This page has been archived by the Computer Arts Society.

Definitions based on procedure


One reason for using a computer is because the artist’s work already has a systematic or procedural element that is apt to be transferred to a computer. For this reason, certain theorists defining Computer Art have restricted their attention to specific computer-based art practices. For instance, computer artist Judson Rosebush invented the term “proceduralism” to refer to the use of computer processes as an integral part of image creation. Rosebush identifies the programmed element as being essential to his area of Computer Art; the artist’s understanding that the art results from initiating particular processes.  As Rosebush puts it:

art made by employing scripted, notational directions that specify processes and parameters; the picture is produced by executing these directions rather than drawing it directly.[1]

“Proceduralism” is useful shorthand for these algorithmic processes, though I prefer Jean-Pierre Hébert’s term “algorism” for its brevity. “Algorism” refers to the act of using algorithms to create art. The Algorist Roman Verostko considers the algorithm to be a “step by step procedure for solving a problem.” In the sense that they encode a series of procedures, musical and choreographic scores and architectural plans could be considered algorithmic.[2] Verostko points out that such procedures predate the term “algorithm”, which is derived from the surname al-Khowarazmi of the 9th century Baghdadi mathematician Abu Ja’far Mohammed Ben Musa, via Latinised variants such as algorismus. When Hébert founded the Algorists in 1995, he introduced Verostko to these definitions through Donald Knuth’s book The Art of Computer Programming.[3]

Rosebush understands the limitations of using procedure as a criterion for defining ‘Computer Art’: “Proceduralism does not claim to embrace all Computer Art: its aesthetics are a subset of Computer Art as a whole.”[4] Indeed, if the aesthetics are so closely tied to usage, how could any term have primacy with regard to computers? However, Rosebush is attempting to legitimise a computer-specific artform in its own terms, by drawing on the one factor which is inherent to computer operations which can also become an aesthetic element in the art itself. Because the algorithm is at the heart of computing, an algorithmic or procedural artform on the computer takes this as a starting point for its organisation and development. Perhaps an algorithmic artform developed without a computer might be said to imitate this process; in other words it is more appropriate to use a computer for algorithmic art, rather than the other way around. Still, as process art demonstrates, procedural elements were widespread in art before the advent of the computer. For this reason, it is still imperative to find a computer-specific artform.

Tom DeWitt proposed a similar concept of “dataism”, taking as its basis the “innate formalism” of programming as a consequence of the “specific processes that are expressed in a defined notation.”[5] This formalism concerns the programmatic elements of the artwork, viewing it as a sequence of directed processes. Like Rosebush, DeWitt recognised that process-based elements were not solely a property of computer-based artworks. Conversely, he decided that not every artwork created on the computer was an example of “dataism”. DeWitt favoured the work of artist-programmers as opposed to artists using commercial software packages.

Identifying Computer Art with programmed art stems from a basic distinction in computer usage. Artists who program are distinguished from artists who use mice and windowing systems to make their art. The Graphical User Interface (GUI) is based on the metaphorical device of “drawing” directly on a simulated canvas, rather than approaching a computer image as a series of scripted procedures. Computer purists claim this merely transplants earlier methods into a digital setting and cannot therefore qualify as “Computer Art”. [ref] They propose that programming not only conveys a deeper knowledge of the computer, but that it is more appropriate for a computer artform and use the computer at a more fundamental level.

Although De Witt’s definition seems to exclude GUI graphics users, he points out that even paint packages may be subject to the “Dataist” definition because they are, fundamentally, programs that act upon an image. But the underlying programmatic aspects would seem to matter less to an artist drawing in a GUI system, which is perceived to bear a resemblance to its physical counterpart. In a similar vein, one could consider whether knowing the chemical composition of oils is essential to a general discussion of an oil painting’s aesthetics. The interaction of the oils is an important aspect of the painting’s creation, but it is not necessarily evident in the final work, nor is it always a conscious component of the artist’s working of the image. Similarly, a GUI-using artist may be aware of the underlying computational processes, but still approaches the image as a drawing on a flat surface.

When creating a category for computer artworks, making programming the cornerstone of a definition implies that GUI usage is “inartistic” in a computer context. This disregards the open-ended nature of the computer and the range of artistic comprehension which may be applied to it. Certainly, the product of a GUI system may be judged inartistic, of itself, and one might criticise direct manipulation with a Graphical User Interface as facile. Yet this is more a judgment on the artist’s lack of vision than an in-built flaw of the system itself. When the computer artist Frieder Nake proposed that only programmatic artforms should be used on computers, Gary William Smith disagreed with this attempt to “formulate what the use of a given medium ‘should be’ […]”, saying “Why limit exploration when it has hardly even begun?” [6] Thirty years later, the diversity of approaches to Computer Art is such that any attempt to limit the definition by practice would be a retrograde step.

[1] ACM SIGGRAPH 89, Judson Rosebush “The Proceduralist Manifesto”, p55

[2] From “Algorithmic Art: Composing the Score for Visual Art” by Roman Verostko (htm document, 1999)

[3] Correspondence with JPH, Nov 2002

[4] Rosebush, ibid, p55.

[5] SIGGRAPH’89 – Computer Art in Context

[6] Gary William Smith, PAGE: Magazine of the British Computer Arts Society, No22 April 1972.