The new architecture wars
aeon.co
aeon.co
We have already figured out how to build beautiful buildings in any number of styles from many eras. That's because a building is a discrete entity whose design is managed by one group of people, who can form a relationship with an architect who pours their heart and soul into the design.
A street network, on the other hand, is managed by successive governments staffed by much more disinterested operators, and often by different governments who may be working at cross purposes. I'm not saying there should be a master plan (I actually don't like those) but rather that the logistics of designing a street network are much, much harder than for a single building, however you go about it.
I also consider the harmonization of automobile and pedestrian traffic to be an open problem. Maximalists on both sides just talk past each other. Car-oriented development is bad, but at the same time we aren't going back to a pre-car world.
Whether your model is Brasilia (modern) or Cayala (traditional), there are just so few opportunities to actually think about the design of the street network of a city. Street networks are just taken as given and then architects try to make the best of them. That's not a criticism of architects or city planners; the street network design largely seems to be on autopilot with very few opportunities for a change in direction. And for established cities, changing the network is nearly impossible.
One of those roads splits my neighborhood into two clear halves with each half only access by one artery, so that commuters don't go gunning through the neighborhood.
The system is great for directions and general movement of people.
https://en.m.wikipedia.org/wiki/Indus_Valley_Civilisation
We had IVC as a topic in school, vaguely remember it being mentioned, could be wrong.
Of course we're not going back to a pre-car world, but I feel like even advocating for relatively minor changes like protected bike lanes or limiting car access in city centers gets you labeled a maximalist.
Drivers have a countrywide pedestrian-free transportation network (the interstate highway system) but there is practically no place at all in the US that you can walk or especially bike for transportation without frequent potential conflict with cars. And even small steps towards establishing such spaces meet fierce opposition.
This is a great observation, and it immediately suggests a vision: An interstate pedestrian path system, and an interstate bicycle network.
The Park Service is already imagining these things, but many of the "trails" are just marked sections of automobile-road. They don't have the resources to go further, but we should go further.
Define minimum standards. The most important is physical separation from vehicle traffic. Mere paint or bollards is insufficient. Ideally some limited infrastructure like occasional benches and water fountains. Fire Robert Moses and put Charles Olmstead in charge.
The funny thing is, when you create nice environments, people want to be in them. A market is created. All of a sudden, cyclists are passing through. I can run a little cafe that caters to them. Commerce happens. Yes, much of it's leisure-oriented, but that's still GDP.
Yeah, that would be a nice part of it. And indeed, the Netherlands and Belgium have such things: https://en.wikipedia.org/wiki/Bicycle_highway
Of course, my reference to the interstate highway system was mostly meant to illustrate the current imbalance. But just as drivers need to leave the interstates to get to their final destinations, so too would cyclists. So it's still important to have safe walking/cycling routes on local streets too, because ultimately that's where we all start and end our journeys.
It won't be possible to completely separate people/bikes/cars in all of these places, but thankfully there are a lot of designs around the world that we can take inspiration from, and with careful planning you can prioritize cars and bikes on different streets so they don't interact as much. But in the US, we're basically not giving anything but cars a serious chance at this point.
But you can humanize the grid a little:
https://tilings.math.uni-bielefeld.de/substitution/pentomino...
https://tilings.math.uni-bielefeld.de/substitution/domino/
>I also consider the harmonization of automobile and pedestrian traffic to be an open problem. Maximalists on both sides just talk past each other.
In its early days, StrongTowns made a big deal out of arguing that parking should always be put behind commercial buildings instead of in front of them, so the storefront reaches the sidewalk. They've spent less time talking about this now because it attracts vitriol from the people who want to get rid of parking altogether. Unfortunately, it seems like the loud people have found the urbanism discourse.
But grids themselves aren't necessarily bad. One can imagine a street grid as a trellis, on top of which life is able to grow in abundance. Alexander and Jacobs both didn't seem to have a problem with grids and right angles, only with poor uses of such.
For example, Jacobs talks about long blocks which prevent cross-mingling between parallel streets. She also describes how staggering streets with dog-legs and other obstructions can improve neighbourhoods by slowing traffic.
> the street network design largely seems to be on autopilot with very few opportunities for a change in direction
You can't have it both ways. The solution to a tragedy of the commons is to manage the commons. How can you say that you dislike master plans, but you want a planned design!? Architects pouring their heart and soul into a building's design is a good thing, but city planners pouring their heart and soul into a neighborhood's design is a bad thing!?
One reason it's impossible to do anything is because large-scale planning is unfairly villainized as "inefficient big government intervention", and here you are villainizing it, while simultaneously wishing for intervention (from, who else? Government!)
Edited to add: Having read so much Hatherley over the years I can almost guarantee that a name drop in the conclusion like that means he's currently developing more scholarship / criticism on the subject...
Architects, we have to admit, are under-performing in the metric of "creating new architectural styles and schools of thought" and over-performing in "number of uninspiring designs."
I have to conclude that either there are systemic barriers in getting out-of-distribution buildings built, or something about the pedagogy of architects that filters or hammers out out-of-distribution designs. Either way, we're suffering from a lack of architectural imagination.
When I see neo-traditional examples[1], I can't help but be jealous we, in NA, aren't teaming with our own. Instead we've got architects churning out designs that fit squarely in confines of The Ten Thousands Buildings of Oatmeal Problem[2]. Taken as a whole, the parts and their interactions, are a system for producing architectural oatmeal.
1. See https://en.wikipedia.org/wiki/Neo-Andean
2. https://golancourses.net/2018_60212f/wp-content/uploads/2018...
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
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-ch... (previous discussion: https://news.ycombinator.com/item?id=35735763)
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.
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.
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".
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.
I wonder when buildings were built as you describe; maybe the 19th century? Isn't the 'golden age' romantic fantasy obvious?
> The bulk of the buildings in El Alto are simple, unadorned brick structures; Neo-Andean structures are typically similar in underlying structure, but with the addition of very elaborate decorative facades.[1]
That's to say that it is not "new", or even architecture for that matter, but an artistic veneer over what is still traditional architecture.
In other words, in the entire history of ever, no one has ever been right when they've called something "not architecture".
It clearly fits into Alexander's pattern language in the same way that neighborhoods, walkways, form, ornamentation, and doorknobs do. We must not confuse ourselves into thinking there is good-architecture and not-architecture.
The glasses on my face are not architecture, nor is the puddle of water outside my window. While they may have been "architected" in a sense, at a certain point we must call a rose a rose and mean a rose if we are to have an intelligible conversation.
I don't see a problem, except in the unimaginative repetition of obsolete styles. There have always been lots of uninspired buildings, just like there have been uninspired songs - the pop music of 50 years ago wasn't better, it's just survivor bias.
They are obsolete now because they were designed for a certain situation: The affordable building materials, skills, engineering, etc. If you want to follow tradition, and innovate new designs now for the new situation - that's what they did, and back then people all complained about anything new, just like now.
We were a culture that embraced innovation, now we're actively killing it, IME, as people seem to have embraced thoughtless, irrational conservatism. (General thoughtlessness - an acceptance of the abandonment of reason - is killing it more.) It's socially acceptable to make these statements, people will even cheer you on, so people make them. You can see the almost standardized rhetoric, including the definitive language as if there is no other possibility they can imagine. If even HN doesn't embrace innovation, who will?
Why is new styles a goal? Shouldn't the goal be a building that is functional, beautiful, appealing to live/work in, appealing to live near, efficient in its materials and design, eye-catching, yet appropriate for its surroundings? None of that requires "new". Architects that are caught up in having to be new make the most monstrous creations of all.
- Roofs are flat. HVAC equipment is put on the roof. This allows a maximum use of zoned floor area.
--> Solution: write a roof/attic allowance into zoning codes.
- Trees are too small. Developers want to avoid maintenance difficulties with trees, so they plant dwarf varieties that don't provide shade.
--> Solution: fund tree management as a public good. Offer incentives for preserving existing trees.
- Buildings are surrounded by Stepfordesque hedges that look unnatural. Green space is crammed into noodle-shaped setbacks that are too narrow to support anything like a garden. Mulch is everywhere.
--> Solution: replace setbacks wherever possible with access easements and open space requrements.
I couldn't care less what the big famous buildings look like. I don't spend much time near them anyway. That's a high-class problem. No wonder it's a duel between the right and the yuppies.
As to why flat roofs are ugly? I'm not sure, but I'm guessing it has to do with flat-topped things being uncommon in nature. It's a structure made possible only by decently sophisticated engineering, what with the stress on corners and the need for drainage. It makes the occipital lobe think it's missing something.
In some city, I forget which, I saw a guide to street trees that said that picking the correct trees solves the root / maintenance problems: roots grow neither too shallow or too deep, etc.
> I couldn't care less what the big famous buildings look like. I don't spend much time near them anyway.
They are part of the visual environment that we live in. The Empire State Building in NY is in the visual environment of millions who don't go near it (in fact, it's hardly noticeable close-up - you mostly see the bottom floors).
Such pubs tend to be part of (run-down) housing estates (what might be called 'Projects' in the US), with attendant clientele.
It's a good rule of thumb if you are unfamiliar with an area!
>Traditionalist and modernist architecture are both mass-produced, industrial and international. Is there an alternative?
Alternative:
>When the historian Frampton called for a ‘Critical Regionalism’, a modern architecture sensitive to place and materials, he found an example in the work of Álvaro Siza in Portugal.
Rebuttle: regional/vernacular designs almost always involves more labour, which is expensive in advanced economies. Small countries/markets can't rationalize engineering entire regional specific building techniques and validate materials. Large countries/markets have multiple climate zones. In the end mass construction will fall back on value-engineered mass produced industrial building components.
If I am ever able to acquire any significant resources, I would like to pursue a new type of city. This city will use technology like open source design software, formats and blockchain to create an integrated environment.
Outer buildings will be public containers for private modular internal buildings. There will be multiple levels of infrastructure including separate levels for roadways and pedestrian walkways. The standard vehicle will be a small autonomous pod.
The outer buildings will be connected on multiple levels by roadways. The architecture will be dictated more by functional needs such as passive or active climate control than by aesthetic.
Here's more of what he's written, in case you're curious like me.
on topic: i think it's a hard thing to reconcile and/or combine and/or make a "building with identity" while being able to accommodate >1000 people in a building. Poundbury seems to be a good example, though in singapore, economically speaking, it's better to have an apartment than to have a house on a ground.
https://commons.wikimedia.org/wiki/File:Geisel-Library.jpg#/...
However, I wouldn't mind seeing more Victorian and Queen Anne homes being built. I realize that because of their enormous level of architectural detail that economics will dictate that they'll be built cheap, bland, and uniform. Which isn't the point. But at least they won't be faux Craftsman or modern Farmhouse. So tired of seeing those.