In this Scientific American article, Armando Fox and David Patterson discussed an unconventional approach to build reliable computing systems. Instead of focusing on improving software and hardware reliability, they consider failures inevitable. Their recovery-oriented computing (ROC) approach focusing on bringing the service back quicker. Some method such as micro-rebooting is considered.

Their study also revealed that operator errors cause most of the system downtime. Perhaps the most important boost to reliability is to improve system usability.

2003.05.31 [tech, computer] - comments

A great interview about Municipal broadband from broadbandreports.com (formerly dslreports.com). Let down by the pitiful "broadband" services provided by incumbent cable and phone companies, legal expert Jim Baller put forward an alternative, municipal broadband. He assert that municipal is in a great position to provide broadband connectivity that private corporations has failed to deliver. And that we should expect superior services such as truly high bandwidth with FTTH (Fiber To The Home) and universal access to communities deem unprofitable for private companies.

Can we trust the government to provide connectivity services instead of private companies? Jim look back to the history of electrification of the United States. When the technology was first invented in the late 19th century, private power companies were concentrating on providing services to the most profitable metropolitan areas. Countless small communities were literally left in the dark. They wired themselves by establishing public utilities. Despite being discredited by the commercial interests, public utilities in general faired well and provides reliable services at an affordable price.

How can municipalities accomplish what private sector could not? Jim explained that local government have different goals and are not tied to producing short term profit as in private sector. Instead they see this as an vital infrastructure for the economic development. Even if the service does not turn a profit it can be justified by the benefits they add to the society.

2003.05.22 [network] - comments

From this Mercury News article, the world is facing new breed of infectious diseases. SARS, West Nile, Ebola and AIDS are all emerged within the past 30 years. This is no accident. Scientists have pointed this to human activities and the alteration of ecology.

Among the factors are microbial adaptation and change; human susceptibility to infection; climate and weather; changing ecosystems; human demographics and behavior; economic development and land use; international travel and commerce; technology and industry; breakdown of public health measures; poverty and social inequality; war and famine; lack of political will; and bioterrorism.

Dr. Julie Gerberding, director of CDC, said, "The kinds of things we are doing for SARS, we can anticipate we are going to do again and again."

2003.05.13 [health] - comments

This slashdot posting points to an interview with Joshua Bloch on the New Language Features in JDK 1.5. Generics is a much awaited enhancement. Today, without the language support, I ask every programmer to minimally document the data type used in a comment, like

  Map userTable;   // username(String) => UserRecord

Still many people fail to do just that. I have to track down the usage of the collection to deduce what data type they are using :(

A generic declaration would be

  Map<String, UserRecord> userTable;

This seems to keep Java collection's light-weight design. It provides some compile time checking and implicit typecasting without the complexity of C++'s template. I still remember spending a lot time debugging the COMPILATION in C++. An compiler error message would spend multiple lines with long and unreadable class names decompose of a template instances...

One reader has pointed out the work on JSR 166 Concurrency Utilities led by Doug Lea. Doug Lea's book Concurrent Programming in Java and the programming library is definitive in this area. It is great that it gets formalized and included in JDK.

2003.05.09 [software, java] - comments

I have been using O'Reilly Network's Safari Bookshelf for over a month. The Safari Bookshelf is an electronic reference library for programmers and IT professionals. It provides online access to over a thousand titles from O'Reilly as well as several other major publishers, like Addison Wesley.

The first question comes into mind is how does it compares to reading a real book. Granted it is more comfortable to read a long essay on paper given the computer screen's limited size (for me a resolution of only 1024x768). On the other hand it is not quite as hard. In the full screen mode usually a full page of information can be displayed. While I cannot flip through the pages as with a book, I can search and navigate with just few mouse click. Moreover I think the reading ability can be trained. For people who read mostly from books and newspaper they would find the online reading experience rather restrictive. For me, after spending a lot of time reading online I have developed ways to navigate and organize online material. I find the Safari format quite satisfactory.

The biggest advantage does not come from the presentation format nor how much money one can save for not buying books. One day I come across a topic on WIFI security, which I am not familiar with. I checked on the 802.11 Security title, browsed the introduction and a few other chapters. Within half an hour I got some basic understanding and a grasp of issues. I realized Safari is a wealth of knowledge under my fingertip. This is a truly 21st century enlightenment.

2003.05.08 [tech, book] - comments

My server is now using its idle time to help search for a cure for SARS. The infectious diseases has already causes hundreds of deaths.

This project is hosted by the Rothberg Institute. The software can be downloaded from D2OL. The program is using models to test the effectiveness of potential drugs to fight the SARS virus. Scientists like to screen millions of drug compounds. This would requires even a supercomputer a long long time to accomplish. Instead, similar to the SETI@home project, they are calling for volunteers around the world to use their computer's idle time to solve a part of the problem. Tens of thousands have already signed up by running the software on their computer, forming a global distributed computing system to attack this problem.

We may not find a magic cure easily. Still I hope we can pool the resource for a worthy cause.

A candidate molecule

2003.05.03 [tech, health] - comments