How can we synthesise human craft and computation?

A Goldsmiths Computational Art Research Project, Term 1 Dec 2020

Authors: Harry Wakeling and Chris Courage

This project is a contribution to the open conversation on the place of craft within a contemporary approach to digital design and fabrication. The intention was to discover new ways of thinking between the computer and the craftsperson.


The objective of the project was to create generative designs that can inform a given craft process, such as architecture or model making, with the intention of producing a series of tangible objects.

Prior to the Industrial revolution, the craftsperson was the prominent mode of production. They produced unique handmade objects, not one the same. Along came automation, with the value of mass-produced artefacts becoming cheaper and of consistent quality; as a result of this, personal uniqueness was lost. The social value of craft was changing.

The shift in attitudes was never more obvious than in 2002 when the American Craft Museum changed its name to the Museum of Art and Design.
-Zoran and Buechley, 2013, Hybrid Reassemblage: An Exploration of Craft, Digital Fabrication and Artifact Uniqueness

What is it about craft that makes it an important part of society? The theory of craft looks at the intrinsic values of human aesthetics and functions worked into and embodied in the craft object.

“Craft practice is rooted in the relationship between materials, tools and techniques as an intricate workflow”
-Loh et al., 2016 Reconsidering Pye's theory of making through digital craft practice: A theoretical framework towards continuous designing

We aim to contribute to this relationship between workmanship and workflow. We will evident this in our exploration phase, fine-tuning a collection of artefacts to produce something unique.

Research questions:

  • What can craft bring to digital fabrication?
  • Can we produce a generative design that is unique and contributes to the given craft?
  • What are both the limitations of the craft and computation?
  • What is the social value of generative craft?


What was our aim?

Our intentions were:

To explore a variety of generative algorithms and to develop an understanding of computation from a craft perspective, producing an artefact that could open up and add to the dialogue between different material practises. We used a fork approach to develop two lines of research and cover more research possibilities.

“Once the exclusive domain of programmers, code is now being used by a new generation of designers, artists, and architects eager to explore how software can enable innovative ways of generating form and translating ideas.”
-Reas and McWilliams, 2010 Form+Code in Design, Art, and Architecture

The two areas being covered were shape and form and image and data. Shape and form looked at 2d generative shapes using a basic algorithm and parameters. Image and data looked at information that is hidden in the pixels of an image and invisible to the human eye. This used a more complex algorithm to interrogate pixels, mapping the data that was uncovered to a visual representation.

Shape and form

This was our initial experimentation where we developed shape and form. The basis of the algorithm contained the following:

  • Form and function
  • Confined square boundary
  • Arbitrary size and position
  • Minimum parameters
  • Minimum points to shape
  • Easily mapped parameters

The idea of developing a workflow that synthesises craft and data infused computation came from the journal article ‘Data Materialisation: A hybrid process of creating a teapot’. The teapot experiment conceptualises data through a 3d form which in-turn is digitally fabricated.

Cinque Terre
Credit and debt data translated into teapot form
Starrett, C., Reiser, S. and Pacio, T. (2018). Data Materialization: A Hybrid Process of Crafting a Teapot. Leonardo. MIT Press. Vol. 51 No. 4. pp. 381–385 [online].

The evolution of a generated form

Analysis: Shape and form

Our first line of research looked at shape and form. The crafts unique systems dictate the workflow being applied to the computation, which in turn constrains the programmer to use parameters that produce some type of anticipated result. From the code it was not possible to guess what shapes would be generated. In hindsight after seeing the generated results, it is now obvious that the shapes would be more triangular and less geometric due to the parameters and number of points in the shape. Furthermore, the aesthetic value of these generative shapes has a certain sense of materiality. There was no bias towards it, but the generated shape lends itself to have a material quality based on its form. This characteristic of the computation was not anticipated.

Live shape generation, data taken from horizontal MOUSE POSITION.
This animation demonstrates that with little data input the algorithm
can generate a unique shape.

Image and data

The second line of exploration started by experimenting with 3D shapes in Processing. We were interested in creating them with code, using techniques such as 3D printing to turn them into real, tangible objects. The starting point was to create some super-shapes inspired by the work of Daniel Shifman:

Cinque Terre
Super shape

The shape and segments of the object were influenced by the movements of the user’s mouse. The next stage in workflow would have been to export this into an STL file, allowing the artefact to be printed on a 3D printer.

Leading on from this the experiment, we investigated generating 3D objects from the pixel values of an image. We started by creating an experiment using a photograph of Battersea Power Station, used famously as the cover for Pink Floyd’s ‘Animals’. By mapping the brightness of each pixel we could move a grid of cubes along their z axis within a for loop, increasing the amount of movement with a mouse. Moving the mouse from left to right produced the explosion that you can view below.

Cinque Terre
Image explosion experiment

This experiment was another development in creating objects from the data of a 2d image. The picture in question was an aerial shot taken from Google Maps, gathered by the programmer Paul Bourke. It is part of a wider series displaying how fractals can be seen across different landscapes. We were inspired by a piece from ‘Form + Code’ that generates 3d explosions by displacing the brightness of each pixel on its z axis. By displaying it in 3 dimensions, the viewer can view the form of the image from multiple angles, interpreting its data from different perspectives. 

Cinque Terre
Fractal information from an image

Cinque Terre
3D visualisation. Pixel brightness mapped to Z-depth

Analysis: Image and data

If we had more time and access to a workshop, we would have liked to experiment with creating physical objects from this data, using techniques such as 3D printing and CNC routing. This is a process commonly used within architecture for creating forms using topographical data, an example of which can be seen below. This from was fabricated from routing topographical data of the Grand Canyon into a block of wood.  

Cinque Terre
Digital fabrication of an image

This allows the observer to the view the form of a previously flat image from multiple angles. The creation of physical forms also allows the creator to give different meanings to the intention of the original data. An example of this would be if the creator were to build a table out of the routed wood. This gives it an audience, a function, and a distinctly different purpose to the original image, which was to display a natural form using data from Google Maps. 


Interrogating the two lines of research: Shape and form; Image and data

As our research topic was fundamentally about craft, we decided that our project should be guided predominately through experimentation. The work produced for our artefacts proved it is possible to generate functional objects through generative designs. We believe uniqueness is the primary quality of computation, with the secondary being the designer that constrains function. The algorithms used to generate the shapes can produce countless combinations, but it is the experience and knowledge of the artist or designer to influence it by combining the uniqueness and functionality which ultimately is embodied in the object. The final fabrication process is dictated by the design unless factored in from the initial stages, thus further adding constraints to the computation. In conclusion, the factors of design and fabrication previously mentioned can only argue that generative design and fabrication are crafts in their own right.

Knowledge and experience are both fundamental to craft and computation, but also limiting. It is not possible to design an algorithm for an arbitrary craft; instead it is necessary to have pre-existing knowledge to inform the design of the algorithm. Having proficiency in constructing algorithms does not mean that you are able to create generative designs without first having prior knowledge of that craft. With the technology of digital fabrication well developed, fabrication is no longer a limiting factor. It is only the wild imagination of the artist that could now limit what is possible.

The social value of craft and computational synthesis can only be a positive one. Computation interacts with humans and their being, animating ideas that an audience can participate with directly. Synthesised with craft, computation can be embodied into physical objects adding uniqueness of past, present, or future data, allowing for the exploration of chance and unpredictability. An example of this can be seen in the aforementioned piece ‘Fractal Dice’. This synergy allows us to reinterpret existing data (such as images or videos), giving them physical presence and new meaning. It enables the ability to make invisible data visible.

Computing can often be seen as something of a cold, daunting process to an outsider with no experience in programming. Using computation within processes such as sculpture or painting can give the algorithm a ‘soul’. We can manipulate this just as a craft person manipulates technique, altering parameters depending on what you want to produce; this is often informed by the craft and the technique.


Agkathidis, A. (2016). Generative Design (Form + Technique). Laurence King Publishing.
Iwamoto, L. (2009). Digital Fabrications: Architectural and Material Techniques. Princeton Architectural Press.
Johnston, L. (2017). Digital Handmade: Craftsmanship in the New Industrial Revolution. 2nd edition. Thames and Hudson Ltd.
Loh, P., Burry, J. and Wagenfeld, M. (2016). Reconsidering Pye’s theory of making through digital craft practice: A theoretical framework towards continuous designing. Craft Research. Vol. 7 No. 2. pp. 187–206 [online].
McGee, W. (2016). Robotic Fabrication in Architecture, Art and Design. Springer.
McWilliams, C. and Reas, C. (2010). Form + Code in Design, Art and Architecture. Princeton Architectural Press.
Perry, G. (2016). Are computers killing off craft? Not a chance [online]. Available from: [Accessed 2 December 2020].
Programming Design Systems (n.d.). Procedural Shapes - Programming Design Systems [online]. Available from: [Accessed 1 December 2020].
Starrett, C., Reiser, S. and Pacio, T. (2018). Data Materialization: A Hybrid Process of Crafting a Teapot. Leonardo. MIT Press. Vol. 51 No. 4. pp. 381–385 [online].
Zoran, A. (2013). Hybrid Basketry: Interweaving Digital Practice within Contemporary Craft. Leonardo. MIT Press. Vol. 46 No. 4. pp. 324–331 [online].
Zoran, A. and Buechley, L. (2013). Hybrid Reassemblage: An Exploration of Craft, Digital Fabrication and Artifact Uniqueness. Leonardo. The MIT Press. Vol. 46 No. 1. pp. 4–38.