by the algorithms that create outputs from them. And this approach needn't rely on digital media or database queries. Other groups at Collision 1 produced paper prototypes rendering 'indecision' using techniques such as origami-like folded paper maps to allow multiple layers of information, cross-hatching to depict overlapping geographical zones, and so on.

These examples give concrete graphic form to the feminist proposal of consent-as-collaboration, in which negotiation does not consist of successful or unsuccessful proposals, but an open space in which possibilities are communicated and can co-exist.

[ how does this relate to work of art digital manipulation - query as re-introducing benjamin-like 'aura' as each query is unique? ]