5 ms·

So one framing for what's happening now is Parametricism, a term coined by Patrik Schumacher and associated with the style of his business partner and firm: Zah

So one framing for what's happening now is Parametricism, a term coined by Patrik Schumacher and associated with the style of his business partner and firm: Zaha Hadid. The basic premise of parametricism is that with fully digital design pipelines we have new ways of addressing the flaws in modernism while avoiding the aimlessness of postmodernism. In this view, architects are no longer designing surfaces or volumes, but rather painting parameters which are then utilized by algorithms and optimizers. This approach results in a style that has both unity and variation, continuous differentiation instead of the mechanistic repetition of modernism. Here's one dry video and two kinda silly ones that introduce the perspective: https://www.youtube.com/watch?v=vIcGTBRoKlg https://www.youtube.com/watch?v=vIcGTBRoKlg https://www.youtube.com/watch?v=Cfg_7ihB0YU https://www.youtube.com/watch?v=Cfg_7ihB0YU https://www.youtube.com/watch?v=Il41AlOBlgk https://www.youtube.com/watch?v=Il41AlOBlgk

kelseyfrog 3y ago
It's interesting that people are searching for the escape hatch, and I appreciate the videos, but I don't have much hope in parametricism. The reason I don't have much hope is that parametricism has arrived, it just isn't evenly distributed yet. One of the key ideas in the videos you mentioned was the digital tooling that enables parametricism. I need to point out that the same tools and ideas are already put to use in digital asset/content creation. We've already experienced the negative side of parametricism in digital content, specifically games that employ procedural generation. But shouldn't procedural[parametric] design of digital content provide endless variety? The reality is more complicated and this is where I have one reason to believe it won't, and one reason to believe it will. The reason parametric design won't provide endless variety is the Ten Thousands Bowls of Oatmeal Problem I referenced above. Just because each instantiation is different, variety must be semantically meaningful. It doesn't matter that each bowel of oatmeal has a unique combination of fruit, nuts, honey, and spices. It's that at the end of the day they are all just bowls of oatmeal. Real variety would be scrambled eggs, crepes, or breakfast in bed. I don't have hope that architects will be able to create salient variation ie: perceptual uniqueness. Adding more variational parameters doesn't guarantee any increase in perceptual uniqueness. In fact, it's an open question[1] as to what does. The reason I think parametric design could succeed is that the equation for parametric architecture is different than that of digital content. Manual curation, for example, is a huge cost compared to procedural generation, but a negligible cost compared to building a building. Curation may pose an effective filter against perceptual sameness. I'll likely remain a cynic until evidence that parametricism indeed offers a way out of the doldrums of the last 80 years of architecture. 1.https://knivesandpaintbrushes.org/projects/why-oatmeal-is-cheap/why_oatmeal_is_cheap_fdg2023.pdf https://knivesandpaintbrushes.org/projects/why-oatmeal-is-ch... (previous discussion: https://news.ycombinator.com/item?id=35735763 https://news.ycombinator.com/item?id=35735763)
- pixl97 3y ago
  The question is, do bowls of oatmeal provide the maximal utility? Or, to frame this a different way, "Just because you're unique, doesn't mean you're useful".
- jasonwatkinspdx 3y ago
  I don't think the bowls of oatmeal problem applies in the way you're thinking. Parametricism isn't just creating variety as filler, as we see with games that use algorithmic generation as a crutch. Parametricism is about exploring a design space created by rules and constraints in context. Love them or hate them, I think you can say that first, there is something that is a consistent style across Zaha Hadid's buildings, and that they're not just all minor variations on the same bowl of oatmeal.
  - kelseyfrog 3y ago
    Consistent styles will converge toward oatmeal with enough samples. The contradiction present in modernism is not resolved with parametricism. And switching from endless variety of boxes, separation, and repetition to folds, blobs, and swarms doesn't resolve the contradiction either. It replaces one pattern language with another. I'd love to live in a world with a thousand pattern languages instead.
    - jasonwatkinspdx 3y ago
      I think you're overfitting the oatmeal argument to the point of pointlessness.
twic 3y ago
So when i was building castles in VRML in 2000, and used PROTO nodes to cut down on repetition, i was doing parametric architecture? Cool.
- specialist 3y ago
  Just guessing: Based on the examples in the videos (linked elsethread), "parametrisism" looks more like NURBS for A/E/C, like if one designed buildings using Rhino3D. Versus "parametric designs", as I had (also?) initially assumed, which generalizes a class of objects by introducing parameters. For example, a "bolt" metadesign that can generate a specific bolt derived from the given parameters. An analogy might be "templates for 3D objects".
  - jasonwatkinspdx 3y ago
    It's kinda both. A lot of the reason Rhino gets used is the Dragonfly plugin, which allows for generative systems style creation of geometry. There's really two layers of "parametric" here, the basic being that CAD tools are now parametric, basically meaning drawings you can revise as needed, and then the second being how design philosophy reacts to this sort of software as a new tool, particularly when coupled to constrained optimizers searching through design possibilities weighted by a creative input.

kelseyfrog 3y ago

It's interesting that people are searching for the escape hatch, and I appreciate the videos, but I don't have much hope in parametricism. The reason I don't have much hope is that parametricism has arrived, it just isn't evenly distributed yet. One of the key ideas in the videos you mentioned was the digital tooling that enables parametricism. I need to point out that the same tools and ideas are already put to use in digital asset/content creation. We've already experienced the negative side of parametricism in digital content, specifically games that employ procedural generation. But shouldn't procedural[parametric] design of digital content provide endless variety? The reality is more complicated and this is where I have one reason to believe it won't, and one reason to believe it will. The reason parametric design won't provide endless variety is the Ten Thousands Bowls of Oatmeal Problem I referenced above. Just because each instantiation is different, variety must be semantically meaningful. It doesn't matter that each bowel of oatmeal has a unique combination of fruit, nuts, honey, and spices. It's that at the end of the day they are all just bowls of oatmeal. Real variety would be scrambled eggs, crepes, or breakfast in bed. I don't have hope that architects will be able to create salient variation ie: perceptual uniqueness. Adding more variational parameters doesn't guarantee any increase in perceptual uniqueness. In fact, it's an open question[1] as to what does. The reason I think parametric design could succeed is that the equation for parametric architecture is different than that of digital content. Manual curation, for example, is a huge cost compared to procedural generation, but a negligible cost compared to building a building. Curation may pose an effective filter against perceptual sameness. I'll likely remain a cynic until evidence that parametricism indeed offers a way out of the doldrums of the last 80 years of architecture. 1.https://knivesandpaintbrushes.org/projects/why-oatmeal-is-cheap/why_oatmeal_is_cheap_fdg2023.pdf https://knivesandpaintbrushes.org/projects/why-oatmeal-is-ch... (previous discussion: https://news.ycombinator.com/item?id=35735763 https://news.ycombinator.com/item?id=35735763)

pixl97 3y ago
The question is, do bowls of oatmeal provide the maximal utility? Or, to frame this a different way, "Just because you're unique, doesn't mean you're useful".
jasonwatkinspdx 3y ago
I don't think the bowls of oatmeal problem applies in the way you're thinking. Parametricism isn't just creating variety as filler, as we see with games that use algorithmic generation as a crutch. Parametricism is about exploring a design space created by rules and constraints in context. Love them or hate them, I think you can say that first, there is something that is a consistent style across Zaha Hadid's buildings, and that they're not just all minor variations on the same bowl of oatmeal.
- kelseyfrog 3y ago
  Consistent styles will converge toward oatmeal with enough samples. The contradiction present in modernism is not resolved with parametricism. And switching from endless variety of boxes, separation, and repetition to folds, blobs, and swarms doesn't resolve the contradiction either. It replaces one pattern language with another. I'd love to live in a world with a thousand pattern languages instead.
  - jasonwatkinspdx 3y ago
    I think you're overfitting the oatmeal argument to the point of pointlessness.

twic 3y ago

So when i was building castles in VRML in 2000, and used PROTO nodes to cut down on repetition, i was doing parametric architecture? Cool.

specialist 3y ago
Just guessing: Based on the examples in the videos (linked elsethread), "parametrisism" looks more like NURBS for A/E/C, like if one designed buildings using Rhino3D. Versus "parametric designs", as I had (also?) initially assumed, which generalizes a class of objects by introducing parameters. For example, a "bolt" metadesign that can generate a specific bolt derived from the given parameters. An analogy might be "templates for 3D objects".
- jasonwatkinspdx 3y ago
  It's kinda both. A lot of the reason Rhino gets used is the Dragonfly plugin, which allows for generative systems style creation of geometry. There's really two layers of "parametric" here, the basic being that CAD tools are now parametric, basically meaning drawings you can revise as needed, and then the second being how design philosophy reacts to this sort of software as a new tool, particularly when coupled to constrained optimizers searching through design possibilities weighted by a creative input.