Peg, who has a talent for interesting questions, commented on my last blog post:
Tien, is this to be a wall piece? My problem with the usual design books, is that designing a piece of fluid fabric that is to be worn has more complex design problems that, so far as I can see, the usual art design books do not address.
The final product will not be a wall hanging, though I may produce one en route to the finished garment. And you are quite right – there are many more design considerations when creating a garment than are addressed in two-dimensional design books. (For one thing, the eye is naturally attracted to faces, so the eye always starts at the head of the wearer.)
That said, I am currently designing and sampling as if it were a two-dimensional piece, because it allows me to identify and solve design issues that, while not identical, will be relevant to the design issues in the finished garment. The motion of the eye down the flat form will not be identical to the motion in a garment on the wearer, but I think it will be similar enough that the flat samples will prove worthwhile.
The purpose of a sample, to me, is to answer a design question. The sample does not have to be the best answer to that question; in fact, it rarely is. What a well-designed sample does is provide an adequate answer to the specific design question, with minimum effort and resources. It models the important aspects of the design problem and ignores the rest. In doing so, of course, it makes assumptions about what is and is not relevant, and those assumptions may turn out wrong. But then one does further sampling and design to correct those errors, producing a set of successive approximations that eventually converge on good design. By making simplifying assumptions, it’s possible to break down the overall problem of design into many smaller and more addressable issues, which can be analyzed and sampled more quickly than if one tried to address all of them at once.
This method has solid roots in the scientific method. The purpose of a scientific experiment is very rarely to derive an entire theory; in fact, the objective of a scientific experiment is typically to acquire data that proves or disproves a very small hypothesis, a portion of an overall theory of how stuff works. The experiment may contribute to the development of the theory – Newton’s apple certainly contributed to his understanding of the law of gravity – but a single experiment or observation is typically designed to answer a very small, simplified version of more complex questions. If you wanted to understand a leaf falling, for example, you’d have to understand gravity, air resistance, wind currents, turbulence factors, and quite a few other things in addition to gravity. Trying to solve all those problems at once would lead to throwing up one’s hands in despair. It is easier to understand the problem by taking out the extraneous variables – so instead of studying gravity via a leaf in air, measure a cannonball falling in a vacuum. This requires making simplifying assumptions, but the simplification is valuable and can (eventually) be taken back to the problem of the leaf falling, as a portion of the solution.
That’s how I see sampling. The design problem I am trying to solve is extremely complex: how to design fabric to be layered on top of other fabric and then cut and sewn into a garment, so that the overall style of the garment (which also involves engineering design) combines with the design of the fabric to produce a specific artistic impact. It’s enough to make you throw up your hands in despair.
I’m approaching this through iterative design: answering small design questions en route to the larger ones, starting with the questions that are most important and/or which need to be answered first. The samples I’m building now are important for answering two questions:
- What kind of fabric do I want to weave? (drape, appearance, etc.)
- What design do I want to use for the front of the piece?
In this case, a really good answer to the question of designing the front would require sewing a mockup, i.e. a muslin, and designing based on a photo or surface design on the muslin. And in an ideal world, I’d have waited for the muslin to start designing the fronts. However, to have a muslin, you have to draft a pattern, and you can’t really design a garment without knowing the fabric – the weight, the drape, the fluidity of the material you’ll be working with. This means designing the fabric and weaving samples…and while you’re sampling for the type of fabric, you might as well put on some more warp and test out some design, colors, etc. Here I’m chaining two types of sampling together for convenience, because setting up a sample warp is very expensive, time-wise. It won’t be a perfect answer, but I think it will be an adequate one, and it will provide information I can feed back into the garment design.
That’s basically my approach to sampling: ask myself what design question I’m trying to answer, and then find the quickest approach to get that answered – not down to the smallest details, but in the rough – knowing I can refine my questions later. It’s an iterative process, and, while slow, is faster than trying to solve the whole problem at once.
