Now you’ve lost me. That’s not the way to sell your product. Why aren’t you involving us in your process, by empathizing with our process? We are all your potential customers, right? What the boss really wants are happy clients. Clients don’t want second rate, product failures in the market place.
DID YOU REALLY JUST SAY THAT?
I know my boss sees design without meaning as a big FAIL. Design without meaning is design without value IMO. This can work for some commodity items, and short term sales, but real return$ come with meaning.
I see a lot of potential for Generative Design, but your argument here is not helping me defend it.
Design without Meaning sounds like you’re suggesting the branding isn’t important, and that would be a mistake.
Are we maybe hitting cultural differences here on some of these points?
This might be a tad off topic,
but I understand what SK means. What I’ve seen from my few internship is, outside of designers, people are more interested in margins and what colors thing are. And much less on the actual design of something. As long as the design isn’t hideous it doesn’t matter what the design is.
This may not be true in all cases, but I feel like this rings true in the corporate environment. Especially where people revamping essentially the same products every year. Something is “good enough” and then you move on. There isn’t enough time to really flush out all the details one would like.
Rather than take sides in this somewhat polarized discussion, I think I’m just going to go ahead an suggest a light read:
It’s an architectural book, and not a very heavy one. There’s many uses of metaphor and many points are very applicable to design. Ultimately, it does not separate meaning from form, or form from program, and I think that might be something that is beneficial to those who have made firm statements in their posts up to now.
If anyone is interested, I can quote a few excerpts, but it isn’t the same as reading the whole thing.
Its just another tool that aid in the design process. Besides, as of yet it shouldn’t be used to design the whole thing but maybe just parts of the object. Like grips, speaker grills or I see it especially pertinent in tread patterns in shoes.
Another thing is that most of us cannot sketch that stuff out when it comes to complex 3D patterns. So in some respects it is smarter then us.
Gets back to it being just another tool for designers and thats how it should be looked at.
Kinda like the blend tool in illustrator. Set the perimeters and let it do its thing in-between.
Im certainly going to use it as its an obvious trend.
Nice discussion. I have experimented with generative design off and on for quite a while now. It has an attraction that in someway a magic will be revealed in the vast array of possibilities that a computer can generate.
Designing snowboards and skis in the 90’s I realized that with loops and parameters, I could generate basically every possible nose and tail shape, after all, it is a finite set of factors, length, width, end flat, tip entry angle, tip bias, etc. you get the picture. About ten variables, in a physical area of 30x30x10 centimeters. All the convex curves that will fit in a bisected rectangle or cube. Write the program, press the enter key, thousands of shapes generated, a book, literally. So what was the result? The first snowboard professional rider that came in to design his very own model of snowboard, hand him the design book. Overwhelmed in the first second. Utility of a thousand generated designs, zero. The difference in shapes is too subtle. The conclusion I came to was that humans don’t see past a certain granularity, too smooth a gradation and we are overwhelmed by the choices. The end result in sports, humans like the product designed specifically for them, in all stages of the process, by someone that understands the subtleties and emotions involved as well as the mechanics. This value permeates down through the consumer experience.
Again, 1990’s, this time a golf club manufacturer, I was writing parametric programs for club heads. A seemingly definable shape in a limited physical envelope. An opportunity to generate a broad spectrum of shape possibilities. Culling and presentation becomes the issue. Humans are overwhelmed by more than ten choices. One hundred, one thousand, ten thousand, while all possible to generate, are lost on the individual. The client, golfer, user wanted a human designed golf club with a computer generated family progression.
The next set of 3D CAD tools allowed the same exploration. Pro-E and SolidWorks this time, again the result was the same. A multi-dimensional range of possibilities is difficult to navigate and differentiate, only by adjusting the coarseness of the steps and narrowing the range of possibilities does the choice become more obvious. However, they become without value. They can operate in functional parameters, they can all be attractive, fitting a certain segment of the user population, but they become without feeling. It is difficult, in my experience, impossible to convince anyone in the decision chain of the value. Personally as well, as the shape designer, there is no sense of accomplishment in picking from a generated spectrum. There is of course the satisfaction of having written a set of macros and scripts that have generated more work and actual designs than imaginable, but that has to go into the obsession box. Interesting, but not applicable to the the task at hand.
The concept of having a computer generate something that would surprise me still continued to intrigue me. Scripting POVRay into generating renderings of hundreds of shapes was a start, trying to tie those into industrial designs remained elusive.
Where the mind sees sameness of shape, wristwatches, snowboards, bicycles, mp3 players, etc, it seems natural to assume that it can be automatically generated. If one could “define” the morphology of a Ferrari and the same morphology of a Aston Martin, could a variety of shapes be generated based on variations and combinations? The answer of course is yes. However would any of those designs have cultural or emotional meaning? Here my answer is no, or highly doubtful. There is a vast population of consumers, some of which would no doubt find a generated result satisfactory. Resonance is the factor, emotional and cultural. While many designs made by humans lack that as well, the most resonant designs come from a skill and knowledge.
Where generative design has appeal to me, is where it removes itself from function and goes into small “a” art.
The following links are examples of personal experimentation with generative art, which were used as shape and form inspiration for industrial design as well. Not only was a shape exploration experimented with, but variations in color, with induced randomness to attempt to maintain the surprise. The images were generated all at very high resolutions in quantities of hundreds and culled down to what I deemed at the time to be attractive.
Some experiments with graffiti-like shapes and colors at high resolutions.
http://www.nxakt.com/generative-2d-design/Bench_Glyph_III
http://www.nxakt.com/generative-2d-design/Individuality_4
http://www.nxakt.com/generative-2d-design/Mix_II_349-391
Ultimately, after generating and culling thousands of variations and attempting to cross-generate the attractive decedents, the beauty/attractiveness became cheapened by the quantity. A single 1.5x1.5 meter print on the wall of a single image is attractive. The individual becomes lost in the choices.
My personal conclusion is that design should be guided by the hand, and generative design has its place in lower value worlds, virtual spaces such as games http://www.procedural.com/
Nice post, nxakt. Nice to hear from someone whos attempted to design whole objects using Generative Design and became, just as I had suspected, disconnected from the process. As others have said, I see a small value in little things, grips, grilles, 3D pattern exploration, etc. But to me, to design an entire object and just pick the best of 1000 variations, takes away from the design process as a whole.
nxakt, I am delighted to hear from some one who has taken the first few steps of playing with generative design. Thanks for sharing your work and your thoughts.
Generative Design introduces a problem that people have never faced before - too many designs that are too similar. But this is only the first phace. Sad to see the disillusionment it has caused. Perhaps you have given up too early.
Both problems that you mention have solutions:
- Similarity of Designs
- Selection burden
Good modeling practices will create much greater diversity than simple parametric variations. The trick is to create powerful genetic representations that are capable of creating very wide range of variations. A good example is the mammalian skeletal structure. We share it with frogs and whales.
The selection burden is a more serious problem. As you have learnt, it is a very very bad idea to show a client all the possibilities; showing them to the designer is bad enough. So the way to avoid this is to automatically kill all the designs that are visually similar. This can be done through many ways. This is an essential step towards making generative design a useful technology. Relegating it to art is a cop-out that I would avoid.
My conclusions are yet to be made; but if I were to make it now. It would be the opposite. I would say that it has no place in the lower value worlds - because, generated stuff devoid of distinction is just noise. If it has a place, it is in the higher value plane where human imagination has reached it’s limit.
Sounds like he’s been at it for a while SK, and that he’s taken more than “the few first steps”…
The selection burden is a more serious problem. As you have learnt, it is a very very bad idea to show a client all the possibilities. Showing it to them to the designer is bad enough. So the way to avoid this is to kill all the designs that are visually similar. This can be done through many ways. This is an essential step towards making generative design a useful technology. Relegating it to art is a cop out that I would avoid.
So now you’re saying that all those thousands of designs should be tossed anyway? If your software is creating massive amounts of similar designs, why can’t you tell it to toss them out? Just like you would in your mind while you’re sketching…
genoform - does toss them out, that is its distinction. Unlike the human mind, the granularity of search is something that can be set in the software. But the over evolved monkey mind has greater problems. Its hard to re-set. It conserves what it learns and gets attached to what it sees. It is aware of it’s limitations and works within it. Why do you think all this resistance to generative design comes from ?
Once it’s locked onto an idea or a work practice, it refuses to budge from it. Our attachment to our own creation is a serious impediment to wider exploration.
Perhaps that is one of the greatest strengths of generative design.
What’s all the hubbub? The OP has a CAD tool that is basically an automated sketch monkey. There is nothing wrong with that. I suppose it would be OK at generating some me too products. But anything beyond that the OP may be claiming is only superfluous hype.
Then again, one could be offended that this thread is essentially spam. But I am no moderator, I don’t police the boards.
nxakt, it’s great to hear first-hand from someone who has tried this in his work and your conclusions.
SK, I think a lot of the resistance happening is because you seem to think this technology is going to replace designers. When it matures, GD does seem like a tool that could be used for less important aspects of products, or to spark some ideas… maybe… but it’s not going to replace designers, just like CAD and CAD jockeys haven’t
Ok, great. But, where’s your proof of success in a product on the market today? The case-study we’ve been looking for? The “case study” on your website doesn’t really show much beyond, as I said before, 9 photoshopped images of 9 different bluetooth headsets. How is that a study of the successes and positives of generative design?
Honestly, I’m really looking to be proved wrong here. I see, and know, there are tremendous possibilities for generative design as a whole. However, I don’t believe that you can get a good outcome in terms of designing a complete product. Prove me wrong. And, I don’t want to hear: “Well, that ship sailed long ago, and the subconscious mind has difficulty adapting to change, so that’s where you’re getting hung up, but you should try it and see the possibilities, etc., etc.”
Is that too much to ask?
I would dispute its value in games or lower value worlds, in fact, forms of generative design have been used in the gaming industry to add replay value to titles at least back to the mid-90’s. Early examples include Diablo and the Age of Empires series of games that would, upon entering a new map or dungeon generate a completely different experience though layout, enemy placement, loot, objectives, etc. The designers set the parameters within the coding that generated these environments and “bred” out the generative qualities that made for poor/impossible gameplay.
Fast forward 15 years and game devs are bringing this in to more and more facets of a gaming experience. Valve’s Left 4 Dead series has an engine driving the game, which they call “The Director”, and it is responsible for creating a different experience every time you play. This deals mostly in placement and mechanics, unfortunately, after a while you realize that there’s maybe 10-15 different looking zombies…I would like to see them take it further and apply generative aesthetic design so that no two zombies ever looked exactly the same. Then there’s Spore, which ironically is a game about evolution from bacteria all the way through to sentient space traveling beings. The game allows you to randomly generate the animals in each stage of evolution and add traits, or you can use the game’s tool to design your creatures and as they evolve they change shape and color. Spore actually is really relevant to this discussion.
When you have studios like Infinity Ward knocking down $500M with the release of Call of Duty MW2 inside a month I don’t think you can really consider this stuff lower value, it is probably more applicable to gaming because there is no manufacturing of a physical object involved, it is more based on user experience and entertainment.
Good example Greenman, I’m trying to figure out why this is different for ID Maybe it’s because the value in generating new game content reaches the consumer directly? If my cell phone actively regenerated itself (or my car’s form) at the push of a button within it’s brand/model’s DNA contraints, that would be f’ing cool.
I’ve been contemplating for the past couple days whether or not to jump into this discussion because I find myself interested & intrigued by the idea of “generative design.” At the same time, I don’t think that this technology fits into the ID design process as it exists now, here’s why:
The fact that generative software is by it’s very nature, software, relies on the input of digital information, whether it be components, user directed constraints or a 3D model to iterate. The moment within the current ID process that most designers could take advantage of this type of “dumb iteration”, the work doesn’t happen within a digital space yet and many of these contraints are not yet known. By the time we are working in CAD we have eliminated many of the unknowns and the form already has intent. At this point we’re past the usefulness of “dumb iteration.” Right now, this is a downstream tool, that is only potentially useful upstream.
The problem really is that I don’t doubt this type of tool could create 100 good solutions if I give it all the right input data, but if I only have half that input data, I could still create 20 good solutions myself and work those 20 down to 5 great solutions while I fill in the missing data through iteration and feedback loops.
Right now, in regard to ID, generative design is a solution without a problem. I can imagine that soon, as new digital tools emerge to replace and assist our analogue tools (carving foam, post-its, napkins, etc.), this type of “dumb iteration” could find it’s place. I can’t know how our design process will change, and how technology will creep in, but there is no denying that it will. Does that make sense?
Maybe the the problem is calling it generative design. The word design implies conscious directed creation, not randomness, even if it is controlled and restricted randomness.
Perhaps it should be called Random Generative Creation, then?
edited because I’m “dumb”
Maybe part of it’s uselessness is the expectation of going from input to output, it’s probably good for assisting with form exploration, but I think it’s a little ignorant to expect being able to put in a bunch of constraints and out pops 100 good designs.
You can use Photoshop to do generative design if you want to. Scott Roberston was using a technique with Photoshop brushes to create forms. The brush tools have a large amount of settings that you can control. He did a demonstration where he crumpled up a piece of paper, took a photo of it, and then made that his brush tip, tweaked the settings to his liking and then started drawing shapes and blobs with the brush. From those randomly generated blobs he was able to pick out shape and form for things he was designing and then use illustrative brushes to add depth and tightness. Not too different from picking shapes out of clouds, the eye wants to make forms from things we see that lack definition.
Would you not consider Lego as a generative design tool with people compiling designs instead of computers?
Or is it the high amount of minute interpretations of data around constraints only possible with computers that make it “generative design” - variations on a theme?