What Is Interface Aesthetics, or What Could It Be
(Not)?
Florian Cramer
9/5/2007
What is interface aesthetics, what could it be, or perhaps:
what could it be not? There is, first of all, a colloquial common
sense understanding of the term that is best illustrated through
screenshots:
- the aesthetics of Windows
- versus the aesthetics of Windows Vista
- versus the aesthetics of Mac OS X
- versus the Macintosh from 1984
- which borrowed from the Xerox Star
- in contrast to the command line interface of Unix
- and its counterpart in 1980s homecomputers such as the Sinclair
ZX-81 and ZX Spectrum
- whose aesthetic has been artistically resurrected in the work
"All wrongs reversed" by jodi
- ... artists who had experimented with the aesthetic of
operating system and browser user interfaces jodi earlier, for
example in their work OSS
- as opposed to the likewise fictitious spinning, blinking and
beeping in-your-face 3D operating system user interfaces in
Hollywood movies as summed up in the Wikipedia article
"Hollywood OS".
While these are powerful, influential and passionately
discussed examples of interface aesthetics, they merely represent a
part of it that too often gets mistaken for the whole. Interfaces are
more than just user interfaces, and aesthetics means more than just
look and feel.
What is an interface?
Literally, interface means something in between. The term
originated in chemistry where it means, according to Webster's
dictionary, "a surface forming a common boundary of two bodies,
spaces, phases". In computing, interfaces could be defined as
anything that acts as a common boundary or link in a cybernetic
feedback loop either between machine components or between humans and
machines. Since the machine called computer differentiates itself
into hardware and software, interfaces can thus link:
- hardware to hardware, such as CPU sockets, buses like PCI and
USB;
- hardware to software. A good example is the Windows key on PC
keyboards, or multimedia controller keys that no longer physically
http://www.pocket-lint.co.uk/news/news.phtml/5940/6964/hollywood-technology-useability-movies-wrong.phtml
control audio hardware, but act as front-ends to software
controls, CPU instruction sets such as x86 that no longer
correspond to the internal CPU design, but act as software
compatibility layer;
- software to hardware; the classical examples are operating
system kernels and drivers;
- software to software: plugin interfaces, file formats,
protocols, and most generally: application programming interfaces
(APIs), i.e. interfaces used between computer programs such as C
standard library API, Unix system calls or the DOM/Javascript API
in Webbrowsers;
- humans to hardware: keyboards, mice, screen and audio feedback,
any controller and feedback device;
- human to software: user interfaces.
Out of these six possible interfacings, media studies tend
privilege if not limit themselves to the latter. One could call this
phenomenon "the restrained interface", deliberately
borrowing from the French literary theoretician Gerard Genette and
his notion of "the restrained rhetoric". Genette observes
how modern rhetoric and literary studies from Giambattista Vico to
structuralism had gradually reduced its notion of rhetoric to merely
figures of speech, and finally just to the single figure of metaphor.
The similarity to contemporary media studies is all the more striking
considering that the restraint of interface to graphical
human-to-software-user interfaces is conversely linked to a focus on
visual metaphors implemented in those interfaces.1
Another problem is that the notion of the user interface is
restrained in itself, since it is based on a politics of artificially
separating "users" from "developers" based on
different priviliges of access to machines functionality granted by
those interfaces. The differentiation of user and programming
interface - in programmer lingo: UI and API - is purely conventional,
and hence arbitrary. In 8-bit BASIC computers like the SInclair ZX
Spectrum, on the Unix command line, in MIT's Lisp machines and
even in Alan Kay's initial graphical Smalltalk system, the user
interface was also a programming interface and vice versa. Live
Coding, such as in Adrian Ward's and Alex McLean's Perl-based
musical performances or in jodi's ZX Spectrum video, shows how an
API can be employed as a user interface.
And of course, aesthetic notions exist just as well for
software-to-software interfaces, hardware-to-software,
software-to-hardware and hardware-to-hardware interfaces. The search
term "ugly API" currently yields [...] Google search
result, "beautiful API" [...]. Similar results can be found
when researching Internet flamewars on the beauty or ugliness of
other technology, in the sense of: beauty and ugliness of the
internal, logical design. This brings us into the field of
mathematical, scientific and engineering aesthetics that has existed
since Pythagoras equated the musical octave in its beauty to the
division of numbers by two. Beauty and ugliness, as judgments of
taste, however bring us into the realm of aesthetics. [Fludd,
Monochord]
What is aesthetics?
Judgments of taste of course involve more than just beauty and
ugliness and, to come back to my initial screenshots, more than just
look and feel. Ever since the Alexander Gottlieb Baumgarten coined
the term aesthetics in 1735, the concept has been philosophically
understood in a much farther reaching anthropological sense as:
- theory of perception
- theory of judgments of taste. Classicist aesthetics focused on
beauty - which its more sophisticated theorists from Winckelmann to
Walter Pater always understood as a inherent tension between
consonance and dissonance -, other theories of art from
Pseudo-Longinus to romanticism and postmodernism tended to focus on
such aesthetic categories of the sublime, the ugly, the disgusting,
the obscene.
- theory of art, notably since Hegel, and thus encompassing
older-than-18th century philosophies of beauty and art, such as
Pythagorean philosophy and Neoplatonism.
In other words, not much is gained if one broaden ones focus
from the restrained interface to a more comprehensive technical and
cultural understanding of computer interfaces if one ends up
discussing what is an ugly or a beautiful API, thus buying into the
post-Pythagorean fallacy of art as conceptual beauty of technical
design. It is just another restrained concept of interface aesthetics
because it reduces of art to beauty and of aesthetics to theory of
beauty, with the implication that beauty is good and ugliness is bad.
Such equations are turned upside down in concepts of the
"hack" as an ugly, yet efficient and ingenious fix, and in
art like jodi's which honors the ugliness of software [pic].
Alexei Shulgin's projects 386 DX and Fuck-U-Fuck Me http://www.mediaartnet.org/works/fuck-u-fuck-me,
for example, point out and value ugliness and obscenity of
hardware-to-user interfaces, too. Edmund Burke's 18th century
"Investigation of Our Notions of the Sublime and Beautiful"
names pleasure and pain, terror, obscruity, suddenness among the
registers of aesthetic perception, and here they also turn into
pleasure and pain of the hardware interface, terror of the desktop,
obscurity of the API, and so on. In other words, the more
sophisticated strands of digital art supplemented the
post-Pythagorean engineering aesthetic of the beautiful interface
that had likewise governed the institutional mainstream of electronic
arts, with a reflection of the computational sublime. Which of course
seems to perfectly match Jean-François Lyotard's
identification of the sublime experience to a postmodern
"incommensurability of reality", a condition for which he
all the more accounts computational information technology.
A problem, however, is that the mere notion of computer
interface aesthetics contradicts textbook aesthetic theory. It is
completely outside the conceptual framework of, for example,
Kant's Critique of Judgement and Hegel's aesthetics since it
neither involves perception of nature, nor of art, being instead
about the perception and judgment of determinist logical systems. As
such, it conflates the aesthetic with what Kant considered to be its
opposite when he writes in the Critique of Judgement:
"That which is purely subjective in the representation of an
object, i.e., what constitutes its reference to the subject, not to
the object, is its aesthetic quality. On the other hand, that which
in such a representation serves, or is available, for the
determination of the object (for or purpose of knowledge), is its
logical validity."
Technological aesthetics, thus short-circuits, or reconciles
Kant's opposites of knowledge and judgements of taste,
respectively of logic and aesthetics. That used to be its hack and
punk appeal.
But what remains when that appeal has worn off? Even an
interface aesthetics that reflects user interface (and) API,
hardware and software, beauty and ugliness might be
unsatisfying as a more comprehensive theory of art. In the worst
case, it boils down to glorification of new technology as the better
contemporary art, be it in iPod lifestyle, Web 2.0 hype or media
theoretical glorification of CPU instruction sets. None of this is
provocative anymore like Marinetti's saying that a racing car is
more beautiful than the Nike of Samothrake used be, but is in
mainstream accordance with the ideology of "creative
industries" supplanting autonomous arts.
If experimental digital art is seen as its mere antithesis, and
playful subversion of interfaces, one runs danger of reducing it to
the role of the court jester of those industries.
Yet interface aesthetics might have its place as a critical
perspective on both media theories and computer science right because
of its insistence on the "aisthesis", perceivability of the
linking agent in cybernetic feedback. If an API can be as ugly as a
GUI, it means that it is also a user interface. Unlike a conventional
engineering angle, an aesthetics perspective thus can shed light on
the arbitrary political nature of interface conventions. The
restraint of the notion of the interface just corresponds to the
technical restraints in the interface proper. This stil means that
"interface aesthetics" remains a limited subset of
aesthetics which will never be sufficient for a critical perspective
on electronic arts, but a seductive trap for critiques to get caught
in. Perception and subjective judgments of taste as a whole cannot be
reduced to cybernetic feedback unless one buys into behaviorism and
naive ideologies of artificial intelligence and artificial life.
In short, the issue seems to be that, for someone whose primary
interest is aesthetics and the arts, not computing, the study of the
beauty, the ugly and the absurd of computers has some hack appeal,
but gets dull sooner or later because it's about deterministic
systems whose states are finite. This why researching interface
aesthetics is not something I would like to do for the rest of my
life.
Aside from that, an even more serious issue lurks behind such
an approach. Ever since the late Heidegger applied existential
philosophy to technology, to the effect that technology became an a
priori of the human condition rather than a human product, there is a
technological sublime that many media theories have subscribed to. As
second nature, technology aesthetically takes the place both of
Kant's natural beautiful and his mathematical and dynamic
sublime. We are, in other words, no longer overwhelmed by mountain
ranges or thunderstorms, but for example by the pervasiveness of
computing. If, with Lyotard, the sublime experience of computing thus
becomes postmodern experience of incommensurability of reality, it
obscures the fact that this particular reality (i.e. the reality of
computing) is based on constructed policies and can be hacked.
There are ways out of those dilemmas. One would be a simple
shift from "interface aesthetics" to an "aesthetics of
interfacing", in other words, towards a perception anthropology
of practices instead of systems. The 386DX software and
hardware interface for example might be dull by itself, its musical
output, too, but Alexei Shulgin's rock star performance which
interfaces with it isn't.
But this should not mean to subscribe to an aesthetics of
social practices either. Theodor Adorno's aesthetic theory
remains, in my opinion, an outstanding critical achievement in its
refusal to bless either practices or objects - respectively the
autonomy or sociality of art -, putting them instead into an aporia
of a negative dialectics that produces no synthesis and trying to
"break the coercive character of logic with its own
means".2 For Adorno,
"The dual nature of artworks as autonomous structures and
social phenomena results in oscillating criteria: Autonomous works
provoke the verdict of social indifference and ultimately of being
criminally reactionary; conversely, works that make socially
univocal discursive judgments thereby negate art as well as
themselves"
Just as this dialectic implies a critique Kant's and
Hegel's concepts of natural beauty, it means that technology -
including GUIs, APIs, protocols, instruction sets etc. - by itself
isn't very exciting from the point of view of a critical
aesthetics. Its objecthood, without any artistic intervention, is
"creative industries" territory,3 lacking the negative
dialectical pole of critical autonomy and asociality. In this context
discussed here, this critical quality is only achied through the
aesthetic interfacing by artists like jodi and Shulgin; a quality
which resists recuperation into what an alliance of
techno-existentialist media studies and the so-called creative
industries might have in mind with"interface aesthetics" or
the "aesthetic interface".
Footnotes:
1Such as in laurel:interface.
2Minima moralia, p. 199
3The meme of the creative industries
should however not be mixed up with Adorno's and Horkheimer's
notion of the cultural industries - since it's not about the
entertainment industrialization of art, but the idea that industrial
product design aesthetically supplants and obsoletes autonomous art.