Toyota learns the tyranny of software complexity

March 2, 2010

A good column about Toyota’s acceleration mess. The author is a former electrical engineer at Ford, and discusses the complexity of the software that runs modern cars. He compares this problem with previous major recalls by other vendors. The comments to the post are good, too. Here’s an excerpt:

The system level error that Toyota made is not letting a brake signal override a throttle signal. I designed speed control systems at Ford, and everything was dependent on having a tap on the brake cancel any speed control function. A throttle-by-wire car like Toyota makes is almost free to add speed control, you just have to have a button to tell the ECU (engine control module) to hold speed and a brake signal,

Read the rest of this entry »


Roger Bohn’s February 3rd talk

February 26, 2010

Hello everyone! This is L. Lin Ong, the Graduate Student Researcher for the How Much Information? Project 2009 report. I have a few photos to share from Roger’s recent colloquium talk.

The talk was particularly interesting due to the post-presentation discussion regarding different philosophical viewpoints on information, an issue we grappled with during the early stages of the report.


ANY product can be improved

February 21, 2010

For my upcoming product development class. A team, following the methods taught in the course, can improve the design of ANYTHING. A strong claim, admittedly.

Building a Better  Mailbox

It is often said that there are no new ideas, but Ms. Troyer and Mr. Farentinos turned that cliché inside out. By correctly anticipating how the high-tech future would change the way we shop, they updated one of the most low-tech items around: the repository of snail mail, the trusty mailbox. Along the way, they responded to a growing concern — identity theft — that established mailbox suppliers had failed to address.

via Prototype – Architectural Mailboxes – A Tale of Determination – NYTimes.com.

The course is based on the textbook by Ulrich and Eppinger:  http://www.ulrich-eppinger.net/


Broadband speeds increase – does it matter?

February 12, 2010

If this increase in broadband speeds is correct, the Internet will (finally) begin to make inroads on the number of bytes people receive, not just the number of words. According to our estimates, average effective bandwidth on the Internet was too low to send many bytes, compared with TV. (Remember how awful YouTube videos were in early 2008?) Partly this is because bandwidth in the last mile, which this report apparently covers, is not the only limit on throughput. Latency delays, limits on originating sites, and pauses by users all reduce average throughput. (This is very visible when I surf from UCSD, where I have speeds above 100 Mbps to my desktop. I still encounter delays.)

US Broadband Speeds Rose 28% in 2009SCOTTSDALE, Ariz., February 9, 2010 – The US residential broadband speeds continue to increase, albeit at a slower rate than in 2008. Between year-end 2008 and year-end 2009, downstream bandwidth rose by 28%, reports In-Stat http://www.in-stat.com.

via In-Stat – Press Releases.

The Internet already has a substantial fraction of average word consumption, because of the higher words-per-minute of reading over radio and TV.

If anyone knows of reliable data on average effective home Internet speeds, please send them along.


Latest development in the battle over vaccinations

February 4, 2010

The Lancet retracts paper linking MMR vaccines and autism By Matt Ford | Last updated February 3, 2010 9:27 AM

This week, after receiving the conclusions of a multiyear ethics investigation of UK doctor Andrew Wakefield performed by the General Medical Counsel GMC, the editors of British medical journal The Lancet formally retracted a study which purported to find a link between the childhood MMR vaccine, gastrointestinal disease, and autism. It was published in 1998 and has been a source of controversy ever since.

via The Lancet retracts paper linking MMR vaccines and autism.

This article is tale of greed and incompetence.  But the harm it did lives on. Here’s a site that claims to present “all sides,” but clearly thinks Wakefield is a hero. There are many like it – conspiracy theories flourish better than scientific analysis on the web. (More fun  and much easier to write, after all.)  For a good article on these hysterias, see an article by Amy Wallace in Wired.  As best I can tell, some  parents cannot handle the concepts of bad luck or Acts of God. If their child gets sick, someone is responsible! And it’s part of a widespread plot!

A second problem, much more widespread than the vaccine phobias, is that people have trouble dealing with small probabilities. (This observation goes back at least to research by Kahneman and Tversky on how humans have systematic cognitive biases.) So you can find nonsensical statements like “If screening for disease X [breast cancer screening under age 50 is the current example] saves even a single life, than not doing it is manslaughter.” What’s the problem? Screening itself causes difficulties, such as unnecessary biopsies. Not to mention that more lives might be saved by spending the same amount of money on something else. So deciding whether/when to get screened is a balancing act; it’s not all one way or the other.

One interesting bit of sociology (which to me is further proof that these health-scare controversies almost never have factual basis): most countries have phobias about vaccines and medicines, but the specific phobia varies by country. For example, the smallpox-eradication effort fell apart at the last moment when smallpox vaccine was rumored to cause infertility. (I’m looking for more specifics on the nation-specific fears of vaccines – I’ve forgotten where I came across it.)

Yet,  there are a LOT of problems with modern medicine, and with drugs in particular. But in an environment where any half-baked theory gets taken seriously, it’s very hard to separate the fear-mongering from the real problems.


The Lancet retracts paper linking MMR vaccines and autism

February 4, 2010

The Lancet retracts paper linking MMR vaccines and autism By Matt Ford | Last updated February 3, 2010 9:27 AM

This week, after receiving the conclusions of a multiyear ethics investigation of UK doctor Andrew Wakefield performed by the General Medical Counsel GMC, the editors of British medical journal The Lancet formally retracted a study which purported to find a link between the childhood MMR vaccine, gastrointestinal disease, and autism. It was published in 1998 and has been a source of controversy ever since.

via The Lancet retracts paper linking MMR vaccines and autism.

This article is tale of greed and incompetence.  But the harm it did lives on. Here’s a site that claims to present “all sides,” but clearly thinks Wakefield is a hero. There are many like it – conspiracy theories flourish better than scientific analysis on the web. (More fun  and much easier to write, after all.)  For a good article on these hysterias, see an article by Amy Wallace in Wired.  As best I can tell, some  parents cannot handle the concepts of bad luck or Acts of God. If their child gets sick, someone is responsible! And it’s part of a widespread plot!

A second problem, much more widespread than the vaccine phobias, is that people have trouble dealing with small probabilities. (This observation goes back at least to research by Kahneman and Tversky on how humans have systematic cognitive biases.) So you can find nonsensical statements like “If screening for disease X [breast cancer screening under age 50 is the current example] saves even a single life, than not doing it is manslaughter.” What’s the problem? Screening itself causes difficulties, such as unnecessary biopsies. Not to mention that more lives might be saved by spending the same amount of money on something else. So deciding whether/when to get screened is a balancing act; it’s not all one way or the other.

One interesting bit of sociology (which to me is further proof that these health-scare controversies almost never have factual basis): most countries have phobias about vaccines and medicines, but the specific phobia varies by country. For example, the smallpox-eradication effort fell apart at the last moment when smallpox vaccine was rumored to cause infertility. (I’m looking for more specifics on the nation-specific fears of vaccines – I’ve forgotten where I came across it.)

Yet,  there are a LOT of problems with modern medicine, and with drugs in particular. But in an environment where any half-baked theory gets taken seriously, it’s very hard to separate the fear-mongering from the real problems.


See Roger Bohn speak on HMI

January 25, 2010

Roger will be appearing at a UCSD colloquium on Wednesday, February 3, to discuss the results of the HMI project.

The event will be held in the Media Center / Communications building on campus (map here), and the nearest parking is just across the way on Muir College Drive. He’s scheduled to speak at 12:40PM in room MCC 201.

Here’s the official blurb:

How much information do Americans consume? At the start we have to define information, consume, and much.  All three definitions are unavoidably controversial. …

I will present and discuss our results, most of which are available in our report at hmi.ucsd.edu. We didn’t have strong expectations of what we would find, but we were surprised anyway. Read the rest of this entry »


HMI Bonus Material: Video Game Screenshots

January 24, 2010

Hello! Blake Ellison here, and I’m Roger’s newest grad student assistant. I’m interested in video games (both academically and personally), so I’m helping the team try to make sense of our findings that video games make up a huge proportion of our data consumption (when bytes are used as the measure).

A simple reason why video games comprise so much of our information  is the sheer volume of pixels that get transmitted to your eyeballs. A game running at 60 frames per second at 1080i on a current-generation console like the Xbox 360 is pumping out a huge amount of data. That’s to say nothing of hardcore PC gamers, who have what amount to miniature supercomputers sitting on their desktops.

Polyphony Digital's Gran Turismo 5

These ’supercomputers’ don’t have all that power simply to push out 1920 x 1080 pixels 60 times per second. They have the power to do all that and make it look good. Read the rest of this entry »


Small is Beautiful for global health tech? A good theme for project proposals

January 21, 2010

(This presentation discusses the “car-parts neonatal incubator” in more detail, which  the NY Times profiled recently.  This is an example of appropriate technology. It would make a good takeoff point for projects in several of my courses.)

Many multinational companies manufacturing medical devices for developing countries focus their efforts on high-end products too expensive to be used in most healthcare settings. Unable to afford their own equipment, healthcare providers in areas with few resources often receive donated equipment from international organizations. Unfortunately, while this donated equipment is usually state-of-the-art, it often ends up falling into disrepair and eventually disuse. Donors with the best of intentions fall into the trap of donating equipment that the recipient cannot afford to maintain. Just as most of us would like to own a Ferrari but would be unable to pay for its upkeep, most clinics in resource-poor areas cannot afford to maintain expensive devices, such as incubators, designed for use in developed countries.

via CIMIT Forum: Medical Devices in Global Health: Idea to Implementation, Successes and Challenges.


Military Deluged in Drone Intelligence – NYTimes.com

January 12, 2010

By CHRISTOPHER DREW  Published: January 10, 2010

HAMPTON, Va. — As the military rushes to place more spy drones over Afghanistan, the remote-controlled planes are producing so much video intelligence that analysts are finding it more and more difficult to keep up.

Air Force drones collected nearly three times as much video over Afghanistan and Iraq last year as in 2007 — about 24 years’ worth if watched continuously. That volume is expected to multiply in the coming years as drones are added to the fleet and as some start using multiple cameras to shoot in many directions.

via Military Deluged in Drone Intelligence – NYTimes.com.

[Actually I'm surprised that the 24 multiplier is not higher, especially since some drones operate at night. On average, only 24 cameras are running.]

Airmen in air-conditioned rooms