If you’ve seen the holodeck on episodes of Star Trek here’s the real-life application. Essentially, Intel is working to create physical, three-dimensional replicas of people or objects, so lifelike that human senses accept them as real. This is interesting stuff for the field of medicine but it’s applications to architectural representation and built form merit discussion as well.

http://web.archive.org/web/20080108083214/http://www.intel.com/research/dpr.htm

Check out the section: ‘Potential applications of Dynamic Physical Rendering’

Tom, Agustina and I had an interesting conversation over beers the other night about the emerging role of computers in the design and construction process.  To my mind two main questions emerged: what is a useful role for computation in the design process and how does architecture respond to the impact of CNC manufacturing processes on the constraints imposed on the builder.  These are both very expansive topics, so I’ll stick to the first.

So this is mostly just cool.  A researcher at IBM has just developed a way to perform computational algorithms on encrypted data, ‘homomorphic encryption‘ in the crypto-parlance.  For example, you could upload encrypted photos to Flickr and Flickr could adjust the color balance without ever decrypting the photo.  This isn’t a particularly useful example, but the general point here is that data being stored in ‘the cloud’ can now be stored and modified in an encrypted form.  Some comments from Forbes:

Homomorphic encryption, like any encryption, mathematically scrambles data so thoroughly that the unscrambling can be done only by someone possessing a secret key. But even though the encryption seems to produce chaos, it might preserve certain mathematical relationships among the possible inputs. If the encrypted version of input x, multiplied by the encrypted version of input y, equals the encrypted version of x times y, then the process is said to be homomorphic with respect to multiplication.

Fully homomorphic encryption would preserve not just multiplication but also addition. What’s the point of that tweak? Any computer algorithm–whether it sorts your mail or figures out whether you qualify for a tax deduction–boils down to a series of arithmetic steps. If an encryption scheme allowed any number of additions or multiplications, any computing application would be possible without decrypting data.

The idea of fully homomorphic encryption was first posited in a paper three decades ago by Ronald Rivest, an MIT professor and the coinventor of the famous RSA encryption scheme now ubiquitous in business transactions. Rivest and his two coauthors also suggested it was probably impossible.

The downside is that the initial aproach is extremely expensive computationally; something on the magnitude of 18 orders of magnitude for large datasets.  Hopefully future developments will reduce this…

This is probably some of the coolest software I’ve seen in years.

I would explain it – but just watch it, you’ll see.  Supposedly there’s a Google Sketchup plugin coming out soon that makes 3d building models using the same basic technology.

I hadn’t heard much about this in a few years so I did some digging and it looks like we’re on the verge of having some very interesting technology for networking.  Back before wireless routers became commonplace there was talk of using a house’s electrical wiring as a data network – after all the copper is already there, why not use it for signalling as well as power.  The problem is that all the stuff we plug into our electrical wires creates a lot of noise, so transmitting data packets at a reasonable speed is difficult (older standards were too slow to use for networking).

Well it looks like they’re managed to solve most of those problems.  There are systems now that operate at up to 200Mbps (Wireless-G maxes out at 20Mbps in best-case scenarios), and there are standards being worked out which will provide gigabit/second transmission rates.  The first chips for these new standards are expected to start rolling out later this year or early 2010.  The most exciting of these new standards, G.hn, is designed to work over power lines, phone lines, or coax (I suspect the transmission rates will vary depending on which line is being used).

The reason this is interesting to me is that it allows practically anything in your house to be networked; toasters, garage door openers, fridges, televisions, lights, security systems, speakers, HVAC systems, PV arrays, etc.  Anything that plugs in could be networked by simply adding another chip to it’s internal circuitry (I would guess these chips will cost less than $10 in bulk).  Think of the possibilities!  This allows control and monitoring, so not only can the lights turn themselves off when you leave the house, you could graph the exact energy consumption of every appliance in your house.  Coupled with demand pricing for electricity, this would enable enormous gains in energy efficiency.

Infoaesthetics posted about some researchers who downloaded 35 million geotagged images from flicker and did some data mining.  The images below show the paths travelled by photographers in New York, based on images from the same camera taken within a few minutes of each other.

The researcher’s site has links to the full paper as well as a few more images including this heatmap:

This seems like an interesting way to analyze people’s movement patterns – you could make interesting video showing the frequency and location of pictures taken at different times of the day , days of the week, or days of the year.

Imagine 20 years from now using the accumulated photos of billions of people to create time-lapse fully immersive 3d environments which would allow you to ‘slide’ forward and backward in time.  The technology to create 3d models from a set of images already exists:

A virtual reconstruction of the Statue of Liberty, created from tourists

The coming ubiquity of GPS-enabled gadgetry will undoubtedly create all sorts of interesting information visible (for better or for worse).

From LinuxDevices:

Marvell Semiconductor is shipping a hardware/software development kit suitable for always-on home automation devices and service gateways. Resembling a “wall-wart” power adapter, the SheevaPlug draws 5 Watts, comes with Linux, and boasts completely open hardware and software designs, Marvell says.

In typical use, the SheevaPlug draws about as much power as a night-light. Yet, with 512MB each of RAM and Flash, and a 1.2GHz CPU, the unobtrusive device approaches the computing power found in the servers of only a decade ago.

If only this had some type of wireless connectivity (although I guess you could put a wireless card in teh USB port).  I’m not sure what this could be useful for, but I’m sure there are plenty of options.

Google just released a service called Latitude, which enables people with certain types of cell phones to publish their physical location.  From Computerworld:

How does Latitude do that? Google is using technology that’s similar to that of Skyhook Wireless in its Latitude service. Like Skyhook, it is a software-only location solution that allows any mobile device with Wi-Fi, GPS or a cellular radio to determine its position with an accuracy of 10 to 20 meters. What sets XPS apart is that it uses land-based Wi-Fi access points, GPS satellites and cellular towers to determine location information.

In other words, Latitude can use any of the three kinds of signals — Wi-Fi, 2G/3G/4G mobile or GPS satellite — that a device can pick up to work out its location. By leveraging these wireless capabilities, the software can combine positioning data from satellites, carrier assistance servers and Wi-Fi base stations to significantly speed up positioning, or TTFF (time to first fix). TTFF for some devices can be up to a minute, but by using multiple reference sites, Latitude can reduce TTFF to a few seconds.

Basically, the software installs on your phone and  periodically checks to see where the phone is located, then send that information to Google.  Aside from the slightly creepy privacy implications, this seems like an interesting way for buildings to ‘know’ where their occupants are located (or their cellphones at least).

I could imagine houses closing the windows, turning off the lights, and dropping the temperature when nobody’s home, then reversing the cycle as the owner starts driving home from work.

I could also see some interesting urban dynamics research using this system if it were widely adopted and the data were public (again, my privacy nerve is twitching…)

After reading and discussing Tom’s post, I thought I’d throw out some more ideas.  I think the biggest challenge to start with is developing a coherent message about how we think the country should go about addressing the ‘green revolution’

Charles Perrow published an interesting book called ‘Normal Accidents‘, which is an anthropological study of catastrophe.  He argues that all systems fail given enough time and that to reduce the frequency and severity of catasrophes we must address the systems which fail (rather than blaming actors in such systems, ie blame the people).  His basic argument is that there are two fundamental factors which influence the probability of a system failure: the complexity of the system and the interconnectedness of the system. Complexity here refers to non-linear causation chains; a complex system is one in which a cause cannot be clearly determined by observing an effect.  Interconnectedness refers to the degree of coordination required between components of a system.  He argues that as complexity and interconnectedness increase, the probability of failure must also increase.

Infoaesthetics just posted a link to a new web page called GeoCommons which allows “non-technical professionals” to view and analyze geo-located data, without the traditional GIS overhead.

As a quick example of why this is cool – here’s a map of the percentage of bridges in each state which are classified as ‘structurally deficient’.  20-25% of the bridges in those states with the big circles are structurally deficient.  Makes you wonder if road trips are such a good idea after all.  Maybe spending some money on roads and bridges isn’t such a bad idea.