h a l f b a k e r y"This may be bollocks, but it's lovely bollocks."
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Please log in.
Before you can vote, you need to register.
Please log in or create an account.
|
Add rigid body modeling to the particle model of interior arrangement. Now objects have a shape, not just a redius. Their asymptotic states start looking more like building interiors and less like monkey bread. A piece of furniture, or even a piece of building materials can be represented as a CAD
"block". A table of points within blocks Attractive and repulsive forces can now act on multiple points in a block, not just a single point representing its middle. Now of course our model has to account for objects spinning around in space and magickally aligning themselves relative to each other. This of course needs gazillions of lines of code. For example a "rule" could be encoded to the effect that brick vertices are attracted to brick long edge midpoints. The result will no doubt contain a lot of chunks of partially assembled corners and pieces of floorboards. They could be used as input for future randomizations. Or perhaps eliminated in further refinements of the algorithms.
particle mechanics meets architecture
http://www.brightid...932F0D2}&bucket_id=
An earlier incarnation. [LoriZ, Oct 06 2001, last modified Oct 05 2004]
Listening for the Fit
http://www.findarti...le.jhtml?term=virus
Describes an interesting combinatorial technique that 'listens' for the shape of a suspended particle. [reensure, Oct 06 2001]
Graph Layout Examples
http://java.sun.com...1/demo/GraphLayout/
These Java examples do something similar, but try to maintain a set distance between nodes, not get close/far. Drag nodes, shake, scramble, view source. [pottedstu, Oct 06 2001, last modified Oct 05 2004]
MIT’s Crazy Materials Could Make for Self-Assembling Ikea Furniture
http://www.wired.co.../11/skylar-tibbits/
As its name suggests, the Self-Assembly Lab focuses on making things that can, well, self-assemble. It has created a series of small wood planks, for instance, that fold into a toy elephants when exposed to moisture. Tibbits and collaborators Christophe Guberan and Erik Demaine are working on products that could morph in response to the weather. In the future, the Lab’s research could make way for Ikea furniture that assembles itself with a splash of water—no Allen wrench required. [LoriZ, Nov 12 2014]
[link]
|
| |
ooooh, I don't have a clue what you're talking about, but 'clumping structural models' sounds a lot like a great new crustal development theory... all the mountain ranges of the world... why are they so concentrated in their specific areas? Is it because all the localities with interesting structural elements feel the need to be together, and so they clump? All the structural models in the world have never yet addressed this empathy that the gniesses feel towards each other! Perhaps that's why they're called gniess! Ahhh. I have found geological inner peace. |
|
| |
Sorry, LoriZ, I think you could explain this a little better. I couldn't make head or tail of your description, though your link makes it a bit clearer. I don't understand why this is under programming: algorithm, though. Surely it should be under Home: Interior Design, or whatever the nearest category is to that. |
|
| |
The goal as I see it is to search for an optimum arrangement of objects in a house (ranging from door to appliances to furniture to rooms) by specifying a series of relationships based on which should be nearby and which should be far away, and assigning a different importance to different relationships. By calculating the distance between each one, and looking at the attractive/repulsive forces you can calculate the quality of the arrangement, and use these figures to move objects and improve the arrangement. |
|
| |
The change from your old idea to the new idea here seems to be: originally, you viewed objects as points, but now you're viewing them as 3-dimensional objects with extension, so there must be a finite distance between them, which will be more accurate, but still not perfect. |
|
| |
Such a system could in theory be used either to score existing designs, or to create new designs - but the latter would fail because you won't take into account things like allowing space between things, and fire regulations, and windows (which have to be on the outside of the house) and water pipes, which must be continuous, and the engineering/physical constraints that limit buildings. |
|
| |
Running a search to find an optimal solution might end up suggesting you stack everything in your house vertically, with your front door on the top. But you probably wouldn't want that. |
|
| |
I think a better approach would be to interview a series of architects, structural engineers and interior designers, try to establish the principles they use, list them as a series of rules, and encode them in an expert system. Part of me thinks that with so many houses built each year, they must have found an optimum by now; but part of me's more cynical. |
|
| |
Well, if you add the local fire codes, and a requirement that all interior parts of the house be reachable by an able-bodied human, and then use the clumping structural models with a simulated annealing algorithm, you should be able to generate workable house plans. |
|
| |
Better, since simulated annealing is somewhat random, you can generate different house plans. Perhaps this way, new housing developments wouldn't have the unsettling uniformity and sameness that they have now. |
|
| |
You'd want some way to calculate construction costs as well. And make sure the WC wasn't in the middle of the living room, and the shower in the front garden. Another problem might be that in an effort to get your garbage disposal as far from the nice parts of your house as possible, it might end up in Arkansas or the bottom of the Indian Ocean. |
|
| |
The way you state your idea reminds me of me trying to explain that, of the multiple ways the human body can move, the coefficient of each movement's efficiency is related to the leverage provided by the most stationary complementary body part. Hmmm. |
|
| |
Got a link for you, though. |
|
| |
Graph Layout Examples: I found some applets (included in the Sun JDK 1.2, and on Sun's website) that illustrate a similar idea. They each use the model of a graph made up of a series of nodes, with a target length specified for each node. Nodes start in a random position, and push and pull to find a good arrangement. You can see how different shapes can arise by dragging nodes around, or reordering the whole model. You can also, if you view the source for the applet and web page, use your own graphs. You'll see in example 1 how it tries to keep the cat and mouse, and mouse and food a fair way apart, but the fleas close together. This is a bit simpler than LoriZ's idea, but should let people get a better idea of what it's all about, if I'm right about what it's all about. |
|
| |
(Note: when you click a button on the applets for the first time, it has to load a groovy sound effect, so it might look like it's crashed while you wait.) |
|
| |
This is obviously a bit more complex than the graph-paper & furniture cut-outs that we used to use for re-arranging our office. I doubt if it's any better though. |
|
| |
I see this as a way to throw all architects and interior designers out of business. |
|
| |
Well, the sunjava example is more or less exactly what we had in mind originally. As for building codes, let's say we have a block representing an eclectical outlet. We can define a point 12 inches below it and gravitate it to the topside of floorboards. As for the door on the roof, that gravitates toward the heliport. The main entry door gravitates to ground level. As a further refinement, elevation of objects relative to ground could also influence calculation of acceleration vectors. I don't know if we're already talking about an unrealistic amount of computing power, though. Water pipes are a componentized system, and what part of what component is compatible with what else is pretty well defined. As for cost guesstimating, let's say we have a number of people from whom we have collected down payments for a subdivision development. We have a block representing each consumer. Each consumer is modeled as having x sub i dollars in 'eir pocket. There is assumed to be an attractive force of between dollars in pockets and dollars represented in the prices of 2x4's. Doesn't account for labor, but that's yet another refinement. |
|
| |
Oh man. Synthetic Spice just won't work. Get your ass to Arrakis.
I'd be interested in seeing a more elaborate post that eliminates the guesswork in what you are saying.
I am not the God of AutoCad. I am only a minor diety & might have been hooked by your problem, yet I don't really understand it. |
|
| |
<comment aimed optimistically at
extremely narrow potential target
audience>I thought this was going to
be about improvements to the SCOP
database</caoaenpta> |
|
| |
For some reason when I read this and could make no sense of it up or down, I started laughing and couldn't stop. I thought this had to be the funniest thing ever written. |
|
| |
Now that I know it's actually a real something, it's not funny at all. |
|
| |