Tech Talk iconTech Talk

Down to Two

The discussion is intense: Men are arguing about why women buy dresses.

GPS fashion tagging will take the form of a web site where a user can upload a picture of clothing they like, on anyone on the street. Then other users will provide the informationâ''she got that dress at H&M for $35.00, say. Ideally, that Buy It Now possibility would get clothing stores involved.

The argument I am now eavesdropping on is about why a clothing store would buy in. â''Why would stores get involved if a user can find one dress for $1100 and then â''tagâ'' a similar dress on someone else that costs $100?â'' asks one argumentative potential entrepreneur. â''No one is going to buy the 1100 dress!â''

â''But thatâ''s not why women buy dresses,â'' argues a kid with a full dayâ''s beard stubble and a big mop of curly brown hair. â''They decide based on, is the hem like this,â'' he holds his hand over his knee, â''is the hem like this... price isnâ''t the most important part.â''

â''Have you met them?â'' the other guy challenges.

â''Yeah! Iâ''m married to one!â'' Curly yells.

But the Internet Scavenger Hunt is going rapidly down in flames. Itâ''s a pretty good idea, but I think itâ''ll lose to the other two. The idea is a web site that helps people figure out how to navigate the internet by way of games. A scavenger hunt would force a person to learn a lot about navigation, scrolling and so on. I think itâ''s designed for kids, but Iâ''m thinking of my parents, who call me every time â''The Computerâ'' does something unexpected.

But I think it was designed a bit too much with corporations in mind. Everyone is talking about how big companies would love it because they could send people on the hunt for an icon on a particular site, so the customer would spend a lot of time thinking about the companyâ''s brand. Whatâ''s the userâ''s motivation? I just overheard someone say, â''You look like youâ''re shilling for SONY.â''

And the votes are inâ''Internet Scavenger Hunt is out.

The Cliff Notes

Okay, now weâ''re going to backtrack a little.


I think I explained Startup adequately. But what does it have to do with Brooklyn Polytech? Mike Wills, (standing on the left) is a Polytech alumn. He read about Startup weekend on BEST would be a good host. Because Poly is hosting the event, 30 spaces of the 100 were reserved for Poly students. The rest of the attendees are entrepreneurs from around the New York area. Registration was first come, first served. Bruce (sitting, left) sent the email on Wednesday afternoon--registration was closed five minutes later. They were flooded. But there was no application process. If you can get a spot, you're in. I ask Andrew if that self-selection process poses any problems-- no one is screening for qualifications. He smiles. â''If youâ''re willing to show up on a Friday night after a busy work week, and work all weekend, Iâ''m guessing you're qualified."

And think about telling a university PR department (and all the other various planning entities and persons who require notification) to coordinate this event with a two-day lead time. A full weekend event, hosting 100 people, making the posters, catering the meals. Apparently it workedâ''thereâ''s a roof over everyoneâ''s head, a turkey sandwich next to my computer, and a poster hanging only slightly crooked on the wall.


And tomorrow, theyâ''re bringing in the big guns. They donâ''t want media coverage for that, but this nascent little idea will be put through its paces by a savage horde of lawyers, accountants, PR consultants and marketing people. How feasible is it to put together a global site that ranks the best cheeseburger? How do you let people know the site is out there? How do you motivate them to add their two cents? How do you decide on the look and feel? The coders will have their work cut out for them. I doubt theyâ''ll get much sleep. On Sunday the graphic designers will design the look and the coders will be frantically tapping gobbledygook into their laptops.

And on Sunday night, the world will have a new web siteâ''either ranking cheeseburgers, harshing on bad dresses, or making you troll the internet for the IBM logo. â''Web-based businesses are probably the easiest to start, grow and maintain,â'' Bruce explains. And of course, youâ''re not going to entice a college kid with the promise of keeping inventories. Patent lawyers will have bestowed upon it a little copyright symbol. Thereâ''s no prize money for â''winning,â'' but Iâ''m told venture capitalists will be here this weekend, flush with startup cash. Thereâ''s a complicated shareholder system that I wonâ''t bother to explain, but essentially every day youâ''re here is a share. If youâ''re here on Friday, you get one share. All weekend, all three possible shares. Andrew makes a point of saying that has no shares in the company.

Down to Three

Well, I picked two out of three: we've whittled down the ideas, and the three remaining contenders are User Generated Fashion tagging, Greatest Menu Meals, and Internet Scavenger Hunt. It's interesting that I zeroed in on two of them-- it makes you realize that some inventions just make sense. I mean, who wants to eat at the world's best Thai restaurant and get stuck with the worst Basil Chicken?

Andrew has a tight schedule for the rest of the year. NYC startup is only 2 months old. Andrew invented it, did it in Boulder, and then again in Toronto. Word spread virally, and suddenly he was getting emails from all over the world-- everyone wanted one of these. So he has a tight schedule for the rest of the year. Fourteen startup weekends will take place in London, Dublin, Hamburg, Boston, DC-- I won't list all 14, but here's the schedule for 2007.

Right now we're discussing a major stumbling block to the success of Greatest Menu Meals: Apparently people really don't like taking surveys online. I thought that was just me, being misanthropic. I feel much better.

NYC Startup Weekend

startup-weekend.jpg Various Spectrum folk will be blogging the NYC Startup weekend, hosted by Brooklyn Polytechnic University's BEST startup incubator, from now until Sunday. BEST (Brooklyn Enterprise in Science and Technology) director Bruce Niswander is a self-proclaimed "serial entrepreneur" who understands the 101 ways 99.9% of startup businesses fail. Startup is the brainchild of Andrew Hyde, a 23-year-old Boulder, C.O. wunderkindwho tells me he is "great at the first 25 percent of starting a business, and terrible at the rest. I suck at day-to-day." He's here to inspire the 100 people from all over New York sitting here in the standard-issue linoleum-floor-and-fluorescent-lights university room where the event is kicking off with keynote speeches. (In his keynote address, Andrew mentions first that 15 people are going to be late.) Right now it's a little bit like Lord of the Flies; we're trying to get a sense of who will emerge as the leader, a process which can only happen organically in an impromptu setting like this one. Everyone is networking with each other and trying to funnel their ideas down to The One Big Idea, which will emerge gradually and finally pop into existence on Sunday night. The goal: A web-based business with a corporate identity ready for prime time, with a functioning web site, a business model, and willing investors. As of right now, the thing doesn't even exist. These 100 people have 48 hours to make all that happen.

It's going to be an intense weekend. They started tonight at around 6:45 (plagued by a particularly "old world" problem that feels like a museum artifact in this new-media-only environment: the delayed insurance certificate). They'll be here tonight until 11 pm, get back at 9 a.m. tomorrow morning, be here until 11 p.m., back at 9 a.m. on Sunday morning, and here again until 11 Sunday night. At 11 pm, in theory, you have a startup with 100 shareholders.

We're listening to everyone's 30-second pitches-- so far I've heard about user-generated fashion via geotagging, 11th hour reservations, Green Ads (where ad revenue gets funneled not to the host but to, say, Greenpeace), and a ratings systems for individual meals. ("Sometimes when I'm in Queens, I don't want the best restaurant-- I want the best chicken parm," says the idea's originator, a clean-cut looking blond guy in a striped polo shirt. He may have just figured out the company's tag line.)

Andrew knows how to whip a crowd into a froth. People are really excited. They're discussing their 30-second pitches in small groups, but the volume is rising steadily towards that cocktail party level right before the moment when everyone goes mysteriously silent just as you loudly say the word "...breast!"

London's public cameras can't solve crimes

For years groups like the ACLU have been fighting to restrict the use of surveillance cameras in public placesâ''itâ''s bad enough to be caught on film every time you use an ATM or buy a convenience store hot dogâ''and they've been slow to catch on. Across the pond, however, the United Kingdom has embraced the technology, setting up more than 10,000 cameras in London alone.

But if you're planning a British crime caper, don't despair: having more cameras in an area seems to have no effect on how many cases the police actually solve there. The Liberal Democrats of the London Assembly published their proof this week, which they acquired through a Freedom of Information Act, according to Theyâ''re clearly trying to paint the camera systems as a waste of money that provides minimal results.

With a quick look at the new crime data, their conclusion seems hard to argue with: whether boroughs had a few dozen cameras or a few hundred, they all solved only about 20 percent of the crimes reported.

But the straightforward numbers can be a bit misleading. As it turns out, researchers have struggled for years to determine if and how cameras work.

Temporarily setting aside any Orwellian fears, itâ''s easy to see why cameras should reduce crime in theory:

1) They make it easier to apprehend and prosecute criminals who are caught on tape (which the Fox network will eventually turn into television specials).


2)They act as a deterrent (this is why stores install fake security cameras and dad's lie about urine-detecting pool chemicals)

This new data seems to rule out the first, as 80 percent of crimes reported still go unresolved even in densely recorded areas. So what about prevention? A 2005 report [PDF] commissioned by Britainâ''s Home Office, found that installing cameras reduced crime in only one of the fourteen closed-circuit systems that researchers studied. Strike two.

But not so fastâ''according to a 2005 review of closed-circuit video cameras and crime, most studies show that such systems do manage to reduce crime. So why can't analysts figure out if they work or not?

The discrepancy arises because crime statistics almost inevitably bundle many variables into one number. In the 2005 Home Office Report, the authors note that crime rates may have gone up in some areas simply because the surveillance systems recorded crimes that would have otherwise gone unreported. Similarly, individuals might be more likely to report crimes if they suspect there's video evidence backing them up.

Regardless of these factors that may suppress the statistical success of video surveillance, the fact that there's still so much debate means that they aren't revolutionizing law enforcement. Even with continued technological improvement, there's a chance that Big Brother just doesn't work.

Harvard Launches New Engineering School

Harvard University is upgrading its engineering and applied science division to the status of a full-fledged school, on a par with its other famous schools such as law, business, public health, and the John F. Kennedy School of Government. Festivities celebrating the creation of the Harvard School of Engineering and Applied Sciences took place September 20 in a tent immediately adjacent to the Maxwell Dworkin engineering building, which was dedicated in 1999 and is named after the mothersâ''using their maiden namesâ''of Microsoft chieftains Bill Gates and Steven Ballmer. The building is on the site where Howard Hathaway Aiken had his Computation Laboratory and built in 1944 his famed Mark 1 electromechanical computer, with help from IBM.

Despite the presence of Gates and Ballmer in Harvard Collegeâ''s class of 1977, Aikenâ''s achievements, and the continued strong cooperation of IBM in the universityâ''s engineering programs, launching an engineering school is rather a departure and a bit of a stretch. Harvardâ''s first engineering offerings were made only in 1847 as part of the Lawrence Scientific School, a couple of hundred years after the collegeâ''s foundation, and even then were looked upon with considerable suspicion. Charles William Eliot, perhaps the collegeâ''s most famous and influential president, felt that engineering comported poorly with the schoolâ''s liberal arts culture, as it â''had a practical end always in view.â'' Harvard abolished the Lawrence school in 1906, and for a hundred years after that, engineering suffered constant reorganizations too tangled and painful to contemplate.

Speaking at last weekâ''s celebration, Harvard President Gilpin Faust, taking note of Eliotâ''s attitudes, credited the universityâ''s engineering dean, Venkatesh Narayanmurti, with maneuvering Harvard toward a more tech-friendly outlook. Venkatesh Narayanmurti is said to know everybody and be known by everybody. He is called so ubiquitously by his nickname Venky that he sometimes is quoted even in print as Dean Venky.

How does Harvard propose to now make a mark in engineering education with a huge and world-famous polytechnic, the Massachusetts Institute of Technology, just down the street? Its strategy, rather than cover the whole ballfield, is to focus strategically on fundamental subjects like applied mathematics and physics, environmental science, and computer theory, with both the betterment of society and the enhancement of basic science in mind. At the same time it wants to emphasize experimental and innovative engineering education.

The pedagogical mission may seem the least dramatic of its objectives, but H. Vincent Poor, the dean of Princetonâ''s engineering school opined during the celebration that â''the best shot at engineering a renaissance in engineering education is for universities of Harvardâ''s caliber to step up to the plate.â'' Armed with the Harvard name, a one-billion-dollar dedicated endowment, and the greater autonomy that school status implies, Harvard engineering plans now to double the size of its faculty and student body.

Tools are not lacking. Engineering faculty and students have at their disposal an IBM Blue Gene Computer, said to be the highest-performance machine self-financed by a university, without government assistance. Just a stoneâ''s throw from the Maxwell Dworkin building is a fancy new laboratory building, designed by a famous Spanish architect, that will provide technical services across the whole gamut of applied science and engineering.

Still, it is a stretch. Keynoter Charles Vest, a former president of MIT, joked that Cambridge natives used to talk about the rivalry between Harvard and MIT, which Harvard seemed to not know about. Will people how jest about the emergent engineering rivalry between MIT and Harvard, which MIT doesnâ''t seem to know about?

Down Massachusetts Avenue at MIT, people arenâ''t completely oblivious. Itâ''s just what MIT needs, one professor commented, â''a sharp kick in the butt.â'' Nah, thought another: after a few years it will just settle to being a very good but very small school of applied science, like those found at a lot of other liberals arts colleges.

Given the tangled history of Harvard engineering, only timeâ''perhaps a lot of timeâ''will tell.

Forward Bias

September 21, 2007

This weekâ''s theme: Donâ''t Tase Me, Bro

First, my defense. Thereâ''s a reason I am jumping on this bandwagon, and itâ''s because Spectrum has an upcoming article about tasers in the December 2007 issue. Mark Kroll and Patrick Tchou explain what exactly is involved in freezing up the human skeletal muscles without disturbing the heart. (Wellâ''theoretically.) Kroll is an expert on heart implants and medical devices who sits on Taserâ''s board. Chou is a cardiologist at the Cleveland Clinic, and the author of a critical perspective pointing out the unknowns and potential dangers of ECDs, also in December's Spectrum.

And now, without further ado, some of the highlights of the literature:

At Wiredâ''s Threat Level blog, Sarah Lai Stirland gives a great rundown of the week's events.

And of course, there was merchandise. So the next time your drooling infant questions the legitimacy of the 2004 election results, you can head off disaster by equipping the child with this handy bib.

If youâ''re over 18 months old but would still like to wear the battle cry of a new generation, get your shirt here.

Suggested use: Wear it while you listen to the dance mix.

And if your interest goes beyond acquiring the resultant pop culture detritus, read a comprehensive Slate explanation of how a taser works. (Itâ''ll tide you over until the Spectrum article comes out in December. Thatâ''s the December issue of Spectrum, coming to a news stand near you in December. Spectrum.)

PlayStation3s power biotech research project

PS3_image.jpgThanks to the power of distributing computing, scientists studying protein folding have a petaflop of computing power at their disposal. Thatâ''s one quadrillion floating point equations per second; kind of like having three Blue Genes (the fastest supercomputer in the world) in the back room.

They hit that magic number yesterday. The project is run out of Stanford, but the call I got today to tell me the big news came from Sony. Because roughly 75 percent of those quadrillion flops are being contributed by Sony PlayStation3s.

Protein folding involves amino acid chains inside cells that, in a healthy organism, twist into unique, three-dimensional shapes. In certain disease processes, the folding goes wrong. Figuring out how and why folding goes wrong may help understand Alzheimerâ''s, Parkinsonâ''s, Mad Cow Disease, and some cancers. But the three-dimensional shapes are complex, and simulating them takes a lot of computing horsepower.backbone_e.JPG

The Stanford-run project is tagged Folding@Home. Like the Search for Extraterrestrial Intelligence project, which pioneered the harnessing of idle cycles of personal computers volunteered for the effort, Folding@Home relies on a volunteer effort. In this case, volunteers register with the Folding@Home network and download the application. The application can run on demand or whenever the computer or game machine is idle.

Since the Folding@Home put out the call to PS3 owners last March, 600,000 have signed on, more than quadrupling the computing power available. Things like this give gamers a good name.

Engineering schools that tie theory and practice together retain more students


Many at Olin College believe that innovative teaching is more than just replacing traditional lectures with hands-on projects. Rather, it is "about the method of presentation," reads a chart in a corridor in Olin's Academic Center.

Engineering students can sometimes be miserable human beings. I know -- I was one of them.

It goes like this: You start engineering school all excited to get your hands on some real engineering stuff ... only to be drowned in theory-heavy lecture after lecture after lecture. Not that there's anything wrong with calculus, linear algebra, electromagnetism -- those are beautiful things, no doubt. But the question is: Why should heavy loads of abstraction precede the more practical, hands-on work?

Why stick with such "boot-camp model of engineering education," as William A. Wulf, president of the National Academy of Engineering, in Washington, D.C., defined it when I interviewed him about a year ago. Wulf and other experts have been arguing for quite some time now that this model should be profoundly revised. I couldn't agree more. In fact, I'd say it should be more than revised -- it should be exterminated.


For one thing, it's not working. Enrollment in engineering programs has been stagnant for the past five years. To make things worse, dropout rates are huge. There are many reasons why students drop out. Some decide engineering is not their thing; others have to leave to get jobs to support their families; and so on. But those reasons also affect biological sciences and other non-engineering programs. So why are engineering students leaving? Is it because courses are too difficult? The answer is no. The main reason is that the engineering curricula are bad: they fail to motivate students.

In a paper titled "Engineering Education Research Aids Instruction" in the 31 August 2007 issue of Science, Norman L. Fortenberry, of the Center for the Advancement of Scholarship on Engineering Education at the National Academy of Engineering, in Washington, D.C., and colleagues report that on average about 56 percent of engineering undergraduates complete their programs. In some schools, such retention rates can be as low as 30 percent -- in other words, two in every three students who begin an engineering program won't get their diplomas. About the roots of the problem the paper says:

Low rates of student retention within the discipline have heightened concerns in the engineering community about the structure, content, and delivery of engineering education. . . . Academic difficulty is not why these students change course, leading the engineering education community to view these loss rates as indicators of defects within the system of engineering education that should be corrected.

So again, students leave not because courses are too difficult (and we're not saying they are not difficult), but because of "defects within the system of engineering education."

But there are good news. Fortenberry and coauthors report that more engineering schools are putting efforts to improve their programs by establishing departments of engineering education, such as those as Purdue University, Virginia Tech, and Clemson University, and by creating experimental courses in which new education methods can be explored.

One such experiment is the First-Year Engineering Projects (FYEP) course at the University of Colorado at Boulder. This course, designed to retain more students, places more emphasis in combining theory and practice in collaborative project-based learning settings. Here's a description from the paper:

The FYEP course connects the conceptual and educational side of engineering with professional practice. This is primarily accomplished through a 13-week project that introduces first-year students to the design-build-test cycle of product prototype development (wherein a new object to meet a stated or perceived customer need is conceived, designed, realized as a physical object, and tested to verify that it meets requirements) in a team-based setting, supported by experimental testing. The innovations at the University of Colorado at Boulder are consistent with the results of previous science and engineering education research on experiential, interactive, and collaborative learning.

The results?

The course boosted retention in the engineering programs to 64 percent. I quote from the paper:

These results add to the growing body of evidence demonstrating that first-year projects-based curricula promote retention of engineering students . . . Although a 64% retention rate is still too low for students that are heavily screened before admission to the major, it represents an improvement on the national engineering retention rate of 56%.

I was able to observe firsthand the impact of combining theory and practice in the classroom, not at the University of Colorado-Boulder, but at another, small engineering college: Franklin W. Olin College of Engineering.

Olin is a phenomenal success story on how to replace the boot camp model with the perfect balance of theory and practice -- and by doing that making engineering students happy. Sure, not all are happy, but the level of happiness there appeared to me to be pretty high compared to other places. I wished my engineering school were like that!

Olin's successful approach to education can be seen in its ultralow dropout rate -- less than 5 percent was what I heard when I talked to Olin officials last year. (Olin is a tiny college and the small number of students, and the fact that all receive full scholarships, can distort statistics.)

About Olin, most experts agree the "experiment is still running," so more results are needed if other schools are to adopt its model. But to me Olin is the best way to understand what's right and what's wrong in engineering education. Who's not getting it? The lesson is obvious. The students who graduate are the happy ones.

Global warming: the undergrad major


Engineering undergrads at Stanford University this year will, for the first time, be able to major in global warming. OK, thatâ''s not exactly the name of the major, officially, itâ''s atmosphere-slash-energy. But global warming is what itâ''s about.

This interdisciplinary major will couple classes like â''Aerosols, Clouds and Climate Changeâ'' and â''Weather and Stormsâ'' with â''Electric Power: Renewables and Efficiencyâ'' and â''Powering the Rim: Energy issues for the Pacific.â''

"In my future career I would like to work toward mitigating global warming, and I found that no other major addresses this issue as well as atmosphere/energy," says about-to-be global warming major Emily Gorbaty. ``Ultimately I want to help implement renewable energy in developing countries, specifically India, China and Southeast Asia.â''

Gorbaty and her classmates are likely to have a wealth of job prospects, since the field of global warming looks itâ''ll be a hot one for years to come. (Groanâ'¿.)

Most Commented Posts

Tech Talk

IEEE Spectrum’s general technology blog, featuring news, analysis, and opinions about engineering, consumer electronics, and technology and society, from the editorial staff and freelance contributors.

Newsletter Sign Up

Sign up for the Tech Alert newsletter and receive ground-breaking technology and science news from IEEE Spectrum every Thursday.

Load More