[jmilloy]

These kinds of tools are mostly about the input constraints, heuristics, and weights. I'd love to see more about this with more practical constraints (rectilinear rooms and hallways, minimum area for cafetaria/gym, etc) and additional heuristics, such as window area in classrooms, number of hallways/intersections for ease-of-navigation, and building costs beyond just material use.

[riebschlager]

> building costs beyond just material use

Yep. The "optimized" floorplans would be an order of magnitude more expensive to build than the original. I mean, drywall alone would take a lot more time with all the cuts, mudding and sanding more joints. The thought of framing out these walls makes my head hurt.

[Semiapies]

And even if you had a building process that made circular/n-gon-based rooms practical, furnishing and laying out those rooms would be awkward. See the problems people have furnishing dome houses.

I'd least envy the people trying to arrange shelving in the library...

Still, interesting. I'd be curious to see results as he adds more practical constrains.

[burfog]

It doesn't look bad at all.

Dome home: most rooms will have 2 or 3 flat walls with 1 or 2 right-angle corners, so put rectangular furniture there.

Classrooms and library at the proposed school: the size of the room dwarfs the size of the furniture, so the problem is much like complaining that the letter "O" can not be made from square pixels. Somehow, we get our letter "O".

[TeMPOraL]

What about concrete 3D-printing?

(I'm not even sure if it's real or feasible; I just remember a bunch of pop news articles with a video of a large CNC machine depositing concrete to construct a building.)

[jethro_tell]

Concrete is manual 3D printing from 100+ years ago. =)

I've seen a concrete 3d printer, but it doesn't look like it would be structural. i.e. not sure how you would print around full lengths of rebar?

[saalweachter]

There's a YouTube channel, "My Little Homestead" where a family builds various house-additions and structures out of stucco-covered dirt-filled sandbags.

Dunno how that sort of construction technique pans out long-term or large-scale but you could get the sort of floorplans the GA generated out of it.

[theoh]

You have to build formwork to construct with in-situ concrete, so there's still carpentry (or equivalent) involved.

It's a good point that existing architectural concrete 3d printing creates "mass concrete" not reinforced concrete, but mass concrete can be structural.

[jethro_tell]

What sort of structure would you use a concrete without a tensile support like rebar or steel of some kind? How would that stand up in an earthquake?

[burfog]

Rebar makes earthquakes worse, as seen in California's freeway collapse. The rebar depends on the concrete to prevent corrosion, but that is only good for roughly 50 years. Rust causes expansion, which cases cracking. The outer concrete falls away, leaving rebar to get crushed under the load. The obvious fixes, like stainless steel rebar, have thermal expansion coefficients that don't match concrete, so instead you get cracking even without corrosion.

Lots of Roman stuff is still standing in areas that get earthquakes. The solution is to use a conservative design, with arches and thick walls. Domes are good. We can improve on this with 3D printing, using a structure like mammalian bone: solid near the surface, and spongy in the middle.

[theoh]

"Primitive" structures and other structures engineered to avoid tension in the concrete. Concrete doesn't have zero tensile strength. Apparently it's a few MPa, so a fifth or a tenth of the compressive strength. http://www.engineeringtoolbox.com/amp/concrete-properties-d_...

The large dome of the (Roman) Pantheon is unreinforced concrete. Most compressive structures built before the modern period (arches, domes, buttresses) were not reinforced with tension elements, just plain masonry.

[PhasmaFelis]

These guys are building small houses with a thing that prints in layers just like a home 3D printer, using mortar instead of plastic. https://www.youtube.com/watch?v=wCzS2FZoB-I It looks like there's manually placed metal mesh for reinforcement.

Dunno what the quality's like, but it's there. It does seem like it would be tricky to scale to structures larger than the practical footprint of their printer, but maybe the thing could be mounted on wheels.

[jethro_tell]

You could probably print in sections as well. This looks like it would be good for single story structures and possibly foundations/footings for larger 2 and 3 story structures.

[rzzzt]

Was this among them: https://youtu.be/DQ5Elbvvr1M

[derefr]

But would it still cost more if the building were made of prefab-stamped “IKEA pieces”, rather than stick-built? I.e., if you had to build the house 100 times (for a townhouse development, say), and so “factored out” the assembly of the framing and drywall and etc. for each curve and corner, would the labor costs then approach equivalence?

[jtbayly]

No, because those joints still need to be created somewhere in the process. It might be cheaper than the one-off build, but it wouldn't be cheaper compared to a more efficient design "factored out."

[dizzydot]

This is essentially what the author said in the summary.

[jmilloy]

Yep. Credit to the author. The page has been updated today since I wrote that comment.