Drawing Machine (Output = Plotter), 2008-12

Installation view, Primavera 2012, Museum of Contemporary Art Australia
Image courtesy Museum of Contemporary Art Australia
© the artist, Photograph: Alex Davies

what is drawing?
Defining a concept is analogous to drawing a boundary line between ‘what is’ and ‘what is not’. The problem with concepts such as Drawing, and more generally Art, is that whenever a boundary is drawn, upon closer inspection the boundary seems inadequate. It either excludes something essential, or else it includes the extraneous.

Benjamin Forster’s Drawing Machine project can be summarised as an investigation of drawing using a specific system-based methodology. This project has two concurrent aims: firstly to explore ideas about drawing, and secondly to raise questions about the authority of reason and logic as methods of understanding.

This is not an investigation of any specific style of drawing, but simply drawing as the act of making marks on a surface; how these marks are made in relation to one another and, most importantly, what knowledge is necessary in order to make such marks. This investigation centres around his attempt to program a computer to draw in a way that is distinctly human, rather than stylistically digital or mechanistic. It is important that his program simulates the human characteristics of drawing because it is exactly the human quality of drawing that he has been attempting to understand. Benjamin believes it is only through comparison and contrast to human drawing that his machine’s drawings reflect the inadequacy of systems to capture the infinite detail of the world.

The Process

Step 1: Look at my own drawings.

Step 2: Formulate formal procedures that encapsulate my ideas about drawing.

Step 3: Express formal procedures in computer code.

Step 4: Compare the resulting machine drawings with my own drawings.

Step 5: If machine drawings = my drawings then STOP, else continue to step 6.

Step 6: Refine formal procedures.

Step 7: Goto step 3.

Note: This machine will never produce the same drawing twice.

Drawing Machine (Output = Plotter), 2008-12
Primavera 2012, Museum of Contemporary Art Australia
Photography: Robbie Karmel

Drawing Machine (Output = Screen)

Film demonstrating the general progression of the screen-based version of Benjamin Forster’s Drawing Machine.

“…a digital environment is an abstract projection supported and sustained by its capacity to propagate the illusion (or call it a working model) of immaterial behavior: identification without ambiguity, transmission without loss, repetition without originality.”

— Kirschenbaum, Mechanisms: New Media and the Forensic Imagination (2008)

relevant [ <09 11 12 Drawing Machine ]

Drawing Machine State 02:11:08

Second iteration of Benjamin Forster’s Drawing Machine. Discarded the concept of the compositional drawing on a plane and instead focused on drawing as the embodiment of thought. Drawing inspiration from impulsive mark making such as, note-taking and doodling.

relevant [ <09 11 12 Drawing Machine ]

Drawing Machine State 24:10:07

First iteration of Benjamin Forster’s Drawing Machine. Examining drawing as compositional responces to a rectangular plane.

relevant [ <09 11 12 Drawing Machine ]

House Drawings, 2010

Selection of ink drawings from a series concerned with architecture, stability and certainty. These drawings were exhibited as part of BLAZE 2010 at Canberra Contemporary Art Space.

relevant [ Drawings ]

Inscribed (Ko Aye Aung), 2010

This work was created as part of the Amnesty International’s ARTillery Festival as part of the MasterPEACE exhibition. Artist’s involved were asked to make work in response to the story of an unfairly imprisoned individual at risk, people whose right to freedom is currently being campaigned for by Amnesty International.

Inscribed (Ko Aye Aung) is an explicit analogy of the process of degradation and loss that has been occurred by Ko Aye Aung. I do not know Ko Aye Aung. If it was not for Amnesty I would not even know his story. All knowledge of him was provided to me as a simple digital profile. Dot pointed and only nine hundred and eighty two words accompanied by a small portrait. His is a story that is singularly unique, however at the same time multiple and all too common. His is a story we never truly know, we never hear about, we are all distant from.

This video is made through a simple process.
Insert Ko Aye Aung’s digital profile, one letter at a time, into and over the code that makes up his JPEG portrait. Each frame of this video is a record of the degradation and loss incurred as each letter is added until at last all his profile is written into the image and his face is lost.

relevant [ Codeworks ]

Computer Watching Television (Channel Ten 04/02/10 23:12)

Made using a simple algorithmic process that reinterprets video input sourced from a computer’s webcam placed in front of the television. The process mimics the slit-scan method that televisions use to render received signals as perceivable images.

The Process:

// for each new frame of video.

for(int jj=0; jj<4; jj++) {

  int x = ((ix+(jj*4))%w)*dw;

  int y = (iy+(jj*4))*dh;

  for(int i=0; i<dw; i++) {

    for(int j=0; j<dh; j++) {

      int t = (x+i)+((y+j)*width);







if(ix>=16) {


  if(iy>=4) iy =0;



relevant [ Codeworks ]

Discourse, 2010

This work is the result of an inquiry into the process of conversation based purely on statistical probability. The transcript actively spilling forth from the two receipt printers is a hypothetical dialogue between the two classical economists Karl Marx and Adam Smith.

The System.

Each sentence is constructed from the statistical probability within a source text that any given word follows from another. The two source text used are Karl Marx, Capital, Volume II: The Process of Circulation of Capital, and Adam Smith, An Inquiry into the Nature and Causes of the Wealth of Nations, Book II: Of the Nature, Accumulation, and Employment of Stock.

Starting from scratch the first word is randomly chosen based on its frequency within the source text. Words with a higher frequency of occurrence are more likely to be chosen then words appearing less often.

The next word in the sentence is then randomly chosen from all the words that follow the proceeding word. This choice is informed by the frequency of those words occurring after the preceding word/s within the entirety of the source text. Then the next word is chosen based on the preceding two words, and so on. It is important to note that this specific instance of the system is limited to a maximum resolution of three preceding words. This process continues until the sentence ends. A sentence is ended when there is no word within the source that follows from the preceding word/s, or given a series of words there is a high probability that a full stop, or equivalent symbol, occurs next in the source.


The system of conversation is very limited. There are only two agents within any conversation. The agents take turns in talking. When one agent talks, the other listens until the end. There are no interruptions.

One agent is randomly chosen to begin the conversation. It constructs a sentence from scratch as describe above. When the sentence is finished the listening agent responds.

A response is constructed by searching the sentence received from the other agent for words or series of words that occur within its own source text, or vocabulary. It then selects between these based on the probability of potential words that follow within its source text. Words or series of words that have the largest amount of possible continuation are more likely to be selected then those with fewer. Once selected, the agent then chooses the first word of its response using the procedure described above.

This process is endless.

This work has also been reproduced as a 180 page book, “Karl Marx and Adam Smith: Dialogue. An inquiry into conversation through statistical probability“. Indefinite edition.
Resale price: $49.95

Discourse won the Non-Aquisitive Prize at the 2010 Fremantle Arts Centre Print Award.

Ambiguous I, II, III, IV, 2010

What makes an image offensive? And how does this relate to the boundary between meaning and information?

Ambiguous is a series of digital prints. Each one contains the exact same pixels sourced from an ‘aberrant’ image, although each image is a unique algorithmic translation of this taboo data. They all are displaying the exact same data; concealing and censoring in plain sight. The simple process for each translation is displayed at the bottom of each image.

relevant [ Rational ]

return to the top