Anthony Barker's Weblog on Linux, Technology & the Economy

VC humour

A very funny take on how Venture Capitalists spend their days.

Also a decent article on why you shouldn't sell your soul to your employer from harvard (HBS).

New game - spot the American sedan?

If Toronto is any leading indicator of automobile sales American automobile companies are getting whiped out of the smaller vehicle market (sub SUV, pickup and Minivan). On Sunday my daughter any I played a game "SPOT the American Car", while driving through downtown Toronto.

Definitely the 60-70% of cars in the city are sedans or smaller - but from our game only 1 in 4 or 1 in 5 cars are made by american producers.

Rules
* Exclude any vehicles that are over 5 years old
* Call overseas subsidiaries of american companies non american (eg Volvo)

The winners? From our informal - completely unscientific measurements:
Toyota is pushing out Honda with its less pricy corolla, sporty matrix, and even cheaper echo
Hyundia is doing extremely well on the less expensive cars
BMW, Acura, Lexus, Volvo cleaning up on the luxury sedans.
Volkswagen is doing fairly well with the young professionals - golf, jetta, + passat

I was in Chicago last weekend and the number of American cars there seemed to be much higher - Particularly the larger sedans.

Lessig & Less Lethal Weapons

Two interesting articles for a friday.

Lawrence Lessig from wired magazine - one of my favorite intellectuals of the 21st centuary makes a great argument for free trade and against protectionism and copywrite laws.

"Protectionism is competition through government favor rather than merit. It is power used to defeat change. Over the past five years, this valley has suffered protectionism of one sort: intellectual property laws out of touch with their animating purpose. If trends continue, it will suffer something worse.

... We must stop using Washington to protect against progress."

Brilliant speech.

And another thought provoking article from Miller who writes for MIT Technology review on Less Lethal Weapons.

Disabling enemies without killing them is very difficult. Truly nonlethal weapons do not exist. When the Lone Ranger whacked a bad man on the skull with the butt of his gun, most viewers assumed (incorrectly) that it causes no more than a few minutes of healthy unconsciousness. In reality, such a blow often results in concussion and sometimes in death. It is a tricky business to disable without killing.

The weapons are getting more sophisticated -

Fire pellets - paint balls filled with pepper spray
Stinger grenades - explode with rubber pellets and pepper gas
Nonlethal gas - opiate fentanyl
foam walls of tear gas
Microwave weapons
Long-range nonlinear acoustic devices
and the old standards -
rubber bullets, bean bags, and tasers

Ironically, the strongest argument against these weapons is that they are not terrible enough, and therefore the military (or the police) are more likely to use them.

"It is good that war is so terrible: we should grow too fond of it." Will LLWs make war less terrible, and therefore more likely? Or will they permit us, in essence, to simply shoot the guns out of the hands of the bad guys, and at least open the possibility of winning over the hearts of the children?

One thing the article doesn't mention is the possibility for use for subjugating people. For example the Chinese government could use these weapons and techniques for repressing democracy. South American or African governments can use them for disabling crowds of protesters.

We need government productivity measures

From Brad Delong -
Simon Kuznets decided in the 1930s that he couldn't do it, and was going to assume that the productivity of government workers was constant--never increased at all. There are estimates of productivity growth for non-profits, but nobody trusts them. The standard productivity numbers are for business and for nonfarm business, because those are the only sectors the government is even half-confident it can estimate...

We have some measures of government productivity from the UN - for example the child mortality rate. I would like to see extensive measures of productivity in government which allow tax payers transparently see what they are getting for their dollars. Currently all we get is rhetoric.

Do interest rates matter?

John Kay echos a blog entry I wrote a couple weeks ago. Interest rates are watched intensely but don't have that much effect on long term prosperity. He suggests that watching interest rates is more a matter of trendiness.

"Other government policies - state versus private ownership, competition versus monopoly, the structure of property rights, systems of education and training - mattered far more. But these microeconomic influences are barely debated in financial markets, while even the rumour of a variation in interest rates can change perceptions of the value of companies by billions of dollars. "

"A recent survey by the ECB suggests that in Europe a quarter-point increase in interest rates might reduce consumption by up to a tenth of 1 per cent, and business investment by a bit more - perhaps as much as 1 per cent over three years. These modest responses are what we should expect. There are not many business investments that are attractive if interest rates are at 2 per cent but unattractive if they are 2.25 per cent. And credit charges are only loosely related to money market interest rates. "

Virtual credit-card number used to increase security

Many Internet shoppers are using 16-digit random, single-use, "virtual" credit-card number to increase security.

In recent years, credit-card issuers have been offering virtual account numbers as weapons in the battle against identity theft. Typically the substitute card number has the same number of digits as a regular credit-card number and is free for customers.

In 2003, around 215,000 cases of identity theft, online as well as offline, were reported to the US Federal Trade Commission, more than doubling from just two years ago. Greater amounts of money are being spent online each year. By 2007, about $104.5 billion -- representing more than 5% of total U.S. retail sales -- will be spent online on items such as clothes and electronics, according to JupiterResearch.

Most companies require customers to register your card at the company's Web site. Some require users to download software other require customers to come back to the site to get a new disposable account number before each purchase. In general, the number can be used only at one merchant, whether for a single purchase or for a service with a recurring monthly charge.

American Express Co., one of the first to launch a single-use account numbers, phased out its Private Payments product earlier this year. "We didn't feel that, on an ongoing basis, it was necessary, There are already a number of safeguards for privacy and security out there.", says Judy Tenzer, an American Express spokeswoman.

The World's biggest problems

Copenhagen Consensus, a project that will bring nine of the world's top economists to Denmark later this month. The expert panel -- including four Nobel Laureates -- will create a prioritized list of opportunities to solve the 10 greatest challenges facing humanity, as we see them.

The economists will examine the costs and benefits of solutions to each challenge. An example of a solution (to the challenge of communicable diseases) could be to provide free mosquito nets to areas affected by malaria. The result will be perhaps the grandest "To Do" list the world has seen, showing us how to spend our money the most efficiently. The list will be concrete, outlining tangible opportunities that can be done today.

* Climate Change
* Communicable Diseases
* Conflicts
* Financial Instability
* Malnutrition and Hunger
* Education
* Governance and Corruption
* Population: Migration
* Sanitation and Water
* Subsidies and Trade Barriers

There is nothing wrong with this list. I think they are excellent. I would restate the question to "What would stop humans from existing 10,000 years from now?"
Some things fall right off. Subsidies and Trade Barriers, for example.
Here is my modified list of things that could wipe out humanity.

* Conflicts
* Undue Environment destruction

Unfortunately the Bush administration through its unilateral actions is ignoring both of these. The Bush approach to peace is “peace through strength, not peace through paper.” If that means shredding two decades of international arms control agreements (most of which were negotiated by Republican presidents), so be it.

Bushes approach to the environment has been backing out of the Kyoto protocol, an international agreement to reduce emissions of greenhouse gases, and for weakening rules intended to reduce emissions from fossil fuel-powered energy plants in the United States.

Microsoft will kill Google... I mean service the customer...

Interesting interview with Gates about his "think week".

At one point he says that "we are going up against google". And then backtracks to - we are going to service the customer in ways they never expected.

Service the customer in ways never expected. Sounds like he is justifying integrating search functionality right into the OS.

Nothing much new here:
* He thinks machine translation is ready for commercial use
* Wireless webbed networks are important
* Webservices
- Yawn

Standards with patents

I really don't understandand how standards commitees can allow patents to be approved that contain patents or are so surrounded by patents that they are effectly proprietary. An example is MPEG4.

Is there a standards standard? Which standard meet the grade ISO, IETF, IEEE?

Do any standards committes have regulations where you need to be approved? For example check that an implementation of a standard meets the grade? This is effectively what happened with the Java Microsoft lawsuit. Microsoft was creating an incompatible version of the proprietary - closed Java.

knock knock please open your port

Ran into the knockd program - neat concept.

http://www.zeroflux.org/knock/

Basically it waits for a socket connection on a specified port which feeds a certain bit combination. Once it receives the combination it opens the port such as ssh to the requestor. The port is by default closed. This would allow you an additional layer of defence on your linux box. New ssh security hole? You've just bought yourself the day or two that it will take to get a patch.

Why not implement it in a language not subject to buffer overflows such as python? Why not implement the knocking as a web service(rest)?

Sales for startups

Interesting set of slides from MIT Forum on sales for startups.

In seems geared to mid sized (5-300K) purchases in selling to business (IT).

Sony's new ebook -LIBRIé

Nice review of sony's ebook.

Major issues:
1) Screen refresh is slow (particularly when searching)
2) Proprietary as hell - can't use anything but Sony rental books

Overall reviewer likes it - which gives a thumbs up signal for philips and eink. It sounds like they have the technology pretty much there. Now if they can get the price point down and eink and phillips don't decide to kill the technology by over charging for it we may see ebooks take off.

I would add a small LCD screen (if the screen refresh is a problem), give the device to have some other functionality such as an mp3 player, and then get the price point down to 300-500 USD.

Throw in Gutenburg connectivity for free and an interface for something like audible.com and oreilly's safari. Definitely the ability to read pdf, html and txt files is key.

Perhaps they should license the technology to Apple for an"iPod eBook".

Innovation at the Interface

MIT has an interesting webcast featuring a couple of outstanding innovators Brooks and Langer talking about the sources of innovation.

The argument is that there are two main sources of innovation.

1) Breakthrough technological change
2) The Interface of different changing technologies

Langer gives the best example of (2) by combining material science with medical research.

Rodney Brooks takes the “play approach”. He likes to combine stuff that he finds interesting and tries to find a use for them.

Personally I think the "play" approach combined with scanning many technologies can lead to breakthroughs. It is hard to predict innovation - the play approach allows you to waste a lot of time - but at the same time find channels that perhaps were not expected. "Playing" also means looking at other types of technology and perhaps leveraging the interfaces.

Unfortunately as our society becomes more and more specialized it doesn't allow for the "renaissance man" people who are required for this type of innovation except for outside the corporate environment (universites?). You simply don't see that many job ads for Biologist/Electrical Engineer or CFA/J2EE developer/Biological Systems Expert.

The exception are the larger consulting companies - Mckinsey etc. They understand that innovation requires different backgrounds and perspectives. Companies have a hard time sifting through candidates. Typically they fall back on stupid measures. Ie must have done the job for 3 years with a large menu of skills with a degree from a recognized brand name school.

John Udell Interview

Nice interview with John Udell over at http://www.itconversations.com/.

He covers:

* Dynamic languages are missing from .NET and J2EE (in particular python)

"Python an interesting example is that it is strongly typed, but is also dynamically typed, and what that basically means is things don’t have types until you assign something to them, so an object can become anything, but once it becomes something, it matters, and Python cares about what the type of that thing is and will complain at runtime in the future if you try to assign an object of a different type"

* Why REST - emailable, human readable web services are so important
* Security has lagged - particularly email security
* More IT choice than ever before
* Email is overused - RSS and other technologies should replace most of it
* The Linux Desktop has a way to go except for Tivo
* XML based web gui's are hot - Flash, and Macromedia’s Flex & the Laszlo product