I’m one of the creators of WikiHouse project (and lead founder of WikiHouse Foundation) – @ben_howes (http://twitter.com/ben_howes) gave me a nudge and suggested we throw some thoughts into the mix on this (it is Hacker News after all). Some really great points made here - I’ll focus on a couple and try to keep it brief.
The WikiHouse project began as just an experiment into digital, distributed housebuilding, but has turned into a serious open source project aiming to change the way building information works. We’re doing it the hard (and slow) way – so currently busy working through certification, regulation, liability, commercialisation etc. It’s already in use in various pilot projects by a bunch of teams around the world already, but as you can imagine, change happens slowly in this space.
In a nutshell, the advantage of digital fabrication is similar to any fabricated components (think IKEA). It brings speed, value, precision, performance-predictability, safety and mistake-proofing (which lowers skill and risk thresholds for small players) to construction projects, plus it’s editable as code. The opportunity we see is to push this from big centralised factories to distributed networks. The lead technology, WikiHouse WREN, is just intended as a starter for ten really. We started designing for machines and materials that are pretty widely available (Plywood, CNC), and then invited others to fork and edit, which is what’s happening. It is strong, buildable, airtight, performs well on cost etc, but yes, at the moment, it is limited in terms of span, height... Please do fork and improve it, or get into developing other open building technologies.
As a couple of comments mentioned, the key issues are around the limitations of sharing static design files, because, yes, price, conditions, regulations etc vary depending on where you are, and yes, there is a legal risk of any specific design shared being interpreted as professionally certified, and therefore exposing professional contributors to risk. So, over the last few months we’ve been working on some browser-based generative software for designing and delivering homes digitally without relying on external proprietary software like Rhino and Sketchup, with support for distributed supply chains and with checking protocols built-in. We’re hoping we’ll have some stuff to release before too long for people to get their teeth into, but as you can imagine, we have to moderate our excitement with a degree of sensible caution and thorough testing. But as a rule, we’ll be open-sourcing everything we can, as soon as we can.
Thanks for posting Alastair! What will be the status of your Rhino/Grasshopper process as you move over to your browser-based generative software? Will it still be integrated and supported, or are you phasing it out?
I ask because it is beneficial to have it integrated with Grasshopper/Rhino (or Revit/Dynamo) for not only the generative design stuff, but also the simulation and optimization plugins. i.e. Energy/environmental simulation, multi-objective optimization. These engines are starting to be ported over to the web, but for now Grasshopper is state-of-the-art in terms of design/simulation/optimization integration.
I'm interested in forking and playing with these kind of features, but would be prevented if this moved over to a browser-based software...
So much time and effort has been put into applying technology into making home bigger and more luxurious, all to further elevate a price point. It needs to stop. Use the technology to lower the price point and make the buildings more efficient.
You read the page, of course, so you can understand how I'm having difficulty with the point you're trying to make given that the only house plans I could find on the site are for a house that could be built in my garage. I put some honest effort into looking, but could you point me to the plans for a "bigger and more luxurious" house on that site?
OK for all those asking about UK building code. It is pretty strict. Thermal insulation requirements are really high. Structural stability and fire resistance are also tough. Yes a UK roof can hold snow! In fact they hold a massive weight of tiles normally.
We don't really get tornado's or earthquakes however. we do get winter gales regularly.
There is a huge need for cheaper, and more efficient (in terms of energy needed to make the materials) homes here. Due to aesthetic requirements (and as a traditional way of making a house that will survive the damp in the UK), most homes are still built with fired-clay brick skins forming a cavity wall, although the inner masonry skin is often a prefab timber frame these days. The roofs are usually clay tiles (or concrete copies if local planners allow) or natural slate. The foundations are a solid trench of 1 metre deep concrete (more if the ground is soft). Ground floors are usually thick concrete onto thick hardcore bases with insulation incorporated either in the concrete or under the floor covering. The whole thing is massive, heavy, concrete and clay focused and expensive. Of course land prices are high in the UK so the resultant houses are really expensive no matter what. There must be a more efficient way of building.
And by that I mean, you have spent a good deal of thought on a pretty marketing page, but your spiel doesn't really demonstrate knowledge of the wooddust and nails that make up homebuilding. You're not being careful with which codes you satisfy, you're not talking physical tolerances (earthquake? hurricane? heavy rain? wind?), you're not discussing the existing prefab systems and what makes yours better.
There is, of course, potential in prefab. But I would suggest more sweat of talking to carpenters and other tradespeople. Hiring a top carpenter for 6 months to throw darts at the thing would be useful. willyt provides some good feedback, but I think its revealing that you don't treat that kind of information, in that kind of depth, in your current information copy. You can get there though!
Seems like they're aware, if not super upfront about it. Under the R&D tab, the #1 challenge listed is structural testing and documentation.
Quote: "In order to make it easier for engineers, building control officers and mortgage companies to do structural calculations and to certify WREN in projects, we need structural engineers to do further R&D, structural lab testing and documentation of the system’s structural performance."
So a couple of points in reply to your comment:
This looks to be an open source development model, and so appearing to be professionally designed may help attract the needed expertise, instead of having said people get the impression the project isn't going anywhere and not see a point in joining/helping.
As for carpenters/tradespeople - and I've worked as a residential carpenter so please don't take this the wrong way - but from what I can tell this project is pushing the mold enough that they need structural/materials engineers more so than tradespeople. Tradespeople have a lot of practical knowledge about how easy/hard it will be to build what the engineers put in front of them, and that knowledge is valuable, but to use a car analogy, you wouldn't ask a mechanic to design you an electric car. And sticking with the car analogy, this project looks far closer to an electric car than your standard internal combustion engine car (ie standard 2x6 framed pine style construction, which most zoning laws assume)[0].
[0] In America. I see this projects hails from the UK, dunno what their zoning laws or standard construction practices look like.
In manufacturing, the best way to build a device for effective fab on the factory floor is to have a senior assembler involved in the creation. They will point out amusing errors like "You can't run wires there or no one will be able to build it", or, "If you rearrange the widget over 2 inches, it'll cut time to put the widget in by well over half". Obviously they don't run the show, but they should be involved in the entire design-build phase to ensure the effectiveness of the system when the rubber is deployed on the travelway.
Physical product development is an entire skillset in and of itself, and the skill gaps are pretty big when you're coming in from the software world. As I'm sure you know. :)
The question with 3D printing is around the materials and the build times. Wood is a great low carbon material, whereas printing in concrete is a different matter.
A low carbon successor to concrete could change this if its low carbon and not so dense that it's difficult to transport (like concrete)
Lego style building recently started to look pretty good. It also has potential of being extremely cheap. It may be even possible to do rearrangements (moving, shortening, enlarging walls etc) later if you feel like or reusing your old house "bricks" to make new one.
This is actually a very good idea. I would personally love an easy reconfiguration that I can pre-design myself: a young kid --> cut off a master bedroom chunk for a semi-detached room; older kid -> his own room, (teen -> full enclosure, enter from the roof :) )
If it is relatively quick/automatic one could reconfigure before/after visiting friends.
For that to be cost effective, I think you would need a pool of other house builders building the same house who need the same parts. Maybe a kickstarter-like project.
If anyone is curious on how to build a house I urge you to volunteer at the local Habitat 4 Humanity. Not only will you help someone build a home but you'll learn loads about the home building and permitting process. It seems a lot of people in this post are overthinking things.
The local school has been working on building a Wiki House for two years now. They struggle with issues around the CAD drawings and getting them into the system their CNC router uses and getting parts made with accuracy.
The thing that I noticed is the material waste is pretty high. Getting things fit to 4'x8' sheets of plywood in the US has lots of waste in the sheets. When you stick build a house, most of the materials are used, with a little careful thought things work out. With the long arches, only a few fit in the sheets.
Buildings are a really complicated mix of design, engineering, regulations and regulated trades. This looks like a nice student concept project with some great graphics for the assembly drawings. Here are some hopefully constructive comments.
Structure. What is the advantage of all the machining that is being done here? It looks like there is not enough structure? I'm not convinced that the clip together frame screwed together on site from 18mm ply will form a convincing portal frame? That stuff costs about £40/sheet what happens to all the wastage from the machining? Once these are assembled and filled with insulation they will be pretty heavy to lift into position without a crane which is a safety/legal issue, so if we need a crane, what is the advantage over other cassette systems like dan-wood, hof house, cross laminated timber or even just prefabricated stick frame panels? A ply/foam sandwich could make for a pretty rigid structure (think surfboard), but the insulation is loose filled here; opportunity missed! I note that Arup are listed as partners in the project, they should be able to advise on this.
Design. Its a massive amount of effort expended on developing a new construction method to create a 40m2 one bedroom single storey house like many others before. Why is it better to build like this? What is the maximum structural span of this machined plywood slot together construction? What advantages does it give me if I want to have walls at angles other than 90 degrees or rooms wider that 4m? How high can I go? It seems like you have chosen eco friendly materials but single storey construction is fundamentally not environmentally friendly because the density is too low to support public transport and efficient public services provision.
Energy. They are proposing a Mechanical Heat Recovery unit and very little heating which means they need airtight construction, but the airtightness membrane is very close to the inside of the house and will be easily damaged when the electrician and plumber fit the services. Airtightness is critical. A more robust design would rely on a taped board as the airtightness layer, preferably somewhere in the construction where it won't be damaged by someone putting a picture up. The integrity of the vapour check layer is also very important as any significant interstitial condensation will degrade the plywood structure quite quickly. Loose fill insulation is notorious for slumping and leaving cold spots that go mouldy.
Cost. Double it. E.g. £1000 labour allowed for the foundations installed. That pays for 4-5 days of a 1 skilled labourer's time, it's not enough. There are no plant hire costs in there either. Also only £1500 allowed for the electrician!
Legal. Who gets sued if this design goes wrong? Who pays for the insurance? You have a disclaimer, but if any of the designers contributing to this are architects, engineers or other construction professionals the construction case law in the UK suggests that a professional can't blanket disclaim all advice given in their professional capacity as potentially incorrect. So any professional contributing to this could still be held responsible for design failures. It's rather terrifying what has been blamed on architects giving free advice off hand. I don't know if the creators of this are architects, but bear in mind that Architects in the UK are required by law to have adequate insurance to cover the full cost of any mistakes they may have made for undefined period after the building is completed I don't remember the exact figures but you are generally advised to carry insurance at least a decade into your retirement. Hence no professional will contribute to this website for free unless you find a way of guaranteeing that they are insured as otherwise it will have an unknown ongoing cost to them into the future.
These are very intelligent musings. I take it that you have some relevant knowledge. Thanks for engaging :)
I'm interested in this stuff (to the point that I've rented a shipping container and plan to non-destructively press-fit a frame and document that process), and yet find the conversations like in this thread don't tend to happen in the open. Building and construction is arcane knowledge. Especially compared to the openness of software knowledge.
Bringing those conversations into the open seems to be the value of a project like this, imho
Many houses in the UK use prefab timber frames and or heavy roof trusses. A crane is a perfectly normal requirement.
The airtightness requirement in UK is not that hard to pass. When the airtightness rules came in it caused a lot of headaches for cowboys, and was a doddle for well build dwellings. The vapor control layer needs to be right on the inside as it is supposed to stop vapour from reaching the structure. On a normal timber frame it would be right behind the plasterboard, right where picture hangers will hit it. Obviously a few holes are not hurting the bulk of buildings out there. Loose fill insulation is very widely used and doesn't have to slump, obviously everything could be made badly and installed badly.
I love to see new ideas for building. Personally I can see the future being about plywood panels laminated to foam in ready made units, rather than this, but good luck
From a quick skim of the site, seems like the intent is that the machined parts would eventually be prefabbed at quantity and you'd purchase the required parts. Compare "machining a single Lego out of plastic" vs "buying a bag of 100 1x2 plates, and one of 100 2x6 bricks".
I see it becoming Github for container house manufacturers. Put some "community" level designs here to market yourself and your "enterprise" level designs or finished products.
Building houses is hugely dependent on local regulation, building code, and also climate. This project has home in the UK, so it probably is adapted to UK regulations, weather etc?
Right. I'm in Finland, and would have the same concerns, with tornadoes replaced by regular extreme changes in temperature and humidity (well, I guess someone building in UK at least has thought of the humidity).
Anyway I'm pretty sure these houses would not be permitted here due to not having enough insulation, so that they need more heating energy than is the norm, thus do not meet our CO2 reduction targets.
I'm really glad github is being adopted in architecture/engineering. That being said, I've contributed to a handful of popular, open-source tools in architecture, and the big problem for us is that architecture depends on large binary CAD files that don't cooperate with github.
The plan with wiki-house is that you'll be able to to draw a plan and have software generate the structure for you. A few years ago they were working on a plugin for sketch up to do that - not completely sure what happened to that!
I remember checking this project out around 2014 and it seemed it was all being done in Sketchup using ruby. Now it looks like it's mainly being done in Grasshopper/Rhino (visual scripting) with some components being done using custom C# or Python scripts. I think its the right move, as Grasshopper/Rhino is a more powerful CAD engine, and has a more useful and extensive ecosystem of plugins for them to develop around.
A friend of mine - a long time ago - told me of a concept he had that he termed "Mennonite-in-a-Box" - basically, some kind of machine that you could load a design into, and it would construct a home on-site. He never elaborated more on it, but it got me to thinking. Unfortunately, I am not someone who ever has built a house (nor do I have the money or time to experiment on such a scale) - so my thoughts never went anywhere either.
I ultimately came up with a couple of ideas in this regard; one seemed potentially doable, and the other seemed absolutely doable. Both rely on the idea of the main building material being plywood or OSB - like 1/2 - 3/4 inch thick 4x8 sheets (or the metric equivalent).
The first concept was something that a "do-it-yourselfer" could own or rent. Essentially it was to be a plotter you would attach (in some manner) to the sheet of plywood, and it would "plot" the outlines of the cuts and such on the wood (at the time, I envisioned an ink-jet like approach - today a small LED laser engraver might work better), with annotations and other such things to guide the builder in fabrication and assembly. These annotations could also refer to areas in a reference assembly diagram or software that could take the user step-by-step in cutting out the pieces (using various tools - a jigsaw and circular saw mainly) and assembling them (perhaps with 3D views, audio, graphics, etc). The plywood pieces would also be numbered or otherwise noted so that each is used in the proper order. The whole thing could be sold as a simple robot, perhaps attached with cords or something to the wood, in a case with software - user supplies a laptop (or perhaps it comes with it?).
This machine actually does exist - in a way nowadays. I know of a CNC router that attaches directly to a piece of plywood, which could cut the parts directly; however they wouldn't be numbered (but perhaps they could be engraved?). There's also another tool that is something like a "user guided" CNC router; the user holds the tool, and guides it based on an on-board reference (I'm not sure, but I think it can also help to auto-correct problems). Ultimately, both of these machines come really close to the first concept I had. It's nice to know someone had the same ideas and took it further - but that's to be expected, I guess.
The second concept I had was something that I could only see as either an owned piece of equipment, or as a rental/lease system. Essentially, this would be a container (ideally a standard shipping container). One end, you'd load an "elevator" of plywood sheets (like a printer of sorts), the pieces would go inside the "magic box", and would be spit out the other end (again, stacked) as mostly pre-cut pieces of plywood containing the parts of the building.
Think of these pieces as being similar in scope as those "pop-out" wood puzzles of dinosaurs and such you can buy as toys, just larger scale. The cuts made by the machine wouldn't be "complete", but would have to be completed by the user using a jigsaw or handsaw. This would also allow shipping pre-cut pieces (which could be made in a factory) to a building site. Perhaps these could be "ordered online". But the original concept was that the container box would be dropped off at the on-site building location, with a load of "blank" pieces of plywood/OSB.
Inside the box would be a simple system of conveyors and other plywood material handling devices that would move the pieces from the input stack, to a cutting machine, then to the output stack. The cutting machine could be either a CNC router (like a 4x8 shopbot or similar) or some kind of large (2-300 watt?) laser cutter. The latter might be preferable, because it could also engrave assembly annotations, etc on the parts.
The downside of this machine (actually both ideas have similar downsides) would be the fact that you'd need to monitor and do tool bit changes as they wear - or in the case of a laser cutter/engraver - monitor and prevent fires from starting. Which would seem to make this kind of a system more suited to a "factory floor" installation rather than an on-site system rented by a diy'er or a contracted builder. Then you'd just go online, select a floorplan/style, make certain mods as available, and order. Completed sheets of the parts would be delivered to your building site (perhaps with all needed fasteners and such).
Lastly - I'd be willing to bet that such a system (and the "WikiHouse" designs) - at least in the United States - will suffer from the same issues that plague people who want to use other alternative construction techniques (aka - rammed earth, straw bales, earthship, etc): That of the hassle of getting permits approved for the construction, then obtaining homeowners insurance for the resulting structure. In some areas, both can be so onerous as to make such construction impractical, more expensive, or impossible...
Your mention of attaching a device to draw the cut-lines made me think of Hektor, Viktor, Otto, Rita projects by Jürg Lehni
http://juerglehni.com/works/hektor
But, getting a mortgage to build an alternative style of house is still all but impossible.
Code in most places still isn't up to snuff on the issue of natural building materials, and the like, as well. New Mexico is one of the better locales, as it has specific code coverage for adobe, earth bag, and a few other kinds of natural building material and design. You still need to have an architect sign off on it, most likely, but it's plausible to get a weird house built in much of New Mexico without a tremendous amount of money/time wasted on code compliance.
I've been researching this, as I'd like to build a house with my own hands, and figured if I was going to do it, I'd want to do something interesting. New Mexico is among the most friendly states in which to do that, but some other states and counties are viable, as well. But, again, banks don't really like weird houses.
I've always thought origami-style cardboard furniture would be a good idea. Cut, fold, put on a blanket and cushions, and you've got a sofa. And getting hold of some cardboard is not difficult.
https://www.opendesk.cc/
I think we'll be seeing a lot of things like this in the next few years.