computer generated form ideation

anyone have any thougts on the recent “innovation” of computer generated form creation.

there have been a number of apps recently sprouted, including on a core77 post today that promise to " explore more design possibilities than what was possible"…

in my opinion, i havent seen a single concept or development that has actually benefited from this kind of software. no doubt a real designer could come up with more workable, usable and realistic concept variations with a pencil given 10min.

so… who are these so called development tools targeted to? non-designers? engineering firms with no understanding of the design process?

Given the relative growth of this sub-category of software I assume there must be a market, but I am left scratching my head as to who…

anyone have experience with these “tools”?

I cant find the link, but i do remeber another demo showing suposed variations on a car design with 100s of wacked out positively ugle squished and squashed designs for a car…maybe a mustang of some sort…


I remember seeing that demo too. They looked like hideously malformed Ford Mustangs. I could see a tool like that being useful for simple things like vases or cups maybe, do one concept and a 2nd, then it fills in all the variations, sort of like keyframe animation, then you pick…seems to take the fun out of it though. Definitely not for shoes or cars though

good analogy.

they should rename the tool as Incetuous Mutation Malformed Design Tool.


I could see this kind of software being very helpful in generating different directions when things feel stale, or you just need another perspective. If it is used for a means to an end I see it being a disaster…but who knows…maybe a detail on a curve or a button that the algorithm comes up with might be that spark for the next great idea that you have.

I would never underestimate a different angle to a box.

My concern is more the effort that would have to go into generating these variations. Does it take 5 hours to model each variant and then 8 hours for the computer to chug on the variants? Not really efficient IF that’s the case.

I saw a demo, these softwares are basically designed to take advantage of parametric modelers.
You can make requirements based on volume, a particular dimension, an overall shape, etc. make multiple requirements and the software will explore a wide variety of permutations.

In that sense the idea could potentially “work” for cars or shoes, you just need to set the proper parameters. This software still needs designers to lay the groundwork, generate the basic geometry, set the basic paramaters, and finally select from the infinite number of permutations generated. Maybe we haven’t seen this thing generate good car design variants because a great designer hasn’t set up the geometry, set all the RIGHT parameters, and then choose the GOOD variants? I mean you could easily set this up so all it did was explore infinite edge treatments to a designated radius.

it does seem like it takes the fun out of it, but I can imagine creating some basic geometry and setting up my parameters before I left work to come in the next morning with a hundred formal variants to consider.

I think I see it more as an inspirational “here are some ideas you might not have thought of / didn’t feel like modeling” tool.
I think inevitably, you’d end up making changes to what ever variant was chosen anyway.
This also has the potential to waste that much more time wading through the numerous design variants with potentially minimal benefits.

agreed that perhaps some software could provide some interesting variations, I just highly doubt it would be something a designer never thought of, or is actually useful. not to mention the usually high cost for these kinds of apps.

in the end, i think it all comes down to the perception of design in corporation thinking that design is only a service that filters through multiples of options to fin the best.

sure, if it were as easy as having a computer “do design” then there would be a lot less crap on the market.

even more interesting, i think, is the difference in todays design process and development timelines vs. those of the past. while some industries surely have benefited from the advances in CAD and rapid protoyping i would argue that others have lacked in keeping up the accelerated timescale of todays design.

sorta off topic, but looking at todays car design where much of the design is compressed into a 1-2 yr timeline with much development in CAD and other modelling techniques, I would argue that a lot has been lost in terms of sheer design skill and refinement. I would take a 1960’s car designed over 5 years through clay and hand sketches anyday compared to the average modern car where every interior and exterior part seems to be put through the same “add 10mm radius” filter. Especially looking at some interiors of US cars I get the feeling a lot of design detailing is left to some automatic CAD apps and filters…


in the end, i think it all comes down to the perception of design in corporation thinking that design is only a service that filters through multiples of options to fin the best.

yes looking at the big picture, using a tool like this in a corporate setting could very well affirm the common misperception of design as service.

I am not a staunch supporter of these apps at this point,
but I do try to see the potential they may offer even if just for purely speculative purposes

i think it would be a very useful application for generating optimal solutions for engineering or mechanical parts. i think there was a feature in popular mechanics a while ago where, a similar tool was used to generate a “design” for a camera lens of sorts and there was also another program that i watched where a computer generated random iterations of a walking robot with certain parameters put in and each “generation” of the robot took successful traits from the previous generation it was pretty cool, though not sure that it is the same type of program being talked about here…

I can’t imagine ever needing such software. It presumes that I don’t understand the problem I’m solving.

Could you imagine writing a song or a book this way?

What’s that quote, a thousand chimps typing on a thousand typewriters, how long before they write Shakespeare?

Funny, wikipedia has an article on the “Infinite Monkey Theorem”

Brings up a very good question though…

Can one obtain great design through brute force?

ie, 1000 designers with 1000 pencils doing ideation. 1000x1000 concepts is a lot of ideas, surely there must be least 1 great one.

It’s an interesting idea, but I can’t see it being truly useful for designers or engineers. At the best it would give the impression that is ‘applying design’ to a form, but this of course assumes that design is superficial layer added to ‘stylize’ a form.

A few years ago I imagined a slightly more sophisticated system, where a form could be merged with style-set chosen from a long choice of pre-designed style-sets. Like brand DNA or a VBL, each ‘style-set’ would contain information about how to treat surfaces, radiuses, etc according to their orientation or priority. These style sets would be created by designers. This system would be less random than the SW system described above, but still has a basic underlying flaw. If such a system were to exist, then the style-sets or approach would become regarded the new ‘generic’ or cliché style, just like the hideous style-sheets that come with PowerPoint or Word (I just shuddered).

The results would be generic at best and would distract the designer from creating a truly good design.

The problem with the infinite monkey theory is that you need an infinite number of humans to find out if one actually wrote Shakespere. Sames goes for software.

The monkeys tend to dribble over the keys too. Its messy.

Interesting idea. I think Karim Rashid must be using that system. Style set “blobby” and pink with graphic + signs printed all over everything.


I asked this a few months ago but didn’t see a reply. Anyone out there actaully seen the MS Vista HARDWARE design guidelines? This style sheet was apparently released to hardware companies to “inspire them”

i’m really very curious…

off topic?

in the end, i think it all comes down to the perception of design in corporation thinking that design is only a service that filters through multiples of options to fin the best.

I hate to piss in everyone’s cornflakes…this software/discussion IS about finning the best solution out of a thousand iterations.

Nobody is to blame except Designers themselves for “The Man” thinking that this is all design does . If the only thing one does in a design capacity is render-monkey their way through the day, or repetitively design unrealistic “blue-sky” concepts, then you have shot yourself in the foot.

This software is another tool. A tool like Pro/E, or Alias, etc. Nothing more, nothing less. If what you do day to day is generate styling options, then this software threatens the foundation of your profession. If you bring to the table something more then it is nothing more than a tool.

I also balk at the elitist thought that only a “designer” can generate good ideas or forms. We’re in the business of risk reduction. Of course a million Mechanical Engineers can pound out The Eames Chair eventually, but someone trained and experienced can do it faster and significantly reduces the risk (AKA $$$) of getting it wrong.