www.nettime.org
Nettime mailing list archives

<nettime> Software Art After Programming
Richard on Tue, 19 Oct 2004 14:28:07 +0200 (CEST)


[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

<nettime> Software Art After Programming


Are people still interested in ART on this list?
Perhaps they are...


Software Art After Programming

Richard Wright, April 2004.

First published in MUTE magazine, no. 28, Autumn 2004
http://www.metamute.com/look/article.tpl?IdLanguage=3D1&IdPublication=3D1=
&NrIssue=3D28&NrSection=3D10&NrArticle=3D1397


The history of computing in arts practice is littered with the mental =
debris of its half-forgotten debates, unresolved problems and anxieties, =
and questions that have now become as obsolete as the Commodore 64s and =
VAX mainframes that accompanied them. Who can remember the art and =
technology projects of the sixties when the question of 'Can the =
computer make art?' allowed a generation of isolated computer artists to =
position themselves as a team of intrepid explorers setting out to cross =
a new continent without first waiting to find out whether it could =
support life. Under what conditions was the question ever first =
considered worthy of posing in the first place? Did the computer offer =
input into specific art issues, such as arts relation to other forms of =
scientific knowledge, to language, representation or the abandonment of =
the object? Or was it just intuitively realised that 'computer art' was =
at the forefront of a slow, inexorable computerisation of twentieth =
century society which would eventually demand access to every facet of =
human culture?

As computer hardware and the programming skills needed to operate it =
became more accessible, the question 'Can the computer make art?' was =
asked less and less often. By the beginning of the '80s artists were =
using the first personal computers to produce more varied kinds of work =
until, with all this activity growing, the question of whether art was =
possible on a computer lost all sense. There was a moment when the =
parameters of the question were redrawn, from 'Can the computer make =
art?' to 'Can a computer be an artist?', redirecting it into issues of =
simulated creativity and artificial intelligence. It was at this point =
that the first cracks of a coming schism in the community of computer =
artists became noticeable; this would go on to form the next stage in =
the debate. It seemed to a growing number of artists that as the =
complexity of software increased, so many new possibilities for the =
human artist were appearing that the prospect of deferring to a machine =
artist seemed almost indicative of a lack of imagination.

Although the computer seemed to have made its case as a machine of =
creative potential, there now emerged the question of how to efficiently =
leverage all this creativity. By the late eighties, the interactive =
interfaces and simplified menu commands of personal desktop systems that =
had helped to cause this ground swell of activity had firmly refocused =
questions on the artists themselves. Were the pre-packaged functions, =
options and parameters of the new art applications sufficient to cover =
all artistic fields of inquiry, all aesthetic nuances, all personal =
idioms? Or would it always be necessary to have recourse to the =
precision and particularities of programming languages in order to =
ensure that no desire was left uncatered for? 'Do artists need to =
program?' became the burning question at SIGGRAPH panel sessions and =
electronic art festivals.

To some extent this divergence between programmers and program users =
masked the fact that they had become two sides of the same coin. As the =
argument went, the artist-programmer would regard '.software not as a =
functional tool on which the "real" artwork is based, but software as =
the material of artistic creation', as the Transmediale Software Art =
jury statement would phrase it much later in 2002. On the other hand, =
for program users, programming was only ever a means to an end. Yet it =
was their fixation on this end that hastened their acquiescence to the =
means of their programs and the reconfiguration of their practice by =
programmers. 'Is the computer a medium or a tool?' Yes, it was true that =
some artists were only interested in software 'tools' that were totally =
subservient to their subjectivity, but it was a subjectivity that was =
now mapped onto minutely variable parameter lists and option check =
boxes, mirroring the remoteness of the artist's precious and peculiar =
visions by burying its origins deep within the recesses of multiple menu =
layers. Aided by the runaway success of packages like Amiga's Deluxe =
Paint, Adobe Illustrator and Photoshop, software manufacturers were =
redefining the creative process as a decision making process converging =
towards a predetermined ideal goal.=20

The problem was also attacked from the opposite direction by a top-down =
system design employing pre-sets, wizards, helpers, macros and plug-ins =
that pre-empted the creative process by offering a one button solution =
to achieve those essential lens flares, ripples, rollovers and drop =
shadow effects. The users of programs now found themselves programmed by =
their very own favourite artistic effects, expressed as a suite of easy =
to use software extensions. In the end, both artist programmers and =
artist program users produced artwork that was about the software that =
had produced it. Both became caught up in a wider move to rewrite =
society in terms of information processing.

By the early '80s the artist Harold Cohen had developed software to =
automate his own personal artistic style. A former successful gallery =
painter, Cohen still works on a suite of artificial intelligence =
programs called AARON that seek to encode his earlier painting practice. =
Cohen had always insisted that the content of his work was the software =
itself, and always exhibited the entire process in the form of a live =
computer connected up to a mobile painting device or 'turtle' that would =
scuttle over his canvases. As he told his students, 'Don't ask what you =
can do with the software, ask what the software can do.' But Cohen's =
work now seems to function more as evidence of a historical transition =
that occurred over his working life and reached its culmination during =
the '90s. While we have been watching Cohen's computer prove it can =
recreate art, other computers have been recreating our whole society in =
their own image. But this new image is not the image of the expressive =
subject that is simulated in Cohen's work. It is the image of the =
subject as a node, a switching station for providing feedback to =
calibrate the central processing system, the individual's expressive =
utterances only called upon to ensure their movements are correctly =
synchronised. The artist programmer of today exists in relation to a =
whole culture that has the computer as its central organising =
technology. The pervasive quality of software culture and the resultant =
normalisation of computer use have made it impossible to maintain the =
conceptual categories that underpinned previous debates. In a world =
where artists use software to write software that will be seen by virtue =
of other software, questions about the 'aesthetics of the code' become a =
symptom of not being able to see the wood for the trees. Programming is =
not only the material of artistic creation, it is the context of =
artistic creation. Programming has become software.

One interesting example of the end game of the debate on 'Computer Art' =
is a piece of artist's software called Auto-Illustrator. Written by =
Adrian Ward around the year 2000, Auto-Illustrator was the prize winner =
of the first competition for Software Art organised by Berlin's =
Transmediale media art festival in 2001. Ward describes the work as a =
parody of commercial art and design packages like Adobe Illustrator, =
specifically of their pretensions to provide functionality and user =
control. In contrast, Ward fills his package with 'generative art' tools =
that explicitly try to automate the drawing process. The appearance of =
Auto-Illustrator when running is much like a typical menu driven art and =
design package with the exception that the tool palette and effects =
filters incorporate generative algorithms. For instance, the Pencil tool =
adds wiggles or sweeps to your strokes, while the Oval tool will use =
settings like 'childish' or 'adult' to control a sprinkling of little =
faces. Some tools like Brush seem entirely random in operation, while =
some filters like 'Instant Mute Design' will reproduce an entire =
iconography designed to appeal to the Digerati generation.=20

In fact, many of these generative techniques are strikingly reminiscent =
of various experiments in computer art from over the last thirty years. =
The line tools generate scribbles using algorithms almost certainly =
related to the stochastic perturbations of Frieder Nake or Peter Beyls =
while the 'bug' tool roves around the screen using the same principles =
as Harold Cohen's turtle graphics engine. Even the icons of the 'Instant =
Mute Design' effect are almost identical to Edward Zajec's permutations =
of cubic modules. In this way, Auto-Illustrator is like a compendium of =
classic computer art programs but now presented as a list of menu =
options with conveniently editable parameters. Presented in this =
context, the individual aesthetics of each of these venerable pioneering =
practices are erased, leaving us with more of a confusion of =
idiosyncratic styles. From this viewpoint, Auto-Illustrator's =
'generative tools' actually pastiche the chaotic 'feature mountain' of =
bloated modern software systems, as they are commonly disorganised by =
the superabundance of toolbars, drop-down lists and floating inspectors. =
Instead of defining a drawing function, it might have been more relevant =
for Ward to have his 'bug' tunnelling into the dizzying depths of =
cascading sub-menus and option boxes to find that single cherished =
function with which the user nurtures their unique individual style. =
Ward actually states that wider issues such as interface design are of =
no interest to him and describes 'consumer-based application software' =
as his chosen medium. Auto-Illustrator is successful in its intention to =
parody the functionality-as-expression of mainstream software design, =
but only at the level of coding. By not addressing the wider user =
experience it is unable to think outside of the window box in which this =
functionality is now defined.

Since Auto-Illustrator's release there has been at least one attempt to =
account for a contemporary digital aesthetic with reference to the =
design of a family of software packages and related technologies. In =
2002 the theorist Lev Manovich published 'Generation Flash', an essay in =
which he tried to characterise a then prevalent cultural sensibility. =
Manovich referred to the prevailing visual style of Flash, Shockwave and =
Java based multimedia as 'soft modernism', a reaction against the =
clutter of postmodern eclecticism that returns to an elemental =
'rationality of software'. Aesthetic motifs are defined by Manovich in =
terms of technologically motivated processes: instead of appropriation =
we simply have the sample, a basic operation in the new mode of cultural =
production. Another cultural building block is the network, and =
therefore also one of the terms of a new critical language. These =
operations (networking, sampling) are applied in new modes of expression =
like data visualisation. This can be seen, for instance, in =
Futurefarmer's They Rule project in which the directors of the USA's top =
corporations are cross referenced to purportedly reveal a web-like =
pattern of interrelated allegiances. For Manovich this kind of work =
replaces older forms of authored representation by giving us the tools =
to objectively analyse raw data and deduce the necessary conclusions.

Although Manovich's detailed analysis of the structural basis of new =
media adds an absolutely essential dimension to new critical tools, the =
approach risks being interpreted as a form of technological determinism =
once we lose sight of a specifically cultural perspective. For example, =
our understanding of the workings of the corporate world order do not =
arise automatically out of its most common data visualisations, such as =
the stock market fluctuations diagrammatically portrayed on the =
Financial Times website. Not all visualisations are equal. At one point =
Manovich argues that the 'neo-minimalism' of the Flash style arises =
quite naturally from the practice of programming - the pixel thin grid =
lines, restricted colour palettes, abstracted symbols 'ALWAYS happens =
when people begin to generate graphics through programming and discover =
that they can use simple equations, etc' (Manovich's emphasis). This is =
indeed the case where programming is taught within a certain computer =
science tradition, but it is now impossible to discount the influence of =
scripting environments such as Flash. Not all programming practices are =
equal.

Other discussions of Flash have merely tended to shift the technological =
focus, such as whether the limited bandwidth of the web was the most =
significant reason for the linear aesthetic of vector graphics. At other =
times it moved on to question the 'openness' of the Flash graphics =
standard, whether Macromedia would ultimately allow programmers to =
leverage the full potential of its functionality. However, the =
'functionality', 'rationality' or 'potential' of software will always be =
strictly unknown. It is the 'user experience' of software, the values =
generated by the way it is meant to be used, how it gives shape to a =
practice, how easily a technical 'potential' can be perceived and =
engaged with that should form the basis of software critique. It is =
possible to trace many formative influences on the Flash style not to =
the code itself, but to the conditions in which it is written. =
Programming is now often practised in the form of 'scripting' languages =
that are integrated into mainstream art and design software =
applications. This makes artist programmers and program users both =
subject to the same philosophies of system design that hold sway in =
point-and-click style desktop packages. By examining these environments =
we can find many ways in which they funnelled Flash Actionscript or =
Director Lingo programming practice into nourishing certain wider =
cultural sensibilities during this period.

Multimedia scripting languages like Flash Actionscript tend to differ =
from conventional programming languages by offering access to a library =
of functions that are specific to that particular multimedia =
application. This easy access to a set of predefined 'events' such as =
mouse clicks, drag actions and rollovers is somewhat analogous to the =
way a software user's practice is structured in terms of the predefined =
configuration of menu commands, option boxes and plug-in effects. These =
library functions that populate the programmers imagination with a =
readymade vocabulary of discrete interactive 'behaviours' can be coded =
up and attached to individual multimedia objects - button triggers, =
sprite actions, sound effects, linkages, etc. Actionscript therefore =
tended to differ from typical program development environments by =
identifying code with graphical and other concrete entities that would =
become principle actors in the interactive scenario. This also tended to =
discourage the writing of long passages of control logic and instead led =
to the writing of terse mathematical expressions to manipulate an =
object's properties, movements and relationships to other objects. When =
combined with the instancing abilities of the Object Orientated =
Programming philosophy, Actionscript became very efficient at applying =
these code segments to multiple copies of 'semi-automated' graphic =
elements, sprites, movie clips and sounds. As implemented in multimedia =
authoring software like Flash, Object Orientated Programming actually =
fostered an 'object orientated' approach to interactive art and =
animation.

The point here is to look at Flash at the moment at which its patterns =
of techniques and processes re-emerge as motifs that can enter =
consciousness and practice on an aesthetic level. To start with we have =
an authoring system that orientated the user towards the replication (or =
'birthing') of multitudes of objects and orchestrating complex yet =
concise interactions between them. It is even possible to identify the =
most common form of mathematical expression that was used to regulate =
this interaction during the millennial Flash period. There is a single =
line of code that appears over and over again, a simplified expression =
that produces a distinctive dampening effect on a moving object before =
it finally comes to rest. It was easy for Flash users to apply this =
expression to any or all of ones objects and events until it produced =
the classic Flash 'wobble'. A Flash site became a constellation of =
rippling, bobbing, trembling buttons, icons, eyeballs, legs and rollover =
items as if someone had poured a bucket of water into your computer =
monitor. In the open source spirit, the Flash community ensured that =
such expressions were quickly disseminated until they became an almost =
universal kinetic attribute.

The Flash style was integrated, via its web browser plug-in, to other =
desktop based work and leisure patterns of activity. By keying into the =
internet gold rush fever, Flash art was turned into a highly visible =
design component of the dotcom boom era. This new informal space imbued =
Flash art with the role of a distraction, a demo or toy, making any more =
demanding appreciation of its fluid stylistic and tactile qualities =
unnecessary. The net culture of the time also provided a preexisting =
discourse in which it's visual aesthetic could be interpreted and =
flourish. Echoing the ubiquitous net-cultural meme of the 'digital Gaia' =
- an ecological interpretation of the web of globally interconnected and =
independent agents - foremost Flash designer Joshua Davis commented: =
'.our work should reflect the nature of a fern and be comprised of tiny =
little objects that all talk to each other. The more we add these little =
objects, the more complex and intense the nature of our work becomes.'=20

There are many more factors that could be marshalled to 'explain' the =
Flash style. But as far as practising artists are concerned, how can we =
get a handle on such a deluge of widely different factors, some of which =
seek to align us with a particular model of subjectivity and others =
which just seem like arbitrary collections of protocols? How can we =
forge a path through layer after layer of designed information to form =
ways of working not pre-empted by the predicates of current software =
culture?

There are some emerging ideas that might help. One of these is the =
'techno-aesthetic' - different motifs that permeate these technological, =
social and cultural levels. The idea is rooted in materialist notions of =
social process, but a society now constituted through IT. The emphasis =
is on how specifically cultural forces can form technology into a means =
of expression that is able to exceed its most obvious properties and =
structures. One software art example of this in action is Mongrel's =
often-cited Linker project of 1999. Developed to support a series of =
story telling workshops for the non-expert computer user, the software =
is a highly stripped down system that simply allows users to load and =
make connections between a collection of digital elements - images, =
text, video, sounds. For a start, this transfers an emphasis on the =
practice of the software to the practice of the user. Compared to the =
other examples, Linker coheres around a figure that unites its levels of =
thought and construction yet retains an open space in which imagination =
can breathe. As theorist Matthew Fuller described Linker, 'It relies on =
the simple function of doing exactly what the name says it does - link =
things. Here, the poetics of connection forms a techno-aesthetic and =
existential a priori to the construction of a piece of software.' This =
aesthetic is made explicit when the software is first launched - it =
displays a map image of its three by three grid of interconnected =
regions. Linker is constructed around this image of itself that =
communicates and instantiates its underlying algorithmic structure, =
creative use and conceptual model. It is this figuration of itself as an =
idea that makes Linker art as well as software.

The debate about Linker was unfortunately always limited to its mode of =
production and the social constituency of its intended user group as =
though it had been designed as a tool of social engineering, ready to =
arise fully formed out of a sub-menu check-box list of community =
'needs'. But discussions of DIY empowerment, Open Source and the =
'sociability' of software are presumptuous without any attention to the =
context in which imaginative ideas can grow. When we look at the kinds =
of applications that have actually resulted from Linux we simply see =
copies of standard Microsoft functionality. The Open Source model of =
production is a dead end without an equivalent 'model of creativity', =
defaulting instead to a wannabe culture. Instead we should look for =
inspiration in practices that could nourish a poetics of data =
'copyability' such as plagiarism and detournement, as noted by writer =
Josephine Berry. But unfortunately free software developers do not =
prioritise this aesthetic context which is what has the power to =
determine whether software will enable or restrain its user's =
perceptions and mode of action.=20

It is not a matter of the different technical abilities of software or =
of how much it costs, but of how easily a technical potential can be =
perceived by the user in a way that motivates engagement. When software =
is written, choices must be made about which data fields carry value, =
how the display of information forms contours of meaning, how the =
modelling of the interface moulds the subjectivity of the user. The =
question of whether artists should learn to program is replaced by the =
question of what kind of programming. Which programming practice has the =
most 'open aesthetic', capable of making software that is not just the =
product of an arbitrary confluence of techniques or a slavish mimicry =
but is aware of all its possible formative cultural and philosophical =
categories and values.

For the first generation of artist programmers there was hardly any =
information society in existence, certainly not one within reach. In the =
early eighties during a period when the launch of the personal computer =
marked a radical shift in computer culture, artist Harold Cohen stressed =
the importance of asking the right questions. Now that we live in a =
world in which his AARON program is downloadable as a screen saver it is =
time for us to extend his question - 'Don't ask what the software can =
do, ask what it can do to other software.'

URLS:
Auto-Illustrator: www.auto-illustrator.com
Joshua Davis: www.joshuadavis.com
Linker & 9: www.linker.org.uk, 9.waag.org
AARON screensaver: www.kurzweilcyberart.com

Acknowledgement
This article was based on research supported by a grant from the Arts =
and Humanities Research Board.


#  distributed via <nettime>: no commercial use without permission
#  <nettime> is a moderated mailing list for net criticism,
#  collaborative text filtering and cultural politics of the nets
#  more info: majordomo {AT} bbs.thing.net and "info nettime-l" in the msg body
#  archive: http://www.nettime.org contact: nettime {AT} bbs.thing.net