Monday, March 8. 2010Dasign: data driven, evolutionary, algorithmic & generative designIn generative design a designer does not create the final product but rather a system, algorithm or tool that in turn generates the final product. Some generative tools create many iterations and others lead to one product. The designer does not design the painting but rather the brush or the method. Instead of the still life one could develop "painting by numbers" for example. Take Joris Laarman. Joris created the Bone Chair using a software tool that simulates bone growth. GM created that tool to model efficient structures for its cars. This caused C. Sven Johnson to ask a rather pertinent question on twitter, "Who is the designer of the "Bone Chair"? And the question is not an easy one to answer. Is it Joris? The researcher that came up with the bone algorithm? The software developers that made the tool? The physical principles behind the algorithm? The scientists that discovered those principles? God or nature? All of the above? I for one have no answer. I do know however that generative design has a bright future. The marriage of tech and design is all around us. In a world where everything is designed a meta "way to design" that algorithmically cuts through the clutter is very appealing. A perfect design algorithm could potentially engender choice in design the same way that Google's PageRank set of algorithms do for the web. And this is what generative design already partially does. It simplifies design by codifying it and somewhere within lies the promise of "true", "simple" & "beautiful" design. With technologies such as 3D printing letting everyone design or co-design things there is also a real need for generative tools. They allow for unique designs but since each is machine made, the marriage is a conceptually comfortable and inexpensive one. Also, rather than forcing the customer into a "blank canvas conundrum" whereby the sheer possibility overwhelms them to the point inactivity, generated models could lead to choice or guided choice in design.
Boston based design duo Nervous System (who will be featured in a Shapeways interview soon) write code in Processing that makes beautiful rings and other items. Jessica Rosenkrantz & Jesse Louis-Rosenberg, who make up Nervous, even provide the tools they use to make their jewelry free of charge to other designers on their site. Belgian design duo Unfold (also the subject of an upcoming interview) who had a breakthrough with their ceramic 3D printing process recently also exhibited at Bits N' Pieces with their Brain Wave Sofa. In this design an EEG was used to "design" the sofa.
Jan Habraken (we will interview him also) & Willem Derks developed Chairgenics for the Bits n' Pieces show. For Chairgenics they created a chair genome and "bred" chairs in order to create the perfect chair. Singapore based software company Genometri makes technology that designs objects. Their Angel's on Shapeways for example are all generated & all unique. Mitchell Whitelaw also made the weather bracelet a bracelet that displays the weather of Canberra.
Michael Piasecki's Cellular Bowl, is beautiful & generated entirely in Processing. Justin Marshall's Generated ring is made with a suite of generated design software. Mendel Heit also makes lovely objects using Shapeways & Processing. Generative design has a myriad of applications. I anticipate that generative design will grow beyond the borders of the systems they are now. One the one hand ever more whimsical "data representations" will fall short of the strictest scope of generative design. On the other hand the nature of generative design will change. Whereas now the designer designs the tool that creates the product more involvement from "consumers" and more "intelligence" in the software will evolve. At the same time the "pure algorithm" will be ignored or augmented by the designer struggling to regain more input into the final object. To me the resulting movement will not be generative design in the purest sense. Rather it will be "Dasign"or data driven design. This design may be evolutionary, algorithmic or generative design but not necessarily rely on any of these for its core inspiration or make up. It will simply have the intersection of data and design as a identifiable characteristic. In the future I would expect to see many different data representations, Mii creators, "quiz design" and ultimately profile based design in a lot of different applications. Data representations could be any representation of data turned into an object. A medallion based on the types of sites you visit as per your browser history. An earing where the thickness of the heart is the number of times it has been broken and the width is the number of times you've broken others. A portrait that does not show you but rather your facial symmetry (or lack thereof). "Mii creator design" refines the sets of different generations of designs based on the users choices. Using the Nintendo Mii Creator it is already possible to with several generations come to a likeness of yourself or someone else. A similar system for bowls would quickly allow a person to, by simply choosing the design that best matched their preference out of the generated subset, come up with their designs. In "quiz design" you answer a few questions, "what's your favorite movie? your favorite song? Which picture do you like best? etc. Based on your answers the piece of software will then design the appropriate bowl. This might seem a bit far fetched but the necessary pieces for it are being built as we speak. Initial implementations would also not be that difficult but just have to have a dataset to start off with. This set could then evolve over time. Profile based design goes further still by assuming, based on previous web searches & purchases, choices and content enjoyed, which designs you would enjoy. This is a rather compelling technology for online retailers to put it lightly and they will drive adoption for it. If you are a designer you should to not sit still until Amazon or Facebook come out with the "you would like to create this" recommendation engine. I urge any and every designer to explore the wealth of possibility that is generative design, right now.
|
QuicksearchCategories |
Latest blogposts29-07-10 | DuannMusical Hacks & 3D Printing 29-07-10 | Duann 20 Material Options for 3D Printing with Shapeways 28-07-10 | Duann The Future of 3D Printing: Birds of a Feather at SIGGRAPH 28-07-10 | Duann Calling the Australian Shapeways Community 26-07-10 | Artur White Strong and Suuuuper Flexible. 25-07-10 | Duann Final Week of Draw Your Own Earrings Sweepstakes |
Latest forumposts3D ApplicationsRe: Sculptris - Drawn-Steel Hero It arrived! Re: Sprocket ring arrived! - stop4stuff Introductions Re: Hello. - artur83 3D modeler needed Re: Need a small part made. (See picture) - Drawn-Steel Hero It arrived! Re: Sprocket ring arrived! - Drawn-Steel Hero |
Want to stay updated?
Get on the list for our newsletter |
For example...
A tool maker, creates a tool that generates a leaf using only nature as a guide. The tool maker made no design choices because he/she copied nature completely. The tool maker owns the design of the tool, but not the resulting design. Nature owns that.
A tool maker creates a tool that generates a leaf, but builds in a complex array of parameters generally offered by nature (type of tree, weather patterns during leaf growth, percentage possibility of disease, amount of nutrients available to that tree, etc). Once again, the tool maker owns the design of the tool, but the resulting product design is not owned by Nature. The resulting product design is now owned by the end user who dictated the parameters.
A tool maker creates a tool that generates a leaf, builds in a complex array of parameters, but some of these parameters are pre-chosen by the tool maker ... let's say the tool maker creates a 3D model of the leaf, and this model is then modified based on the parameters chosen by the end user. In this case, the ownership of the final leaf is shared between the tool creator and the end user. I believe that since the tool creator started the process with their own 3D model, that they have a right to decide whether the end user has the right to share the content - but only if the end user is made aware of this decision.
-Whystler
As for the direction of generative/computational design, I see it more as a medium than a trend. It's a new, ill-defined medium that people are just beginning to explore. Frankly, I see data as becoming less important. Unless that data is relevant to the function of the design. Right now, you can generate a sense of novelty by creating something which you can say "this object is based on me," but I think as more of these sorts of products emerge consumers will realize that these sorts of changes do not affect the desirability of the final object and therefore do not matter. If I need a certain fit or if I like blue, that is a change that matters to me. But that kind of data is more like a shopping user interface. Having more choices just means having a better shopping experience. Even to the extent of being able to put your own text in something, it is just like one of those places where you buy something with your name on it but better.
I totally get it now, thanx! If I typed the above blog post in word pad or Open Office I would not be asking myself repeatedly "who wrote the blog post? Was it Joris or Open Office?" I do think however that such a discussion about authorship and contribution is very important to us. My judgment is completely clouded by the fact that I models such as this appearing on the site: http://www.shapeways.com/model/80951/hoodie.html
The Hoodie is created by Culprit Art, modeled by DotSan and the pattern is by Jon Burgerman. With such collaborative creations who is the 'author'? It will get more complex if we throw software tools into the mix. I think about this kind of stuff a lot, so this might account for my own child like fascination with authorship. This has fundamental implications for us when designing ways by which people can collaborate and also the intellectual property concerned. What stance shall we take?
Thinking of it as a medium is a great way to do it. "your name on it but better" is interesting also. Isn't the data crucial though? In the Cafepress: "put your name on a mug" the overlap between the imagined "word on a mug", the onsite representation of same and the final mug would seem to be crucial. But with more abstract representations is not the believability of the process crucial? The method rather than the result? If the algorithm leads to an ugly bowl I will not buy it in both cases. But in the abstract example if I do not believe that the process actually does refer to me I will not believe in the designer or the company selling the designs any more. On Cafe Press I could care less how they get my name on a mug as long as it looks exactly as I want it to.
In this light the believability of the algorithm, process and data is crucial in an abstract generative design process but negligible in a purely visual representation of what I want. This to me is the emerging gap between customization by choice and customization by algorithm.
what about this then: "[...]My judgment is completely clouded by the fact that I models such as this appearing on the site: [...] The Hoodie is created by Culprit Art, modeled by DotSan and the pattern is by Jon Burgerman. With such collaborative creations who is the 'author'?[...]"
Since I copied and pasted the above text that Joris originaly wrote (i assume). who is the author?
Joris is!. but I did edit it. So I can credit myself as editor.
But here's where it gets tricky for me (even when i used to write essays in Enlgish class back in college) .DotSans is mentionned by Joris. yet "DotSans" is a potential trademark, a brainchild of the person behind the 'brand'. Should he/she be credited every time the name is mentionned?
or is mentionning the name enough by itself?
what if applying my name on a ring, i apply a graphic i designed on a photoshaper with a frame?
frame was provided by the vendor, the image inside is mine.
who gets the credit?
This is why I like open source. (and COMMU part of the word communism
I think people and companies should stop being hung up on their 'rights' and focus on how they can better serve humanity.
Nice post, and with some great examples. I'd like to pick up on a couple of points though.
In the original article you say about generative design: "It simplifies design by codifying it and somewhere within lies the promise of "true", "simple" & "beautiful" design." It's the use of the word 'true' that bothers me, because it seems to imply that somewhere there is one right answer, if only we can find it. That's quite a Modernist idea, that there is one true design which will answer the needs of everyone; I think nowadays most designers understand that there are a lot of equally true solutions depending on the context. Unfortunately, right now, we're still hampered by Modernist means of production which means, for the most part, manufacturing lots of identical things. Rapid prototyping / rapid manufacturing offer the possibility that every product can be unique, but at that point it seems to me the notion of "true design" is somewhat meaningless. Hopefully simple and beautiful will still retain their power though.
I was interested as well to read Jesse's comments regarding authorship and what we expect the contribution of the designer to be within a Generative Design process. I think there is a parallel with the role of the artist and the artist's assistant, who will often be more technically skilled than the artist. Within the art world it's pretty much accepted that the idea is more important than the skill required to realise it: Jeff Koons' paintings or Antony Gormley's sculptures aren't criticised for having been created by their assistants. But the general public are often disappointed, they expect artists to be technically adept and feel cheated when they find out the artist didn't actually paint / sculpt / craft the thing that they put their name to. I wonder if the same thing will happen within design?
but in my opinion there's only one truth per individual.
and customizing will bring that truth to you, rather than dissolving it with mass production for everyone. Also, Joris was saying 'promise of true' which is not an absolute.
second point. I actulaly don't have a beef with Koons about what he does or does not do, my negative sentiment is towards him calling himself an artist, while what he really is, is a designer. In my opinion any 'commercial artist' is a designer - when selling your art is a criteria that you work with while creating the said art, then you're just designing to satisty the market.
again, this is just my opinion.
I meant "true" as Artur is using it. I don't believe that a generative tool could reach "true design" or "true beauty." The "promise of true" design, beauty and products is so seductive though. I know that many will try to reach it by creating tools and algorithms that set out to in vain create the "most beautiful" bowl, ipod cover or glass, either for one individual or absolutely. Imagine a tool that could create the definitive "most beautiful" version of any object. A rigorous scientific, mathematics based method to ascertain the most beautiful object from all the possible variants. No humans involved, no "eye of the beholder", just a calculation that proves that this one thing is the most beautiful. I believe that such a search is ephemeral and sets out for an unreachable goal. But, in an age of technology people will set out to vigorously search for such chimeras.
The Koons point is a wonderful one to make.
With regards to modernism. I do believe that we are at the cusp of a division within society between a neo-modernist techno optimistic view of the world versus a "post post modernism" view of the world where technology might exist but will be irrelevant for anything that matters. The "we will live forever because of science" on the one hand versus "the material, the advances of science and in the physical world is irrelevant" on the other hand. Reflexively one might think the latter is romantic (or what the heck neo-romantic) but crucially I believe that this movement will not look to nature for inspiration because the awe is gone. Something so malleable and fragile as we consider nature currently could not be as inspiring as it once was.
The analogy with Eugenics is clear in this piece, but I do take issue that the ideas used are also described as "Darwinistic" since it is an essential component of his evolutionary theory that random mutations are introduced, which are then tested by their environment. Both theories would be more clearly understood by making this distinction.
From today until the 11th of April we will be holding the Co-Creator Creativity in Co-Design Contest. We are looking for the most interesting & inspiring co-creators that you can come up with. The winner will be a fun, interesting, beautiful interac
Tracked: Mar 11, 13:02